Canon C300 Mark II Lab Test – Dynamic Range 2 Stops Less Than Expected

canon-c300-mark-ii-review-featured

The Canon C300 Mark II is here and comes with a lot of high expectations by customers and a hefty price tag of $16,000. Many are willing to put down the investment as they anticipate superb image quality in return and a new flagship cinema camera that has been several years in development by Canon. In our test lab we took a closer look at the C300 Mark II’s performance and found it less powerful than expected.

[UPDATE:] Canon responded to our article.
Read: How Canon Measured 15 Stops of Dynamic Range on the C300 Mark II

At cinema5D we conduct camera reviews and comparisons. As manufacturers are not limited in the way they advertise their camera’s performance we strive to put the numbers in perspective with unified tests on the latest cameras on the market.

Being a shooter myself I know it’s actually really hard to objectively pinpoint the performance of a camera. In this Canon C300 Mark II lab test I want to give you some insights into our findings about the new Canon C300 Mark II and show you how we tested.

canon-c300-mark-ii-review-03

The Canon C300 Mark II’s reputation

The Canon C300 Mark II has raised a lot of eyebrows since the announcement in April. The first Cinema EOS camera, the C300 has been very popular and remains the go-to option for many TV, film and independent productions. With the C300 Mark II Canon introduced numerous advancements, most notably 12-bit 2K as well as 4K 10-bit internal recording and improvements in terms of sensor technology and image processing. According to Canon:

“a 15-stop dynamic range is provided by a new photodiode design that simultaneously lowers the noise floor”.

Considering that the defacto cinema standard Arri ALEXA was announced with 14 stops of dynamic range the Canon C300 Mark II comes with a bold claim. The information surrounding the new Cinema EOS camera seems to suggest that Canon’s new sensor could actually compete with the image quality of the good old Arri ALEV III sensor that is used in all 9 versions of the ALEXA Arri has released over the years.

As with any new camera on the test bench I was curious, but eventually what I found during the lab test wasn’t what I was hoping for.

canon-c300-mark-ii-review-05

Testing the Dynamic Range

So I looked at the dynamic range. At cinema5D we measure this using a DSC labs XYLA-21, an LED-backlit transmissive chart that displays 21 stops of dynamic range. Each vertical bar represents one stop of light. This way it’s very easy to judge dynamic range just with your eyes. At the end we evaluate the recordings with a software by IMATEST that spits out a dynamic range value. There’s some more science behind it, but I’ll spare you the details.

As we are recording each ISO value with each camera using the identical very sharp Zeiss 50mm CP2 T/2.1 makro lens with interchangeable mount we can compare all cameras to each other. In our database we have about 20 cameras on record so far.

In our tests and according to our workflow it turned out the camera actually has 12.3 stops (measured) of usable dynamic range.

Usable dynamic range. What is that? That means within this range you have picture information that you can use. Anything beyond a certain “Signal-to-Noise” ratio is so noisy we think it’s unusable. I must say 15 stops of dynamic range is something I cannot find in the Canon C300 Mark II. There’s always a chance there’s some hidden setting in the menu to unleash the HDR potential of the sensor, but I couldn’t find it.

canon-c300-mark-ii-review-08

Reading comments of other camera enthusiasts it seems that many expect the new C300 sensor to hold up to the performance of the ALEXA cameras. What I see is that the C300 Mark II is still far away from matching this reference standard and almost on par with the Sony FS7 (measured at 12.4 stops).

An additional surprise was that I saw a lot of noise in the dark areas on the Canon C300 Mark II even at its base ISO of 800. While the camera has a very clean image in the brighter areas and has a really neutral tone with minimal color noise, there is a point in the dark areas where the noise kicks in strongly. When I did some test shots of natural subjects I realized that images shot on ISO 800 that are just 1 stop underexposed can quickly become a little too noisy for my taste especially in comparison to a camera like the Arri AMIRA (our reference camera). That is rather the opposite of what I expected after all the talk about the Canon C300 Mark II sensor and processing being so new and advanced. Ultimately one has to admit the 5 year old Arri ALEXA sensor is doing better…

Looking at the Test Charts

Below you can see 3 different cameras compared with the method explained above. The Canon C300 Mark II, Arri ALEXA and Sony FS7. You can see the usable dynamic range highlighted in red. The two fields (stops of light) to the left are overexposed, the range to the right is too noisy (underexposed).

dynamic-range-C300-mark-ii-vs-fs7-vs-alexa_B

It is interesting to see how close the Sony FS7 and Canon C300 Mark II are in terms of dynamic range performance, while the Arri ALEXA clearly has a wider dynamic range and much better noise performance. Furthermore you can see that the ALEXA has a much nicer highlight rolloff and a very nice “looking” noise overall even in the far blacks. The Canon C300 Mark II has very little color noise, which is nice, but at one point in the darks, the noise is getting very severe. You can see this much better in the following image. These are the same shots with gamma lifted equally so we can better see what is happening in the shadow areas, or what would happen when we push the image in post:

dynamic-range-C300-mark-ii-vs-fs7-vs-alexa-gammalift_B

Camera Settings:

  • Canon C300 Mark II: C Log 2 – Cine Gamut | ISO 800 (native) | F/4.0,5 | 4K (downscaled to 2K) 10bit
  • Arri ALEXA: Log C | ISO 800 (native) | F/2.8,5 | 2K, 12bit
  • Sony FS7: Slog 3 (EI) | ISO 2000 (native) | F/5.6 | 4K (downscaled to 2K) 10bit

I am aware that some people might have questions about our tests. Here are a few answered in advance:

  • Why did I compare 4K to 2K? I decided to shoot the Canon C300 Mark II in 4k (not 2K) for this test as I felt the image looked better when downscaled to 2K. I am aware Canon says they achieve best image quality at 2K in 12 bit. I did not see any improvement for the tests conducted. 12 bit will not give you a higher dynamic range, only better gradations.
  • Why didn’t I shoot the Sony FS7 at ISO 800 also? The Sony FS7 has a base ISO of 2000 and performs best at this speed. I did not see any improvement in dynamic range shooting at lower ISO’s. Same goes for the Canon by the way.

C300-mark-ii-horizontal-stripOne more thing I’d like to mention is that strip of light that is visible on the Canon C300 Mark II recording. In the very dark areas you can see a faint horizontal strip across the image coming from the opverexposed fields on the left. I have no explanation for that phenomenon.

High ISO’s?

The C300 has a reputation of producing nice lowlight images and the Canon C300 Mark II is said to be even stronger in that regard, with an available ISO range of up to 102,400 it sounds promising. In my initial tests however I felt that the camera also didn’t perform as good as I expected. Personally I thought I would probably not go beyond ISO 6400 for most projects, which is more or less in the ballpark of the Sony FS7’s lowlight capabilities. Other DP’s might of course go much higher.

High ISO Canon C300 Mark II vs Sony FS7

When I compared the Canon C300 Mark II and the Sony FS7 with the chart I found that indeed they look quite similar in terms of usable range. The Canon I would say performs only slightly better at the same ISO speeds. You should know though that ISO values and the way they affect the image brightness are often varying on different cameras. Thus I had to open up the aperture about 1 stop more for the Sony to get the same exposure, meaning the Sony FS7 is actually about 1 stop less light sensitive at this ISO speed.

lowlight-C300-mark-ii-vs-a7s_B

Camera Settings:

  • Canon C300 Mark II: C Log 2 – Cine Gamut | ISO 12,800 | 1/840th | F/4.0,3 | 4K (downscaled to 2K) 10bit
  • Sony FS7Slog 3 | ISO 12,500 | 1/1000th | F/2.8 | 4K (downscaled to 2K) 10bit

canon-c300-mark-ii-review-01

Conclusion

Even though the Canon C300 Mark II seems to be quite on par with the Sony Fs7 in terms of dynamic range, I must also say the image of the C300 Mark II is more neutral and the noise is less saturated and mushy in comparison to the Sony FS7. So while unfortunately there is a lot of noise in the shadows, at least it doesn’t look so bad. Also the C300 Mark II seems to be about 1 stop more light sensitive than the FS7 at high ISO speeds.

I must also mention: Curious about the 120fps (crop) mode in 2K I found that it’s actually very soft (think 720p) making it much less usable. Personally I expected much more from the Canon C300 Mark II.

Of course there are other aspects to consider about this camera, many of which are not related to the tests discussed in this article. A big point is that the internal codecs are much stronger than on the Sony FS7 and thus it can retain details much better. Also the ergonomics are different and can be a buying argument if you have gotten used to the first Canon C300. There is built-in Genlock and a few other features professionals might enjoy. Check out this comparison of specs between the C300 Mark II’s main contenders for a basic overview.

Personally I think the FS7 and C300 Mark II can compete well in terms of overall performance, but the Sony FS7’s 240fps and other external recording options make it very competitive in regards to the Canon C300 Mark II that requires a much bigger investment. Eventually it is up to every user to decide which features and aspects of each camera’s performance is important to them.

We’ll conduct some more tests at cinema5D and keep you up to date. Subscribe to our newsletter to stay informed (no spam guaranteed), or like us on facebook if you enjoyed this article.

We hope you liked this Canon C300 Mark II lab test. Please do share your own observations and thoughts about this article in the comments.

Note: We contacted Canon to get a response about our findings. A Senior Canon Representative involved in the development of the camera came back to us and let us know that they will wait to assimilate more technical evaluations before commenting.

avp-logoThanks to AV Professional for lending us their camera.

 

Tests conducted to the best of our knowledge. Errors and omissions excepted.

 
BUY IT
€11,340.20 exc. VATOrder now and we'll deliver when available - Normally ships in under 7 working days but may...
 
BUY IT

Leave a Comment

You are not subscribed to this post. Follow new comments

Login to comment

Reply
Einar Davíðsson September 24, 2015

Very interesting findings. I’m curious if there was any difference between Canon LOG and Canon LOG2 as regards Dynamic Range?

Reply
Einar Davíðsson September 24, 2015

I’m also not tech-savvy enough to know weather the chosen Color Gamut would effect the dynamic range.

Also, as I understand it, the native ISO of the old Canon LOG is now 400 on the Mark II. Was the old Canon LOG tested at all?

Sebastian Wöber Reply
Sebastian Wöber September 24, 2015

The old LOG wasn’t tested thoroughly. Quote from Canon: “Canon Log 2 – optimized for the 15-stop dynamic range of the C300 Mark II”

Barry Goyette Reply
Barry Goyette September 24, 2015

Many of Canon’s reps have been quoted as saying that the original C-Log has “15 stops” as well. Likely without the lifted shadows of C-Log2.

Florian Klaes Reply
Florian Klaes September 24, 2015

I really would like to hear Canons explanation for this otherwise people have to reconsider their preorders. Huge huge bummer.

BX Miklos Nemeth Reply
BX Miklos Nemeth September 24, 2015

Why would they do that? The Canon C300ii is at least as good as the FS7, and it’s a “Canon C” camera, the preferred camera for many producers. I can see the result of this tests absolutely supportive to my decision of buying a C300m2

Sebastian Wöber Reply
Sebastian Wöber September 24, 2015

Indeed I wouldn’t worry too much about the strip. It’s really in the darks, but for some professionals it might be a problem. Will have to be investigated further and see how apparent it is in live environments.

Barry Goyette Reply
Barry Goyette September 24, 2015

Sebastian,

thanks for the test…still evaluating the whole of it, but on the strip…. In all the demos we attended, we saw this strange patterning in the very darkest tones in every ISO (really below what would ever be used in a shot-for instance underneath a wheel well on a jeep in a top lit set at the Canon booth). The curious thing was that the pattern changed dramatically from iso to iso. What would be a band in one, would be a series of streaks in others, or would go vertical in the next. There was always the assumption that Canon would mask this in the final camera. (which they apparently didn’t). Looking at your 12,800 strip, magically it’s gone. Did you run DR strips at various ISO’s?

Sebastian Wöber Reply
Sebastian Wöber September 24, 2015

Interesting!

Florian Klaes Reply
Florian Klaes September 24, 2015

I have pre-ordered this camera and one of the biggest claims and the most important one was the Dynamic Range…Canon claimed the price tag this high because it should be far superior to the Fs7…hearing thinks like the strip or a noisy black even at low isos or a soft cheap looking slow-motion (heared some voices about this softness before)isnt something that i would call “far superior”… i could have sworn that when canon claims 15 there are 15 usable stops…i mean if black magic says 15 i take that with a grain of salt.It doesnt matter if we are talking about +-1 Stop but more than 2 stops is indeed a thing for me…next to these other wunderfull “features”.

Florian Klaes Reply
Florian Klaes September 25, 2015

Do you have the chance of comparing this unit against another c300 mk2, a c100 mk2 and the old c300 mk1 in terms of noise levels. Thanks for the tests.

Alex Leith Reply
Alex Leith September 24, 2015

Very interesting article… I saw the highlight smear on my C300 (original) too. It’s reminds me of the old days of CCDs (except the smears went vertically with those). Subjectively I’d say that the C300 looks a stop noise than the FS7, but the C300 noise looks prettier.

Reply
Joseph Wilkins September 24, 2015

As someone with one of these cameras on pre-order this is VERY disturbing news. Canon better have some very quick responses to this. Thinking of canceling my order and going to get an FS7… URSA Mini, or Red Raven. Too many lower priced options out there for this not to live up to EVERY claim canon has made about this camera.

Bgd Videography Reply
Bgd Videography September 24, 2015

Interesting. That strip worries me! Maybe canon use the same testing equipment as Volkswagon. I have one on pre order, but would like to see Canon’s response before cancelling. Also, I’d like to see real world tests.

Alex Leith Reply
Alex Leith September 24, 2015

The original C300 had the strip too… I had two C300s (and a loaner while I had to get the first one fixed) and all of them had that issue. It wasn’t a problem in 99.9% shots.. but occasionally reared its head when shooting in a dark room with a light in a doorway.

The Canon Cxxx series had that problem in high contrast scenes; the Sony F100/700 had highlight aliasing on high contrast edges; and the BlackMagic cameras had the black hole in high contrast scenes…

Possibly a sensor technology limitation that is being worked around slightly differently by three manufacturers?

Alex Leith Reply
Alex Leith September 24, 2015

I should add that with the highlight strip – you REALLY had to be pushing the grading a long way before it actually interfered with the shot.

Reply
Kemalettin Sert September 24, 2015

Why didnt you put Dragon too?

Sebastian Wöber Reply
Sebastian Wöber September 24, 2015

I think it’s really a different audience looking at the RED Dragon cameras.

Rodrigo Prata Reply
Rodrigo Prata September 24, 2015

Different then the ones that shoot on the Alexa? Don’t forget about the upcoming Raven camera, that will be more affordable then the C300MK II (even when kited out with RED acessories)

Sebastian Wöber Reply
Sebastian Wöber September 24, 2015

The ALEXA is really the reference, that’s why it’s in there, though there is emphasis on C300 Mark II vs FS7.

Reply
Trevor Meier September 24, 2015

A Weapon or Scarlet Dragon reference at 4K would be useful as a stand-in until Raven get’s released. Be good to see where that camera might sit since it’s competing for a similar price point.

Roberto Alieno Mettifogo Reply
Roberto Alieno Mettifogo September 24, 2015

sensor made by wolkswagen?

Mel Bourne Reply
Mel Bourne September 24, 2015

A canon sensor fails to reach 2015 standards?
Say it ain’t so! :D

Alejandro B. Martin Reply
Alejandro B. Martin September 24, 2015

Great article. Do you have a tests results of Dragon sensor (epic and or scarlet dragon) ?
It will be great to know aboute the Dynamic Range of Red compared to

John Marchant Reply
John Marchant September 24, 2015

The horizontal strip is a CMOS smear. Common to all cmos cameras to varying extents, depending on sensor design and gain levels.

The counting of stops is off by one in the shadows IMHO.

At the highlight end you should hard clip the first two chips, then stop down until the second chip is on the cusp of clipping. You then count from chip two to three as your first stop. So far so good.

At the shadow end, the last clearly discernable chip is not the last stop – you’re counting between the second last and last visible for that chip. There is another stop at minimum in all cases between the last visible chip and black.

On that basis with a lifted gamma I can go along with Canon’s claims. Arri’s claims are very conservative on DR. Most other manufacturers go with a best case methodology.

Sebastian Wöber Reply
Sebastian Wöber September 24, 2015

Hi John,
That’s a legitimate approach, different to ours. I like to believe ours with two windows is “safe”. If you are open to use the additional range in the shadows as you suggested you should be aware that you will see this kind of grain in your shots eventually.
I think our S/N threshold is conservative enough and at the end of the day the software decides on total range which we confirm with a subjective evaluation.

Barry Goyette Reply
Barry Goyette September 24, 2015

Hi Sebastian,

When I first viewed the test on my phone at 5am with my blurry, morning vision, I noted that it appeared as if the exposure given the C300 II was somewhat less than that given both the FS7 and the Arri, due to the amount of flair around the clipped patches on the left of the chart. Now that I’ve got this file open on my desktop, it does seem that this is the case, measuring the first stop approximately 1/4-1/3 stop darker than the other two cameras. I’m wondering your opinion on this.

Another question is the raised black level of the Arri Strip. I’ve seen other transmission DR tests of this camera and gamma, and they don’t exhibit this. An elevated black at this end of the range would affect noise readings. When I normalize that chip, the apparent DR difference shrinks by about a half stop.

Still disturbing is the CMOS smear extending well into the 11th stop on the C300 II. Viewing on a calibrated monitor, it’s not as bad as it looked on my laptop, but it’s certainly something (actually, to me, the only thing) of concern.

Were these strips shot at the same time? I’m curious as to the color shift on the FS7.

Finally, I’m curious about how you determined that the C300 had .1 stop less DR than the FS7. When I look at the charts, I see very little difference. The Sony seems to have more color noise in the shadows versus the C300 having more luma noise (which I think is generally considered preferable), but the canon definitely seems to be seeing a stop or or more into the noisiest part of the images. Realistically .1 of anything only matters in a penis measuring contest…but I’m just curious where that .1 came from.

Thanks again for the test. I’m sure it’s valuable to everyone looking at this camera. While transmissive tests like this are the simplest/most accurate/comparable tests of DR, they also give us very little information regarding how the camera responds to colors and detail across that range. (something Geoff Boyles +7/-7 color chart test did). I look forward to additional tests in this respect to gain a fuller understanding of what this camera is capable/not capable of.

Barry

(one final question — do you have a link to the step chart from the original C300 camera. For those of us looking to upgrade, this would be an important piece of data. I looked and couldn’t find it.)

Sebastian Wöber Reply
Sebastian Wöber September 24, 2015

Hi Barry,

I haven’t raised black levels on the Arri. What you see is exactly how it comes from the sensor, other than I raised white levels to max for all cams in the visual.
For the gamma lifted version which is only just for people to get an idea of the differences, the Arri black level was slightly lowered to put things into perspective.
Lastly you cannot manipulate the image and then measure again. The measurement software measures what is in the image. The software measured the S/N ratio steps over the treshold .1 stop later than on the Canon. I think the difference is negligible.

Barry Goyette Reply
Barry Goyette September 24, 2015

Not saying that you’ve raised the black levels on the Arri, only that they are raised compared to the other two cameras, and that other transmissive tests of the Arri that I’ve seen don’t show this. As you mentioned (relative to your gained versions), normalising the blacks does increase the appearance of the noise.

I agree that the .1 difference is meaningless, except on the internet. :-) C5D’s methodology is to involve a visual check of the imatest data, so I just wanted to hear you say that it was the machine that spit out the .1 difference. thanks.

I’m curious what you think of the exposure issue I raised.

Conny Fridh Reply
Conny Fridh September 25, 2015

Hi,

Have you done the same dynamic range test on the old C300? If so, what’s the dynamic range on that one? As this is done now, it actually sounds like the two C300 & C300 IIhave the same dynamic range.

Is it so?

Alex Leith Reply
Alex Leith September 24, 2015

By that reckoning I count 16-17 stops on both the C300ii and the FS7… But you’d never seriously use those last stops – they’re too noisy to be counted as useable DR. I think what Sebastian and the C5D team are doing is setting a benchmark of acceptable signal to noise levels, and testing the cameras comparatively using that same benchmark.

Alex Leith Reply
Alex Leith September 24, 2015

What Sebastian said :-D

Shangwei Liu Reply
Shangwei Liu September 24, 2015

The big brother is the big brother

George Sealy Reply
George Sealy September 24, 2015

Thanks for this test. It really helps in doing pro and con assessments when technology and price points are changing so fast.

Art Sanchez Reply
Art Sanchez September 24, 2015

Thanks for testing it out!

Jeff Regan Reply
Jeff Regan September 24, 2015

Geoff Boyle of CML tested the mark II at 14 stops and used it for a commercial(pre-production unit). He was also impressed with the autofocus features. The mark II is supposed to be 10db cleaner than the FS7, 67db vs. 57db, which in turn is supposedly a little quieter than an F5. It’s clear that Canon has the preamp gain set higher than Sony, explaining the 1-stop difference in sensitivity.

I would like the author to write a bit more on the perceived differences between XAVC and XFAVC codecs. Is there more high frequency detail in the latter?

Rastko Vukovic Reply
Rastko Vukovic September 24, 2015

You say that you have about 20 cameras on record so far. Where can we find those tests?

Sebastian Wöber Reply
Sebastian Wöber September 24, 2015

Haven’t gotten around to creating a platform, but we plan to do that, so you can compare all cameras yourself manually.

Aaron Almquist Reply
Aaron Almquist September 24, 2015

1: To have a more accurate assessment of what the c300 mk ii is capable of I would have clipped the first two chips and shot in 12bit (4 times the amount of gradation than what was tested in. If you’re going to also compare to the Alexa which was tested shot in 12bit you need to match.

2. Please elaborate on your post process and how you pulled and pushed the “film”

Sebastian Wöber Reply
Sebastian Wöber September 24, 2015

Hi Aaron,
As mentioned in the article I also shot C300 mkII in 12bit 2K. The tests had the same results so I went with the 4K as I found the grain looked nicer.
I just raised gamma to 1.8 and on the ALEXA I lowered the blacks a little to make the grain more visible. There is no post process on the charts other than exporting an iFrame and importing in imatest.

Reply
sanveer mehlwal September 24, 2015

Sebastian Wèber you broke Barry Goyette’s heart. Couldn’t you have just cheated a little on the tests???

Heartless technical genius. Hehehe

Barry Goyette Reply
Barry Goyette September 24, 2015

Hardly my friend. I know you think you know me..but you don’t.

Reply
sanveer mehlwal September 24, 2015

hehehe. It was sarcasm.
But, you called me a friend, so it’ll all good as new. :-)

Clayton Moore Reply
Clayton Moore September 24, 2015

First let me say off point, expect to see some interesting acquisitions in the next 24 mo. Not all manufactures of cameras are on solid financial footing – its not being discussed that much publicly but its true.

The Canon C line and the 300 in particular has always been out of my price range personally but thats only because I don’t shoot at a level that could easily pay for that. For people who’s work can pay for it, Canon has done very well. Its a combination of an overall solid performer and ergonomics.

That having been said, I do think there are opportunities for the competition to eat some of Canon’s lunch in the next 2-3 years if only due to the fact that the competition does not tend to be as slow and conservative as Canon.

Reply
Einar Davíðsson September 24, 2015

Looking at this, the most striking thing is, once again, how much better the Alexa is then everything else. Not just the DR, but look at the noise, both the amount and character of it…And yet it’s 5+ year old tech. Can someone explain to me why companies with much larger R&D departments and more to win are unable to reach them?

(Yes, we’re comparing cameras at vastly different price points, but this also holds for other manufacturer’s flagship cameras).

Barry Goyette Reply
Barry Goyette September 24, 2015

I think it’s a really good question. Again, this type of test doesn’t tell you a lot. The appearance of the noise in the Arri chart is being affected by the atypically raised black levels of Sebastiens test. When we normalize the black levels with the other charts, we see about 1/2 stop less DR, and possibly some noise reduction being applied below the 11th stop (11th stop is noisier than the 12th stop). The Arri has a 16bit ADC compared to the C300 II’s 14bit ADC, so that should explain some of it, and we know that in some of canon’s still cameras they don’t typically apply any Noise reduction in the raw files, whereas some other manufacturers do. If we go back a few years and look at C5D’s comparison of the C300 and 5D mark 3, we see that noise reduction/low detail can skew DR tests like this that primarily look at a certain S/N ratio.

Sebastian Wöber Reply
Sebastian Wöber September 24, 2015

No, I did not raise black levels!
And no, you cannot raise black levels and get a better result. You can’t tamper with the footage before measuring it.
Yes, NR will definitely change the test results. But often you cannot control this, so the footage is what the camera operator gets and the test chart shot is what the software gets. A camera might have better dynamic ratings due to NR, but then this should be mentioned as the shots might not look acceptable due to NR.

Sebastian Wöber Reply
Sebastian Wöber September 24, 2015

I believe there is always some kind of NR going on even in the most professional cameras. I know the C300 mark II for example has a NR that is happening directly on the sensor. This could be the reason why the noise kicks in so suddenly.

Barry Goyette Reply
Barry Goyette September 24, 2015

Definitely NOT suggesting that you’ve done anything here with the black levels, only pointing out that they are elevated compared to the other cameras, and other transmission tests I’ve seen of the Arri. http://www.provideocoalition.com/alexa_dynamic_range_its_all_in_how_you_use_it

There are lots of reasons black’s could be elevated. Camera Gamma, Setup level, flare…

I’m not sure how much an elevated black would affect an imatest result, but it definitely can affect a visual examination of the test, which is all we have on this side of the wall.

Sebastian — I am not questioning the veracity of your tests, only asking questions about what I see. I’m here to learn.

Reply
Jim Martin September 25, 2015

Two questions, when did you do this test and was this a prototype camera or was it a production line camera (that just arrived last Friday here in the US)?

Florian Klaes Reply
Florian Klaes September 25, 2015

That would be nice to know + do you have the chance of comparing this unit against another c300 mk2, a c100mk2 and the old c300 mk1 in terms of noise levels. Thanks for the tests.

Tor Johansen Reply
Tor Johansen September 24, 2015

Sebastian – Excited for you compare and evaluate the new Blackmagic 4.6K sensor, hopefully this will happen soon!

Reply
Crimson Son September 24, 2015

How did you record the signal from each camera?

Sebastian Wöber Reply
Sebastian Wöber September 25, 2015

I recorded it in the best / highest compression (codec) and format the camera could achieve internally. Furthermore tested several different formats (2K 12bit, 4K 10bit) and compared and chose the best performing one (for this particular test) which you can find in the article.

Ben Annesley Reply
Ben Annesley September 24, 2015

I think you could give the FS7 a bit more based on that chat and the noise

Martin Del Vecchio Reply
Martin Del Vecchio September 24, 2015

In the first test chart photo, one would expect the right-hand side to be 100% black for all three cameras. But it’s not; the Arri ALEXA is distinctly lighter than the other two, all the way to the edge.

So is the ALEXA overexposed by a stop? According to the camera settings, the C300 was at F/4.0, while the Arri was at F/2.8. Same ISO.

Am I missing something?

Sebastian Wöber Reply
Sebastian Wöber September 25, 2015

Hi Martin,
The ALEXA has a different Gamma curve. It is irrelevant to the test software. The software evaluates noise for each individual patch, no matter if you brighten it up or bring it down. If it helps for you to evaluate visually you could lower the black levels in photoshop.
The ALEXA was not overexposed. the F/2.8 merely indicated that the ALEXA has more latitude towards the highlights before it clips.

Reply
Kaster Troy September 24, 2015

Great article as usual, thank you. Pretty much just as I expected. I knew the DR wasn’t as high as they claimed, and that ISO is sure disappointing. Still can’t get over the cropped slo-mo but for $16 what do you expect ;)

Reply
Adrian Bacon September 25, 2015

I want to see a comparison between the C300 and the C300 Mark II using the same test. That will tell you how much the dynamic range has increased by with the Mark II.

Comparing it to the Alexa and the FS7 is useful, but frankly, subject to your interpretation of what you consider to be usable dynamic range between the cameras.

Doing a C300 to C300 MkII comparison and using the same interpretation of useful dynamic range will show where the improvements are, and since the original C300 has been around for a while, how it performs is fairly well studied and documented, that will give a good interpretation of the improvements to the Mk. II.

Without that, it is a subjective test that is open to the interpretation of the tester.

Daniel De Silva Reply
Daniel De Silva September 25, 2015

Agreed 100%

Conny Fridh Reply
Conny Fridh September 25, 2015

Yes I also would like to see them do this test. Right now actually. Because to come with a claim like this, you need to back it up all the way.

Reply
GELAX STUDIO September 25, 2015

That is why cinema5D built their TEST LAB!
C300 have tested data,and u can check it and compare it to the C300M2~

Bgd Videography Reply
Bgd Videography September 25, 2015

I too would like to see the c300 and c300 Mark ii tests compared. Personally, I have decided to go ahead with the camera, despite what we’re seeing here. I think the real world tests will yield very different results. And really? Do you think Canon screwed up this bad? They sold this camera to us based on its amazing DR!

Reply
GELAX STUDIO September 25, 2015

Does the C300M2 shoot 120P FHD in 10bit or 8bit ?

Sebastian Wöber Reply
Sebastian Wöber September 25, 2015

No it shoots 120fps 2K 12 bit max.

Conny Fridh Reply
Conny Fridh September 25, 2015

Canon must respond to the facts about dynamic range and The doft slowmotion crop image.

And they have to do It tight now. IF not I Will chancell My order of This Camera.

By The way, what usable dynamic range number did you get from the older C300? Also 12,3?

Thanks
Conny Fridh

Sebastian Wöber Reply
Sebastian Wöber September 25, 2015

We will do a direct comparison article soon.

Conny Fridh Reply
Conny Fridh September 25, 2015

Great,, please do IT soon!

Reply
Tobias Mennle September 25, 2015

Sebastian, about the noise. Any log gamma needs a proper delogging with an S-curve before one can reasonably judge noise. You are dealing with 3 different log curves here, comparing Canon to Sony to Arri. You need to delog and colour correct. On my first test with the FS7 I had noise that reminded me of high speed S-16mm, but it completely vanished when I delogged. Most people just don´t seem to get it, and just start to colour correct log material to taste. You see it in Sonys first FS7 clip, and in Philipp Blooms C300 demo, too. So IMO delog both the C300 and the Arri and Sony footage and then judge the noise.

You write: “12 bit will not give you a higher dynamic range, only better gradations.” Again, you need to judge that in colour correction when manipulating the image. I bet a 12bit log will give you way more dynamic range than 10bit log. You really need to try it and pull up the shadows of both and then compare.

About slomo, Canon is so fair to describe the camera in their white papers, and there it is: At 100 frames per second the bottleneck in the processing is 1024×540 pixel. WOW. I know my FS7 150 frames and they look mediocre at best, but this sounds even worse. 540 Pixel…

I bought an early C300 MK1 against better knowledge, but this one seems even more underspecced.

Nevertheless I am sceptical about your test procedures and seriously doubt they do the camera justice.

Reply
Tobias Mennle September 25, 2015

You write: “Looking at the test charts…You can see the usable dynamic range highlighted in red.” No, we can´t. We are looking at the outcome of three different log curves that need their specific delogging so we can judge the images reasonably. One simple point is, a proper delogging will make the shadows even darker, but also get rid of the noise. Part of the lights will get brighter. Etc.

More here: “Understanding log grading” by Mike Most is a brillant short explanation.
http://mikemost.com/?p=251

If you you for example check Shane Hurlbuts tests, of course he has always has perfectly graded images when he compares cameras, in whatever respect. We can totally forget about log images, we need the rec709 grades.

Please check your test procedure, you are not doing yourself a favour.

Sebastian Wöber Reply
Sebastian Wöber September 25, 2015

Hi Tobias,

Applying a LUT will not remove the noise! It will mask it. Furthermore we are not solely looking at noise levels here, we are looking at the noise in each single f-stop patch. No matter if you brighten up a patch or darken it, the software will always recognize and evaluate the noise for each individual patch and in the end spit out how far in the blacks it finds the noise acceptable (according to our S/N threshold that we apply to each camera equally).

Now, in practice, when you shoot something, do you just expose it perfectly and then apply a LUT and that’s it? No, you grade each shot individually and if the camera you used has a smaller dynamic range then you will more often see noise in your dark areas and blacks and the quality of noise the sensor has will determine how this noise looks that you pull.
That said, there is only so much you can get with any given camera. When you use your mobile phone’s camera (extreme example) you will realize you can’t pull off high contrast scenes very well, because dynamic range is just too low. Your highlights will clip, or you can’t pull up the shadows because they are too noisy. This is why we often need cameras with high dynamic range, especially in those situations when you are unable to control the light. The iPhone has “HDR mode” for a reason.

Feel free to LUT the test shots, but it will defeat the purpose of the test unless you make several versions differently graded and evaluate the results.

Concerning 12 bit: I did the same test on the Canon with 12-bit in 2K and the results (from the software) were the same. And judging with my eyes I found the 4K downscaled more pleasing for this particular test and it made sense to compare it with the FS7 in that way which was also shot in 4K.

Barry Goyette Reply
Barry Goyette September 25, 2015

Tobias — just a correction on the Canon HFR. That 1024 x 540 “frame” that canon is referring to, is a color component frame. Meaning they are breaking down the 2k capture into 4 components (red, blue and two greens) with the pixel dimensions above. Resolution wise this is the same as any 2k sensor. Canon uses this novel process rather than a typical debayer. and it gives them a great 2K image when sampled downward from 4k. In this situation, the process is giving them largely the equivalent of a standard debayered 2k image, meaning something less than their stellar 4k downsampled 2k. It looks very much like what I see coming off the FS7 at 150 and 180 FPS, which is doing something internally to avoid processing the entire 4k image at that rate, probably binning. Two different approaches to largely the same problem.

Reply
Adrian Bacon September 26, 2015

Agree completely with this comment. The dynamic range is extremely subjective and this test does not take into account the post process. I’d really like to see what that chart looks like after it’s been properly graded.

Noise removal, contrast, saturation, tones/curves, sharpness, etc. all the post stuff we do to get it into a deliverable h.264 file.

Personally, I find that I’d rather see a rating of discernible stops, regardless of noise level because I know I’m going to be doing some amount of noise removal in post, and discernible stops gives me a ball park of what I have to work with in that regard.

With that being said, I’m seeing 17-18 stops of discernible range for the c300 Mk II on your chart. After post noise removal and grade, I can usually get a good looking image that is within 1-2 stops of discernible DR with either no visible noise, or dramatically reduced noise.

With that being said, I do believe canon has a fairly solid 15 stops and then some depending on how much post noise removal you’re willing to do.

We’ll see once people start posting graded real world footage. Philip Bloom has already posted some test footage and frankly it looks like there is a lot more DR than the first C300 (which he also has/had).

Clayton Burkhart Reply
Clayton Burkhart September 25, 2015

Sebastian, please see if you can do this same test with the new RED RAVEN when it comes out. I am entirely fed up with their claims that their sensor does 16.5 stops, when in fact I have noticed that it has never surpassed the DR of an Alexa, which as we know is a touch more than 14 stops.

Sebastian Wöber Reply
Sebastian Wöber September 25, 2015

Hi Clayton,
Good to see you around!
We will most certainly test the RED Raven when we get our hands on it. But I bet you we will also have to get it ourselves and won’t be one of the first to have it from the manufacturer, just like the C300 mark II.

Conny Fridh Reply
Conny Fridh September 27, 2015

Heres is an other interesting test. How do you feel this hold up to the idea of your evaluation of the C300 II?

Still no comment from Canon about the dynamic range and soft slow motion picture?

Kind regards
C
https://vimeo.com/140441330

Kendal Miller Reply
Kendal Miller September 25, 2015

Not good.

Peter Hybæl Reply
Peter Hybæl September 25, 2015

Kudos to all folks on the forums who have seen this coming based on the highlights in the example footage that was available for Mk II. They have brilliantly called it.. no 15 usable stops anywhere.. not even in the Alexa territory. The only excuse could be that this is a prototype C300 Mk II camera an not the final one.

Also haha it really shows that Arri does have some special sauce or some super clever tricks or hacks (I don’t believe in the strange physical low-con filter theory).. my best bet is that Ursa 4.6k will be 1 stop shorter than Alexa.

Reply
Einar Davíðsson September 25, 2015

Cinema5D have stated here and elsewhere that this was a full production camera, not a prototype.

Jonathan Gentry Reply
Jonathan Gentry September 26, 2015

Interesting that the smear is exactly the width of the first blown out chip.

Reply
Joachim Hedén September 26, 2015

Hi there,
Before your test, did you verify the chart with a spot meter? Did the light source perhaps dim, is there dust/dirt that could affect the levels?
I havn’t done any controlled environment tests, but my feeling after shooting with a production model 300 mk2 for a week, is that I’m holding highlights easier than with a Sony F5 wich I’ve shot with a lot in the past few years.
Also, about the term “useful stops” – it’s a very subjective term. Here’s how I think about “useful stops”: The useful stops are what’s available above middle gray, not below. What’s down there in the bottom 2 or 3 stops you’re not really “using” for anything – but you are always protecting your highlights from clipping. The Sony F5 for example has nominally 6 stops of lattitude above middle gray, but in general the F5 needs to be overexposed by 1 stop to look good. So, that leaves you with 5 “useful” stops above middle gray. Like I said – I haven’t done any controlled environment tests, but my feeling is that the 300mk2 is doing rather well in the real world “above-middle-gray-stop” -aplication of the term “useful stops”.
My 2 cents

Theodore Lederman Reply
Theodore Lederman September 26, 2015

I’ll take 3 Ursa Minis and a shot of Jack for the same price, Alex.

Reply
Joseph Wilkins September 26, 2015

Still no response from canon???

Jeff Regan Reply
Jeff Regan September 26, 2015

These two videos show that the C309 Mark 2 has very good DR:

https://vimeo.com/140441330

https://vimeo.com/137221219

Conny Fridh Reply
Conny Fridh September 27, 2015

https://vimeo.com/140441330

What to you all think about this test? Does it leave us with any news concerning the dynamic range and soft slow motion picture?;)

C

Florian Klaes Reply
Florian Klaes September 27, 2015

Looks promising, i think all this is subjevtive and each online evaluation is diffrent it would really help to see the C300-1 against the C300-2 with original footage for download of both and a statement by Canon about this.

Reply
Adrian Bacon September 28, 2015

The Arri’s performed about what I’d expect to see on the Radiant test. Fairly good all around. The Red’s were about what I’d expect to see, also fairly good all around.

The Sony A7R2 had a ridiculous amount of clean shadow detail. Totally unreal. The downside though was by 1.5-2 stops over it was already completely blowing out highlights. This is in line with what I generally see with their stills cameras. They’re strongly biased to shadow detail at the expense of really ugly highlights. It’s just how Sony distributes their DR. Nothing wrong with that, though, something to be aware of and expose for accordingly.

The C300MkII was not as clean in the shadows as I was hoping, but with that being said, it looked to be mostly luminance noise and was not visually super ugly. In post it would clean up quite nicely and retain a fair amount of detail. I was actually hoping for more highlight range. It looked to be losing the highlights by 2.5-3 stops over. I really hoped it would have been more like 3-4 stops over. Nothing terribly wrong with that, just something to be aware of and expose for accordingly.

In all honesty, I thought the Sony FS7 didn’t look that great. The C300MkII looked way way way better at 5 stops under and had way more/better looking shadow detail there. Also, it started losing highlights by 2.5 stops over again. Again, nothing terribly wrong here, just something to be aware of and expose for accordingly.

The URSA was meh. The Panasonic was better than I expected. They’d do just fine if lit properly on a TV set, but that’s really about it.

My take away, all the cameras had their strengths and weaknesses, a few of them I’d probably only use in certain scenarios that played to the strengths, otherwise, they were all fairly competent if used by a skilled user that knew how to get the most from them.

Sebastian Wöber Reply
Sebastian Wöber September 27, 2015

I find it remarkable how our test results get washed out in some of the comments here and elsewhere, so much I feel it’s pointless to respond to some of them. It seems many people simply “want” the C300 mark II to be awesome and thus resort to accusing us of conducting our tests wrongly, simply not believing our “claims” or criticizing single irrelevant points in the article. This is coming mostly from people who it seems don’t know what testing a camera means.
We take great responsibility and care when conducting our tests and we published the original tested charts here so everyone can comprehend the results our software provided.
It is very clear, you can see with your own eyes that there is substantial noise in the dark areas and the light streak problem of the first C300 persists. There is no need to discuss or critizise our “testing methods”, just look at the charts and see.
The tests purpose is to compare C300 mkII to ALEXA and FS7 and the results are very very clear and easy to see. We did not tamper with these charts and would have no reason to. Look at what we do here, we take great care in providing the latest news, testing cameras and making our own original content. Thanks for keeping the discussions civilized.

Barry Goyette Reply
Barry Goyette September 27, 2015

Sebastian — There’s a saying here in the US that “if you can’t take the heat, get out of the kitchen”. There’s not a respected scientific study published anywhere in the world that doesn’t undergo “peer review”. While you consider the details of your article irrelevant (just like you previously described Dual Pixel Autofocus as irrelevant to the value of the C300 II because you don’t use it — funny…in the room full of hollywood DP’s, operators and producers who sat through a seminar yesterday on the c300, the biggest “ah-ha” moment of the day was the demo of this feature, followed by several outbursts of “kiss my 1st AC goodbye”).

The fact is the devil IS in the details, and your test really doesn’t pass muster. First it starts with some of the skewed comments you made on twitter, here and other places that indicates that you aren’t really looking at the data to learn, but to prove what you already know. Your inflammatory article title is just click bait, because we soon learn that you’re not measuring Dynamic range, but rather “usable” range (also heard at yesterday’s seminar “what the hell does that mean?…why does C5d get to decide what usable is?”… Your twitter outburst suggesting that this is somehow the same sensor as the original C300 is so thoroughly off base (take a look at the Radiant test and see if you can see ANY similarity between the C300 II and the C500.)

A better title for your article. C300 Mark II Dynamic Range: Exactly as I suspected.

“There is no need to discuss or criticize our testing methods” —seriously…you just wrote that? Who are you…the f**king Wizard of Oz?

Take a close look at radiants test. Designed by someone who knows how to design a camera test (and like all camera tests, it has it’s faults). You’ll see that the Amira bests the the C300II by about a half stop in the highlights, with the Canon showing about a half stop more range in the shadows, (exactly BTW what you’d expect given the companies stated DR distributions) albeit with a fair amount of noise. Looking closer we see that the C300 shadows are shifted upwards a little more than the Arri, but more importantly..looking into the deepest shadows we see that Arri is simply applying a fair amount of noise reduction down in those shadows, with the detail either smeared or simply gone.

Why do I bring this other test up. Because your test has too many problems to be considered valid. Usable DR is so subjective it’s not something that can really be tested for. But you insist it’s scientific. OK…then why did you rate the Amira at 13.1 stops last year and 14 this year? Why did you rate the original c300 at 10-11 stops (indicating in the text that it was at the lower end of the range), and then later rate it at 11.4 (just barely below canon’s stated 11.7). Why in the past did you describe the 14.5 stop A7s (Usable 12 stops) as nearly as good as the Amira, and describe the C300 as disappointing in comparison at 12.3. Why have we never seen the charts for any of these tests….(except the A7s as compared to the GH4)?

I think I know why we haven’t seen those charts before…because they aren’t that important…to you it’s what imatest spits out…based on the S/N ratio you consider ‘usable. The thing is…that on this test…the way the Arri’s blacks are rendered, makes it appear that there is substantially less noise and more DR. So lets haul out the charts and show ’em to everybody. So wait…is it the imatest number? or the visual that matters. See everybody I talk to says that noise in a DR chart only matters when you can no longer see the steps. When we normalize the blacks in the arri chart and line it up with the C300…we get nearly the EXACT SAME STEPS (the arri has a marginally visible step at 16), and the noise is barely different, (which we know from the radiant test is being reduced in a greater amount in Arri’s processing–even your test shows that the Arri gets ‘magically quieter (and blurrier) in stops 12-15). So if this test is based on measuring a certain signal to noise of a flat density patch and one manufacturer is applying a greater level of noise reduction to that patch, you can’t see how that that little detail is…um….relevant?

The take away from my discussions with a number of experienced camera testers is that noise is really a separate issue here, especially in the deepest shadows where it can easily be graded out or reduced as Arri (and um….everybody else) does with little effect on important detail.

Back to the Radiant test. It is absolutely true that the Alexa family of cameras bested everyone in the under/over/DR test. It should…it’s the most expensive of the bunch. I didn’t see another camera in the group that clearly bested the C300 II. The Varicam came closest. While the C300 has more noise in the deep shadows than many of the other cameras…it becomes clear that it’s maintaining more detail in those same shadows (the only camera to maintain the fine ribbing in the back wall around stops 2.5-3.5) This indicates that the competitors are simply masking out noise (and there’s still plenty there) where canon isn’t…yet.

For myself, I pulled together a composite of all the “portraits” in the over/under test, so I could see what 10 stops of “exposure” range looks like. I found that the C300 was easily gradable from about 4 stops over to 2 stops under. That’s a really good range, with plenty of headroom at either end to roll off highlights and shadows…all of the noise and “smear” dropped away in each of those grades with little coaxing.

The takeaway of this is your test is relying too much on noise at tone levels that don’t really matter, and that every sensor on the market either displays, or masks.

The only achilles heel that I see is that sensor smear…again…it’s something that generally gets graded out (along with the noise) so as you said earlier (before backtracking) it’s really not that big a deal….unless you think underexposing by 3 or more stops is a great technique. (interesting that you mention that the c300 had this problem. I’ve been shooting with that camera since the day it launched, and I’ve never ever seen it before)…if it’s the same then, probably I won’t see much of a problem with the new camera either.

A few days ago, you said you wanted to hear what I think of your test. Well now you have. Lets keep it civil, my friend :-)

Reply
John Trusk September 27, 2015

No need to be rude Barry…. we can disagree without being uncouth. They guys spent time doing these tests…. lets at least show them some appreciation for that.

Barry Goyette Reply
Barry Goyette September 27, 2015

Actually, John, I’m being pretty polite. Sebastian is pretty much calling everyone out who’s questioned his tests. That’s not polite. He’s chastised me more than once on this forum and on other sites for simply asking questions or engaging in polite, spirited conversation. The only thing impolite I said was a reference to a child’s book in which man controls a city of drones with a mantra of “don’t peak behind the curtain”. Read his previous post and see if it’s not apropo. Otherwise, the rest of my comment is just riffing on a lot of what has gone on here (and in a few other places) the past few days, AND to point out a much more valuable test on another site.

I have been incredibly polite and have received a lot of crap from him for it. I tend to have a really good sense of humor about this stuff, but Sebastian’s attitude towards a certain segment of his commenters (the ones that disagree with his conclusions) gets a little tiring. If you go back and read all the comments on this and the original “is the c300 still competitive” thread, I think you’ll see what I mean. If you go back and read, you’ll see that I have already thanked him for conducting the test, and noted that it was valuable.

Reply
Kaster Troy September 28, 2015

Barry, do us all a favor and just zip it. You obviously have no respect for the time and effort Sebastian has put into doing these tests, all you do is complain and say how inaccurate they are. Go do some tests of your own if your not happy with the ones provided here. In your mind this is already the best camera ever invented and you’ve never even used it. Its like you take it personally when someone has something negative to say about this camera. For the price this camera is a joke!! Accept it and move on.

Barry Goyette Reply
Barry Goyette September 28, 2015

Oh…Kaster…I was starting to like you.

For a guy who started the conversation with “Canon sucks bro, accept it and get over it, case closed lol”….. you’ve sure grown. Actually in my world, I look at a lot of sources positive and negative and use the lot of it to make my decision. It’s worked out for me. Apparently, despite all the shite that was spoken by geniuses about the original C300, it somehow managed to dominate a big section of the industry. I’ll wage you a gentleman’s bet the new one will do just as well.

I’m happy to head on down the road, as this place is an unpleasant place. The forums I belong to, people debate and discuss and then make plans to meet up for a beer. Around here, if you don’t believe in the Wizard, they toss you out on your can.

And seeing as you’re feeling so supportive of C5D right now…go take a long hard look at that FS7 again. Because there’s plenty of bad news in Sebastian’s test if you care to “take a look”.

And as far as I know, the best camera invented is the Alexa….can’t afford one. I’ll have to settle for the next best thing. (oh…and I have used it before).

Cheerio mate.

Reply
Kaster Troy September 28, 2015

Barry, you and I soulmates you know this! Apologize for coming across harsh, a lot of it is my dry sense of humor. I just hate seeing all this bickering over a DR test. I read the article and looked at the pictures and then moved on. I don’t know why people have to continually question Sebastian. He took a lot of time to do this and the test is what it is, simple as that. I think everyone should thank him and wait for the footage or test to come out. I don’t wanna sound like I’m a Canon hater, my first real cam was a 5DM2 then the 5DM3 which are awesome. I never had any interest in Sony until somehow I got my hands on an FS100, FS700 and a7s and they changed my life lol. Sony is killing it right now and the amount of camera you get for the money is just incredible. But there are definitely things on the FS5 and FS7 I absolutely can’t stand.

Barry Goyette Reply
Barry Goyette September 28, 2015

:-D

You know, I’ve published camera tests (it was a long time ago in much simpler time)…so I know where Sebastians at with this thing, and I know what it’s like to have everyone call you on every detail. (like that other test over at radiant, which was a huge undertaking…they screwed up a couple of cameras (of course the FS7 is one), and so that’s all people are talking about in the comments..you can’t win.

I really have less of a quibble with the C5d test than it’s conclusions (a hallmark of most ASC supervised tests is that the tester rarely expresses his opinion, and lets the image speak for itself) and Sebastian’s general reaction to commenters that disagree.

Nice singin’ on “Goodnight”.

Reply
Kaster Troy September 28, 2015

Thank you sir, that’s very kind of you to say ;)

Barry Goyette
Barry Goyette October 12, 2015

Hey Kaster…thought you might be interested in a test that Geoff Boyle has been cooking up. Comparing the range of lower priced cameras from Sony to Bolex to Canon and many stops in between. He had generally “fair to midland” things to say about all of the cameras except for a few. Here’s the part that’s relative to this forum….

He starts with this…..”I don’t like the budget cameras, it’s not that they can’t produce perfectly usable pictures, they can. They just don’t produce stellar pictures and they are a pain in post.”

and then this….

“Sony FS7 it’s a Sony so it has a lot of menus :-) the pictures are fine, if you like Sony cameras you’ll love this. It may help to make me like Sony more!

F55, not really part of tis series but we missed it last time, a good camera, simple if you want it to be and great pictures, a truly professional camera. Worth the money.

C300-2 I’ve kept this to last because it is a clear winner, miles ahead of anything else in the tests, easy to use, lock it in cine mode, easy to read the rushes, easy to grade, just drop the appropriate supplied LUT on and marvel :-) This is what a camera should be.
They should have called it the C400 justas the C500-2 should be the C600. It’s that big a change. In such a short test we didn’t have time to play with the looks function but if they work as we’ll as everything else…

It’s very simple really, if you want great images and a pain free life use Alexa or Amira, F65 or F55, Dragon, Varicam, C300-2.

If you fancy a challenge then maybe include the KineMax.

Anything else is a Fiat 500…”

Just one man’s opinion, I know…

Reply
sanveer mehlwal September 27, 2015

Why don’t YOU conduct a test, which is more Scientific and Precise and let us all see the results.

Barry Goyette Reply
Barry Goyette September 27, 2015

I would…but then I’d have to fight with myself. :-)

Reply
GELAX STUDIO September 28, 2015

websites like dxomark and dpreviw also does camera tests,what if c5d does their tests base on the same standards of dxo or dpreview ?

Reply
sanveer mehlwal September 28, 2015

You would have to eat gumble pie, obviously.

Also, it’s easier for critics to call a film rubbish, rather than ever really make one themselves.

“The whole problem with the world is that fools and fanatics are always so certain of themselves, and wiser people so full of doubts.”

I guess Bertrand Russel was thinking of you, when he said this.

Rastko Vukovic Reply
Rastko Vukovic September 28, 2015

Cinema5D.s test turned out to be spot on. https://vimeo.com/140441330 Now I knew that some people will be, like, hey you didn’t turn the log/gamma/hyper/camera up side down but the truth is this is THE test and canon IS lying and IS overpriced :) deal with it. Unless you work for canon. In that case go F yourself XD
peace
ps https://vimeo.com/140441330
pps https://vimeo.com/140441330
use your eyes

Sebastian Wöber Reply
Sebastian Wöber September 28, 2015

Hi Barry,

Thanks for your comment, I think there is some misconception here: There is no dynamic range test without taking noise into account.
And it is also not cinema5D’s method based on “usable” and the others do it differently. No, all dynamic range tests are based on a threshold and if this threshold is set with your eyes (subjective), then it is still a threshold.
Our scientific test is based on:
1. Usable dynamic range (S/N ratio) as claimed by Arri for their ALEXA at 14 stops. So this was our initial way we setup the software parameter.
2. A professionally acknowledged testing chart and testing software.
3. We always confirm it with our eyes and show you what we saw in the article, but this does not affect the numbers.

And that’s it. At one point in the past the software had a bug that was acknowledged by the software company. This is why for at least two cameras the numbers were changed (Haven’t re-tested all of them yet), because of the software bug.

And no, we don’t change numbers around, you are misinformed. The Arri AMIRA had a different DR than the ALEXA. So we’re not claiming 13 and then 14, they are merely different cameras, I suppose the different processing made the difference.
And for the C300, I think you misunderstood. I once mentioned that the Canons have 10-11 stops, which still included the C300 at 11.4.

I don’t think it does our tests justice if you write it off by saying “your testing methods” are flawed. You always have to decide on a testing parameter. We apply it to all cameras equally.
We admit we consider “usable dynamic range”. Which testing method did Canon use when they found 15 stops?
I think Arri were the only ones to get it right and reflect best what a DP actually has at his disposal in the field.

Barry Goyette Reply
Barry Goyette September 28, 2015

Hi Sebastian,

Thanks for the response to my comment.

I think what I meant relative to noise/DR is that by setting a very conservative number like you have, that results can often be skewed by increased noise reduction, and so while your test ends up being a better test of “image quality” than a less conservative noise number, it may be less accurate in determining actual dynamic range. As we see in the Radiant test, all of these cameras show a lot of noise in Log when pushed to their limits. Each one deals with that noise differently. Your test is totally valid, but does benefit cameras that reduce shadow noise (and detail) to a greater degree.

At the Canon seminar this weekend, they showed us a lot of clips that displayed how the C300 II was holding detail in the deep shadows on very high contrast scenes. This is something that your test doesn’t consider, and it’s an import an aspect of dynamic range that DP’s do consider.
There were a lot of experienced (old, like me) operators and DP’s in that room. Most seemed genuinely impressed, I also think that most would have considered that detail “usable”.

Again…thank you for your answer on this. I’m more than happy to agree to disagree.

Bgd Videography Reply
Bgd Videography September 27, 2015

Doesn’t look like LOG for the FS7 and A7S on the Radiant test?

Barry Goyette Reply
Barry Goyette September 27, 2015

Absolutely correct. That test, at least on the surface, appears flawed in that respect.

Sebastian Wöber Reply
Sebastian Wöber February 20, 2016

It was all tested in log (which were determined to offer the highest DR on all these cameras)

Bgd Videography Reply
Bgd Videography September 27, 2015

I’d just like to say thanks to the author for taking the time to carry out the tests. I think she should change the DR test to something we can all relate to. I look forward to her future tests on up and coming releases.

Bgd Videography Reply
Bgd Videography September 28, 2015

This article is getting a slating on other forums! Silly isn’t it. Will this article tarnish Canon’s sales figures? In the short run, maybe. Canon aren’t in any rush to discredit these findings. Maybe it is all just subjective, and this article amounts to an equally subjective discussion as to whether the Beatles are better than Elvis. Beatles rule by the way!

Barry Goyette Reply
Barry Goyette September 28, 2015

No man. Elvis is the king.

Bgd Videography Reply
Bgd Videography October 5, 2015

Lensprotogo are claiming the C300 Mark II is on par with the FS7 in terms of Dynamic range! Here: https://m.youtube.com/watch?v=C1uo7lQ87-A

Did Canon promote and push this camera on a spec that is subjective? It’s a little sly if they did. No comebacks, no proof, no response to this article.
Surely the price must be called into question, if the FS7 can compete in DR as well. The other features the C300 offers are great, but without the DR and slomo advantage, are they twice the price great? It’s all gone very quiet regarding this camera!

Reply
Jim Martin October 5, 2015

No, it has not gone quiet….a large amount of orders and people calling wanting to know when they will get their camera. Again,long time trusted sources (ASC level) who know cameras are solid on the 15 stops with no subjectivity…..

Barry Goyette Reply
Barry Goyette October 5, 2015

BGD –FWIW…canon has never responded to a blog post in the years I’ve been following cameras. I will say that a canon person told a room full of people “off the record” ( with a wink) that they were comfortable with the 15 stop characterization. The reality is this camera is going to take great images. From what I saw in the LP2Go review, (which, by the way, in the overall, he came down on the side of Canon, not the Fs7), the color on the C300 is superior, the low light is superior, the focus system is “groundbreaking” and superior, the usability is superior. The only things he gave to the Sony were price and frame-rate.

I got my first look at some mxf files off the C300 this weekend (lensRentals blog). I think they look stellar. The color is phenomenal, and the image is amazingly clean everywhere but the deepest shadows, (all cineon gamma camera show this, including the fs7).

Received another email from Geoff Boyle over the weekend in which he described the C300 II as “remarkably good”. Tall words from a guy who’s tested every camera that matters for more years than some bloggers have been alive.

Bgd Videography Reply
Bgd Videography October 5, 2015

Cheers Barry. I look forward to seeing Geoff Boyle’s tests.

Jeff Regan Reply
Jeff Regan October 5, 2015

Not fair to say Mark II is 2x the cost of FS7. If time code and genlock are useful, a $2K adapter needs to be added to FS7, plus V-mount batts, V-mount charger. The Sony proprietary media costs more, and if doing a lot of 4K, many cards are needed. The author of this article stated the Canon codec was superior. Most likely color science is as well. The package price difference is only a few thousand more for the Canon, but the FS7 is a great value, no question.

The Radiant test showed the Mark II to have the best shadow detail and highlight detail was competitive with all but Arri models. The Sony E-mount is not a rigid mount. The autofocus modes are class leading on theMark II

Reply
Kaster Troy October 5, 2015

It definitely isn’t fair but the fact is the Mark II costs double the price of an FS7. Sony isn’t the only brand that makes memory cards for the FS7, Lexar makes them for half the price. If a person needs timecode then your out an extra $2k but not everyone needs that. The side-by-side tests will be fun to watch once the camera is released.

Jeff Regan Reply
Jeff Regan October 5, 2015

Kastor, you’re right, non-Sony media is cheaper, but hopefully CFast cards will go down in price faster. If time code in or out is needed, it’s not just the $2000 for the adapter, but also a need for a 2nd type of battery system, which for V-lock batteries and chargers, adds up quickly. The white balance issue in CineEI mode is a problem, no LUTs can be app,I’d in playback, the build quality is not fantastic.

To me, the Mark II and the F5 are the real competitors, but obviously the FS7 has some very features vs. pricing. Just can’t get excited by that wobbly E-mount.

 Amir Kh Reply
Amir Kh October 16, 2015

E-mount has a great advantage. It can use optional speedbooster that gains one stop light and gives it full-frame look.

Bgd Videography Reply
Bgd Videography September 28, 2015

Seems Prokit.com have the Canon in stock, if you’re in the UK. When are Bhphoto getting them? It says late October on the website. I hope not!

Steve Oakley Reply
Steve Oakley October 2, 2015

I’m seeing 16-17 stops of DR. no idea where the claim of 12ish comes from… at all.

Reply
Crimson Son October 5, 2015

FYI – when determining dynamic range it is BOTH subjective and objective.

Objective is you use scopes like a waveform to show if each stair step is discernible electronically.

Subjective, compare image qualities of the extreme ends (usually shadows as digital still tend to hard clip highlights) and noise patterns.

Not sure why there is waveform in this testing.

Reply
sanveer mehlwal October 5, 2015

I can see 30 Stops of Dynamic Range. Not sure what the others are on.

Hahahahahahahahaha

Maarten Spoek Reply
Maarten Spoek October 6, 2015

So, if it’s a stop more sensitive, wouldn’t the DR be better if you measured using ISO 1600?

Sebastian Wöber Reply
Sebastian Wöber October 6, 2015

No

Reply
Bradley Stonesifer October 14, 2015

I’ve just done some side by side testing with the Arri amira and C300mkii. I’ll be sure to post the results in the coming days. In terms of image quality and dynamic range, one would be hard pressed to see the difference initially. My guestimate on my end is the Amira holds about 1/2 to 1stop more dynamic range in the highlights. The scene that i was metering was 7 1/2 stops over Key. Color sampling in 4k is also better in the amira due to the 4444 12bit vs the 422 10bit. One interesting note is that the CanonLog2 gamma curve has a few options in color space and even though there is more info in the cine gamut color space, the rec709 color space is the most similar to the Alexa/Amira Color science. I’d use the c300mkii as a gimbal drone camera any day. I’m currently testing the canon raw to the odyssey 7q+ and discovered that it is in fact capable of duel link out. You have to set the mon out setting to 4k raw priority to achieve this. Only time will tell if Canon and Convergent will be able to get us 4k60p, fingers crossed.

Reply
Bradley Stonesifer October 14, 2015

I’ve just done some side by side testing with the Arri amira and C300mkii. I’ll be sure to post the results in the coming days. In terms of image quality and dynamic range, one would be hard pressed to see the difference initially. My guestimate on my end is the Amira holds about 1/2 to 1stop more dynamic range in the highlights. The scene that i was metering was 7 1/2 stops over Key. Color sampling in 4k is also better in the amira due to the 4444 12bit vs the 422 10bit. One interesting note is that the CanonLog2 gamma curve has a few options in color space and even though there is more info in the cine gamut color space, the rec709 color space is the most similar to the Alexa/Amira Color science. I’d use the c300mkii as a gimbal drone camera any day. I’m currently testing the canon raw to the odyssey 7q+ and discovered that it is in fact capable of duel link out. You have to set the mon out setting to 4k raw priority to achieve this. Only time will tell if Canon and Convergent will be able to get us 4k60p, fingers crossed.

http://www.bradleystonesifer.com

Reply
Patrick Zadrobilek October 20, 2015

Hey Sebastian,

you should have tested clog instead of clog2. It’s far more noise free and also has a dynamic range of nearly 14 stops. Clog2 is a bit strange on the C300MkII on my tests, I agree on the noise at ISO 800, bit when going down to ISO 400 or ISO 250 the noise dissapears, but you loose about 1.5 stops of DR.

Sebastian Wöber Reply
Sebastian Wöber October 21, 2015

Hi Patrick,
The noise comes from the sensor, not from the log curve. Log2 gives you the best gamma encoding for organic grading and also as per Canon’s whitepapers it is the one to choose for best performance (dynamic range).

Reply
Patrick Zadrobilek October 21, 2015

Hi again ;-)

That could be theoretically true, but it looks more like there is going on some gain boost via signal processing only on clog2, or else every Gamma settings would have been that noisy. So the conclusion is, to achive that “marketing 15 stops” we boost the sensitivity for clog2, but all not for all the other Gamma settings, which are clean. Check out the normal clog on the Mark II and you will achieve also about 14 stops of dynamic range – the true dynamic range of the C300 MkII, which is not bad at all and actually very competitive compared to other cameras in that price range ;-)

Sebastian Wöber Reply
Sebastian Wöber October 21, 2015

No 14 stops on the C300 ii in our tests in Clog1 sorry. And the noise is the same, it’s just encoded differently, with a different gamma characteristic. The signal is the same from the same noisy sensor. If you use Clog2 and then lower the blacks you will have no noise either.

Barry Goyette Reply
Barry Goyette October 21, 2015

I think you’re wrong on this Sebastian. As you have stated previously in this thread, ALL manufacturers apply noise reduction as part of the processing of the image…so yes…all sensors are noisy…you have no way of knowing whether this sensor is noisier than others…you are only looking at results post processing. CLog2 appears to have less noise reduction than CLog, and BT.709 gamma has more noise reduction. All of these gammas have approximately the same number of stops, from numerous tests I’m calling it 14. Others have called it 15. You can call it what you want..but if you’ve tested CLog and found that it has the same noise characteristics as CLog2, I’d like to see that test.

Sebastian Wöber Reply
Sebastian Wöber October 21, 2015

Please show me that test. It is your claim so you should prove it, not me. Also I’m in line with Canon’s whitepaper.

Barry Goyette Reply
Barry Goyette October 21, 2015

Actually, Sebastian. It’s your claim that we’re talking about…two posts back…. “the noise is the same” (clog as compared to Clog2). As far as you saying you’ve got the white paper on your side. I’m not finding anything in there that claims the noise characteristics are better/the same between the various gammas. All discussions of noise appear to be relative to chip design and signal processing, not gamma. Please enlighten me.

For proof of my claim…which is not a claim…just an observation….I’ll offer this. A BT.709 gamma video that has been online for weeks now. Plenty of areas there where we should be seeing that awful noise you go on and on about. (underneath the tires and shadows within the interior of the car). Except…it’s just not there. There’s a download of the original file if you care. https://vimeo.com/140570879

The noise you point out in the shadows of the C300 CLog2 exists to varying degrees in every camera that produces a Cineon type gamma curve, including your beloved FS7 and frankly even the vaunted Alexa, but certainly in the Dragon, the F65, Varicam etc. Manufacturers have a choice of leaving that noise alone (and preserving detail) or smearing it knowing that most of that detail will get crushed anyway.

Reply
Jim Martin October 21, 2015

Barry, you are right on target with your posts…along with MR. Boyle (ASC) and others here in LA that have a lot more experience in testing, evaluating, and using cameras on a daily basis.

Sebastian Wöber Reply
Sebastian Wöber October 21, 2015

Barry, if you believe a different log curve will magically change the noise behaviour of the sensor you can do so, but please spare the readers. You remind me of a politician who twists stuff around and distorts, but I still don’t understand your aim other than that you really want the C300 to be better than my findings. But hey, life goes on.
Does Clog give you more dynamic range? Test results say no. Canon whitepaper says no.
That’s it.
I respectfully ask you to stop confusing unless you have prove for your ideas and observations in terms of a scientific test. Thank you

Reply
Kaster Troy October 21, 2015

I can’t believe we’re still talking about this, it’s absolutely ridiculous. If anyone has a problem with Sebastian’s tests than go do some of your own. If you can’t do a test of your own which nobody here will, than please thank Sebastian for his time not only for writing this article but also for him wasting his time replying to these idiotic comments. I’ve never seen a more annoying nit-picky comment thread in all of my life.

Barry Goyette Reply
Barry Goyette October 21, 2015

Who said CLog gives more DR? Not me. (I don’t think the white paper says anything about it…but…wait a minute…aren’t you the guy who doesn’t trust anything Canon says about this camera….so how come it’s bible now?.). FWIW, Canon has shown a chart (at the EVS demo a few weeks back) showing that Clog and Clog2 have identical DR characteristics, with a different distribution of it’s 15 stops.

” if you believe a different log curve will magically change the noise behaviour…”

Magic?…no…noise reduction. All cameras have it. You’ve said it yourself. All I’ve said is that Clog and BT.709 gammas on this camera don’t exhibit the noise you’ve discussed in CLog2. I’ve even shown you a sample. How am I twisting anything? You’re the one who claims it’s the same. Got any evidence of that, buster?

Finally, You don’t understand my aim? Well Sebastian, I really don’t have an agenda until I see someone with podium spouting shite. Believe me, every word you’ve spoken has made me research this camera and others at increasingly deeper levels. Spending $16k on a camera is not something I do lightly. And as I wait in the cue for mine, I’ve appreciated what you’ve said, but from the beginning, really couldn’t figure out your angle on this camera. I still really don’t know what you’re on a bout. All I know is this blog, by producing the thinnest of analyses (a single step wedge test) has generated the most noise.

Do I “want the C300 II to be better than your “findings”? What a stupid question. If I’m planning on buying one, I want it to be the best camera for the money. Funny, those are almost the same words Geoff Boyle used to describe this camera. You know him…they guy who’s website doesn’t take advertising. The guy who shoots feature films. The guy who’s been testing cameras since you were in diapers.

Frankly, though, I’ve really had it with your constant name calling. Whenever someone disagrees with you, you start saying they are spreading misinformation. It’s pathetic and it really shows your immaturity in this business.

Bgd Videography Reply
Bgd Videography October 21, 2015

Does anyone in the US have this in stock?

Reply
Jim Martin October 21, 2015

We dealers will get more cameras at the beginning of next week….

Bgd Videography Reply
Bgd Videography October 21, 2015

Thanks Jim.

Steve Oakley Reply
Steve Oakley October 21, 2015

So something no one has said : the C Log settings all have very high black levels. if you got out your old timey waveform monitor, or just looked at the camera’s you see the black levels floating at about 13 IRE. this floated black level means the codec will never crush blacks, but you can in camera. So if you REALLY want to BEST record what the camera is making, the stock LOG settings are basically rubbish.

I’ve been using either a custom setting using C Log as base, then adjusting the black levels down to 2-3 IRE. this expands the gradations recorded by the codec. this would also tend to reduce noise. While the Mk1 using 8bit codec, this is more critical, a 10bit codec is going to be more forgiving. Either way, the base C log settings are junk.

The other setting I use most often is simply a tweaked out Wide DR. Image reaquires the least amount of grading and looks good out of the camera. Its the same info as C log, it just looks better, take s a LOT less time to grade or mess around with. Even Wide DR can come out looking flat and Clog under the same shooting conditions would crush the image to basically 7bits.

The reality is shooting in sunlight, the DR of a 12 stop camera is NOT being pushed at all. in fact sunlight may actually be about a 8-10 stop DR in reality and the camera can capture easily everything from shadows to highlights.

Its far more shooting interiors with daylight that you run into more extremes of light that will start to push the DR of a camera…

the reality is, for what it would cost to shoot a dozen rolls of 35, you can own a camera that exceeds it. the pixel peeping here has gotten silly and it bears very little on what you can actually shoot in the real world. in fact the problem I find now is subtracting light from a scene with flags / solids / pieces of black rather than needing to always be adding tons of light.

On my last national spot shoot I used a silk to diffuse direct morning sun light, one 4X4 white board, and a 75W LED dimmed down as the tiniest amount of fill for the talents face. A far cry from what I would of used 10 years ago shooting on film. it would of been an entire circus of gear and crew.

So here is what I know: the Mk2 will be a better camera just for 4K over the Mk1, it’ll have more DR and if its 14 or 15 or 16 stops, it mostly doesn’t matter. my C300 paid for itself faster than any camera I’ve had before and its still making nice pix. Having the 4K option will be better. It will have fast ROI and be usable for 2-3 years which is more than you can say about a lot of cameras these days…

Reply
Jim Martin October 21, 2015

Well said Steve!…..a real working pro with real working knowledge.

Florian Klaes Reply
Florian Klaes October 21, 2015

Hey Steve, i feel the same concerning the diffrent profiles and black levels, but i didnt dig deep into the menu or touched the black levels – could go further into details? Do you have the time to create a video about it. The highly raised black levels seems to be overlooked by a lot of people.
Did you just compensate for the higher black with negative 13 or touched black gama aswell?

Steve Oakley Reply
Steve Oakley October 22, 2015

just adjust the master black level down while watching the waveform. I think it took around -5 or -7 to get the blacks down to about 2-3 IRE. Its a simple adjustment to make, and an important one on the Mk1 to get every last bit of gradation of out of the 8bit codec. I’d still do the same adjustment on the Mk2 as well. anything that improved the gradation recorded by the *codec* is a good thing.

FWIW at last NAB 2015 I was looking at Mk2 4K on 4K screen where they had dark blue / magenta wash on BG. perfectly smooth, no banding. 10bit helps a lot for shooting something like that, as does a tiny of of noise to break of the banding edges…even in 10bit.

The other thing no one is talking about is actual practical shooting. I’ve gone to ISO 4000 or 5000 for some shots. sure there was some noise, but nothing horrible. nothing anyone would turn their nose up at and say unacceptable. in fact quite the opposite, folks loved the low light shots very much. the noise is very fine, its not like the ugly globs you’d get on a old SD BVW400 or something similar… then of course folks slap film convert onto everything. go to the 3 perf 35mm setting and the grain added by FC (400-500ISO stock) is vastly more significant than what you get out of the Mk1 @ 4000-5000 ISO by a margin. So there is just some really silliness going on here…

Florian Klaes Reply
Florian Klaes October 22, 2015

Thanks! I find the noise feels more agressive and non-organic than on the C300-1 but is overlooked aswell it is a complete color free luma noise – i am waiting for a black friday sale on red giants denoiser. I should be easily to get rid of in higher ISO-sections.

Florian Klaes Reply
Florian Klaes October 22, 2015

Another thing i found in the manuel that isnt so common known is: Clog/2, base ISO: 800 – Wide DR, base ISO: 400.
Did you know that?

Steve Oakley Reply
Steve Oakley October 22, 2015

no. haven’t had a chance to look at the manual… speaking of which, where are people even getting Mk2’s ? no on seems to have them in stock right now…only pre-order

Reply
Joachim Hedén October 22, 2015

Hello all.
I did a test that maybe should shed some light…
You can draw your own conclusions.

https://joachimhedenworkblog.wordpress.com/2015/10/22/some-dynamic-range-tests/

Kind regards,

Joachim

Barry Goyette Reply
Barry Goyette October 22, 2015

Nice Test Joachim. Especially this part:

“You can draw your own conclusions.”

Reply
James Manson January 30, 2016

Thanks for a test done the way it should be.

 Chris Westerstrom Reply
Chris Westerstrom October 31, 2015

Anyone know how to get C300m2 2k12bit 4444 footage into an editing program like Final cut x?

Bgd Videography Reply
Bgd Videography October 31, 2015

I don’t think 12 Bit 4444 is supported in FCPX yet.

 Chris Westerstrom Reply
Chris Westerstrom October 31, 2015

thanks for the answer, I guess it’s Premier then?

Bgd Videography Reply
Bgd Videography October 31, 2015

I have no idea about Premier sorry.

Barry Goyette Reply
Barry Goyette October 31, 2015

I saw a video last week that said premiere is supporting all flavors of xfavc in the latest update. And canon is giving away a free year of premiere CC with purchase of the camera.

Reply
Patrick Zadrobilek November 1, 2015

PremiereCC is supporting MXF 12bit 4:4:4 (there is no additional :4 which would be an alpha channel).

Reply
Emrah Cahit OZER November 26, 2015

What do I get from the charts? Alexa 16 stops of DR?

Reply
Dan Evans November 27, 2015

For anyone interested, here’s some initial tests I did with Canon Log 2 and slow motion (2k).
https://vimeo.com/147038575

Reply
James Manson February 1, 2016

Thanks. Really nice.

Sebastian Wöber Reply
Sebastian Wöber November 30, 2015

Hello everyone,
Canon responded to this article. You might be interested in our follow up article:
https://www.cinema5d.com/canon-measured-15-stops-dynamic-range-c300-mark-ii/

 Olivier Sarbil Reply
Olivier Sarbil January 31, 2016

I have been filming with the Canon C300 for the past few years. I have just picked up a brand new C300 MkII and I am getting a terrible high level of noise… at 800 ISO, the picture is just unusable. CPS Europe AGREED that my footage is getting far too much noise and actually their technician was pretty embarrassed… but in the end, he just blamed it on using Wide DR Gamma. They asked me to do more tests which I have done… and I am still getting the most awful noise. I did more tests shooting C-log, BT709 Preset, with CP setting… the footage is just bad… Let’s speak frankly, the footage is just terribly crap. So far it’s 18k in the bin and still no help from Canon… Not to mention that the picture is also soft… If anyone had similar issues?

Reply
James Manson February 1, 2016

Soft? C300MKII? Lens, operator or defective camera. The C300MKII produces amazingly detailed sharp images.

 Olivier Sarbil Reply
Olivier Sarbil February 1, 2016

Definitely not the operator, thanks ;)… the lenses are fine and work perfectly ont the Mark1. If you want to check out what I get form the Mark2 at the moment, please check out the link bellow.

the lens is a 70-200, ISO1000, Gamma Wide DR

I just upload a slow-mo testclip… check out the noise and the “sharpness”…

https://vimeo.com/153733071

PASSWORD: mark2test

CPS technician in France find it sharp, my camera dealer find the softness cinematic. ahahah

Steve Oakley Reply
Steve Oakley February 2, 2016

Hmmm that does look noisy for 1000. however, it also looks *really* under exposed too with not a lot of dynamic range. stretching that back out to a normal range could easily pull the noise out a lot. for such flat light I’d be shooting gamma 709 and 2020 or cine gamut color space. as example a 4K test here https://www.youtube.com/watch?v=_LefhSERi1I which is quite clean and sharp. I’d suggest a brighter exposure. the Mk2 likes more rather than less exposure

 Olivier Sarbil Reply
Olivier Sarbil February 2, 2016

Hi Steve,

The shot needs a brighter exposure but it was shot to highlight some of the camera problems. At 1000 ISO, even under exposed and using the Wide DR gamma… I should not get such a high level of noise, that’s just a joke. Side by side with another Mark2, it takes 5mn to realise that camera is definitely wrong… You just can’t even get a sharp image! Yesterday Canon took back the camera and they finally admitted that the camera is defective (good for destruction). No matter the setting used, the quality of the picture was always terribly poor. Canon will exchange the camera and that’s the good news! But I can’t really hide my frustration… 18k camera and defective out of the box!

 Andy Krucsai Reply
Andy Krucsai February 19, 2016

Fyi:

“Canon has announced that the EOS C300 Mark II digital cinema camera has been independently tested in accordance with European Broadcast Union (EBU) recommendations for both HD and UHD content acquisition…1

The report, by broadcast industry expert Alan Roberts, recognises image capture of up to 15 stops of dynamic range using Canon Log2 and qualifies the EOS C300 Mark II for R.118 Tier 1 in HD and Tier 2 in UHD2.”

Bgd Videography Reply
Bgd Videography February 19, 2016

Who do you believe, Alan Roberts or Sebastian Wober?

Reply
Jim Martin February 19, 2016

That’s easy, Alan…..impeccable reputation….

Barry Goyette Reply
Barry Goyette February 19, 2016

I’ll believe the person without the website that gets paid by the click.

But really, can’t they both be right? Alan Roberts is certainly an esteemed reviewer with more experience in his left pinky than Sebastian, but Sebastian’s review was really faulty in it’s “angle” and possibly his lack of understanding at how Arri cooks the books (sensors that show more noise in the highlights generally don’t show less noise in the shadows unless there is some serious work being done — noise reduction and loss of detail, followed by the adding in of some “pretty” noise). His tests didn’t compare detail in the shadows, only noise, and so we’re left with an incomplete conclusion that unfortunately, Sebastian really wanted to insist was very complete. (that’s that angle part). Otherwise, his results are generally in line with his previous testing, and that’s all fine. For him, most of these cameras are over-rated, (except the holy grail Arri of course), and as long as he’s consistent, it’s all good. His site got more play and PR over this than it’s probably ever had, the only downside being Sebastian having to deal with knuckleheads like me :-) and possibly some cold shoulders over at Canon at the next few trade shows.

Jeff Regan Reply
Jeff Regan February 19, 2016

I really wanted to like this camera, as an owner of two C300 PL’s, but I see too many compromises out of the gate for the price point. I was disappointed that ProRes wasn’t an internal record option, the form factor is a pain, lots of little bugs in operation, color space doesn’t match Mark I(looks more like Sony magenta), didn’t see signal to noise improvement, more dynamic range–but only in shadow detail. Changing lens mounts still seems to be expensive and difficult, high frame rates not competitive with Sony, crop is not ideal quality or operationally.

No question the resolution even in HD is superior with the Mark II, as well as the codec. More ND options is nice. AF options seem to be a good feature for some. 4K, of course. Just feels like Canon is playing catch to Sony at a premium price.

At this point, I’m more excited about the Varicam LT, except for the potential difficulty marketing anything with a Panasonic name plate on it.

Barry Goyette Reply
Barry Goyette February 20, 2016

Hey Jeff, as an original C300 owner, I can really say that this camera is a marked improvement in so many ways. Up to 12,800 this camera is distinctly cleaner than the original camera, and beyond that it’s about the same. Canon’s codec is really superior as a shooting codec to pro-res, and the dynamic range extension goes both ways..the highlights on this camera look nothing like the oversaturated highlights of the C300. The AF features will truly change the way many of us work. I don’t know what bugs you’re talking about (canon not making magnify work in during doesn’t qualify as a bug, more like a difference of opinion). The color on this camera is superior in every way to the original. The magenta you’ve seen (I’ve seen in too) is more of a white balance and rec 709 LUT issue. Once you set up the camera properly it simply isn’t there. And I’m getting a completely different response to my Arri L7Cs and Area 48 LEDs with this camera, which is a great thing, because the C300 always struggled with them. Lastly, the Cfast combined with the new codecs and better LCD and viewfinder (in particular the EVF is amazingly good) have made this camera much more self contained. No need for a recorder/monitor in most situations, although I will probably add a SmHD 702 sometime soon.

The Varicam LT looks like a great camera, pretty much designed to compete in the same market as the C300 II. The lowlight footage I saw at NAB last year was quite impressive. The best move they made was putting an EF lens mount on it, that should help it on the marketing side.

Reply
James Manson February 19, 2016

I still shake my head at the damage the reputation of the C300MKII suffered due to Sebastian Wober’s flawed test. I witnessed orders being cancelled because of an online article that has now been shown to be incomplete. Bad news is good news for websites. Not much love for Canon in these here parts and the article certainly had an ‘Angle’

Sebastian Wöber Reply
Sebastian Wöber February 20, 2016

Excuse me James, but our test was neither flawed nor incomplete in regards to dynamic range. I stand by our findings, we even re-tested and compared them to Canon’s own testing methods which they had disclosed much later. Also, would you please just look at the Alexa in comparison to the C300 II. It couldn’t be more obvious what many don’t want to see. For us bad news is not (!!!) good news, we risk a lot with this and we will continue to do so. We have no reason to be biased. Other times people accuse us of being Canon fanboys. I hope it’s obviously ridiculous to most readers. Thank you

Reply
James Manson February 21, 2016

I have looked at the Alexa in comparison to the C300MKII and the final image from the Canon C300MKII matches the Alexa incredibly well. It is a camera that can actually compete with the Alexa which is twice the price but all we hear is that the C300MKII is too expensive. Testing or no testing, the C300MKII is a monster of a camera that has been unfairly tarnished by your article that basically went viral. It was syndicated across practically every camera website there is. I’m not sure why you are unable to see the damage that was done to the reputation of the C300MKII by this.

Seriously, what is that you are actually risking?

Barry Goyette Reply
Barry Goyette February 21, 2016

I wouldn’t overestimate the effect of this article on C300 II reputation or sales. The reality is the price puts it out of range of the bulk of those who might have been influenced by the article (it was obviously written with FS7 fans in mind). There were and are certainly many other respected voices speaking out with a completely opposite opinion from Cinema5d. The reputation of the camera will come from talented people putting it to work on projects they care about, not from some overhyped step wedges.

Sebastian Wöber Reply
Sebastian Wöber February 21, 2016

James, I invite you to place a C300 mark II and Alexa next to each other and do some test shots to observe the difference in dynamic range, even better, please work with both cameras and then come back here. This is what we aim to do at cinema5D, besides the tests. I think anybody who says the C300II can compete with the Alexa has, I’m sorry, no idea. Also did you note that they are two very different cameras for different applications? The C300II excels in other areas like weight or price.
And who told you the C300II is a “monster of a camera”? Makes you see that advertising works… and this is exactly why we published this test.
In terms of dynamic range the C300II is closer to the FS7 and VariCam 35 and this was well presented here. If you think you can simply look on these images on your iPhone and judge them, please think again. We use an industry standard DR testing software for objective results and well tested, consistent and thought through testing environment. Also I would like to add that most informed DP’s would not choose the C300II over the Alexa for a serious shoot where DR is critical to him. I am certain this is why some users cancelled their orders, because the C300II clearly doesn’t compete with the Alexa, though otherwise claimed in Canon’s advertising. Note that the tests by some other reviewers don’t compare to the Alexa, but are based on a different benchmark and Alan tested for broadcast standards. Video and television have other prerequisites than professional film applications. People have shot on cameras with 8 stops of dynamic range in broadcasting. Of course the C300II is a great camera in this regard and also in many other regards. But if you measure like that, then the Alexa will score probably around 18 stops.
I find it sad that people defend a company on an aspect where the company advertised something in a misleading way. Are you a Canon employee or are you just in love with the camera’s specs list?
I try to be critical and realistic. This is the only way I can make a shooting decision as a DP and is the reason we do this here. We get a lot of criticism and sometimes hate from commenters. That’s the tradeoff, but we stand by what we do.

Reply
James Manson February 21, 2016

I came to the conclusion that the C300 MKII is a monster camera by using it along with many other cameras filming many episodes for two television shows both lifestyle based in many, varied conditions, some brutally harsh. Auto focus that you completely discount makes the camera a monster for me alone. I am basing all my comments on my own experience using the camera since early November on an almost daily basis not because I work for Canon which I do not.

Sebastian Wöber Reply
Sebastian Wöber February 20, 2016

@ Barry: Sorry Barry, but my review was not faulty. I think you’re quite misinformed about measuring dynamic range with your “noise in highlights” idea. I want to see a camera that has noise in the highlights that makes the data unusable. I conduct my tests so they make sense for a shooter, because I am a shooter and I’ve studied this and been shooting for a while. I believe the issue is that some people try to science dynamic range out of a camera where there is none. I guarantee you, you can make the C300 mark II have 18 stops of dynamic range if you want it. I try to give people perspective where manufacturers and some biased reviewers have none. We don’t care how much DR a camera has, we just show it. And no, the “clicks” on this article are pretty much irrelevant for the income or popularity of our site. If anything it puts us on the line and you can be sure, nobody ordered a C300 II through our sponsored links…

Barry Goyette Reply
Barry Goyette February 20, 2016

Sebastian,

As usual, you have a hard time seeing the forest for the trees. I was generally supporting your review, saying that it’s ok in Cinema 5d’s methodology for the camera to have tested the way it did.

However, your angle, and your placing of the Alexa on a pedestal doesn’t assist your review. Testing DR with an singular test like this DOESN’T tell the whole story. A step wedge can’t tell you how much detail is being retained or lost due to noise reduction. Arri’s sensor shows too much noise at 50% of peak white to be considered Tier 1 by EBU, yet it’s noise profile below 10% seams to go in the opposite direction of what it should. No other camera exhibits this. (EBU is considering rewriting its specification because of how the Alexa performs).

From Newsshooter– “At closer to peak white the Canon is actually quite a bit less noisy than the Amira (although film is actually noisy in the whites which is where the negative is at its densest). At the lower end of the curve, at below 10% peak white, things get noisier with the C300 MkII. The ARRI does something clever here which makes these lower levels less noisy.”

The ARRI produces a beautiful image, but it isn’t the engineering benchmark that you want it to be. For instance, it is only a 3k sensor and it’s measured resolution maxes out at 20% fewer lines than the C300. Generally higher resolution sensors exhibit higher noise, yet the C300 exhibits the same or less noise than the Arri everywhere except in the area below 10% where Arri is cooking the books.

Your article, it’s headline, your twitter posts, your methods for detailing your results (pushing the step wedge image in post so you can “see” whats going on in the shadows), …all of it was slanted. Inflammatory. This camera easily has 2-3 stops more DR than the original C300, yet your review indicates about a half stop. Your twitter posts suggested that it was actually the same sensor as the original.

I’ve been shooting with the C300 II now for months. I’ve never seen noise on this camera that was objectionable at the ISO’s we’re discussing. Yet somehow your article would make one believe that the noise on this camera was extraordinary. Actually, the noise levels and DR in Clog 2 are on par or better than what we see from the F65, F55, Red Epic, Varicam, et al. All cameras that are significantly more expensive. You article, in its weird way of comparing it to both an $70k camera and the $7k FS7, argued that it was simply too expensive, ignoring the fact that it performs similarly to many cameras at 2-3 times it’s price.

Sebastian…Buddy…sorry. I have to disagree that you simply don’t care how much DR a camera has. If you didn’t you would have simply posted the step wedges, and shut the hell up. Instead you editorialized the process in a number of ways. Which is the only reason you’re getting beat up over it.

Sebastian Wöber Reply
Sebastian Wöber February 21, 2016

I see where this is going. I’m sorry what we found obviously offended you and your new camera. Please try shooting on a different one and get a perspective. I have explained myself and I have no interest to discuss this with you further, as after what you wrote I think you have just read a lot of stuff on the internet and have little knowledge about proper camera tests. Bye

Barry Goyette Reply
Barry Goyette February 21, 2016

—I think you have just read a lot of stuff on the internet and have little knowledge about proper camera tests. Bye—

….writes the man who published his “not-faulty, complete and unquestionable” test of the C300 mark II within hours of it becoming available.

….writes the reviewer who’s only “real world” mention the camera at the time of the review includes shock that a camera with a cineon type log curve would exhibit noise when underexposed (uh…maybe you should sit down with Sony Guru Alistair Chapman and have him explain that one to you).

…..writes the “creator of proper camera tests” who has tested so few cameras that he didn’t recognize CMOS smear until several commenters explained it to him.

Yes…Sebastian….I do read a lot of stuff on the internet. It’s a cool place to learn things. I also have worked professionally as a photographer/DP every day for the past 26 years. I’ve owned and used more cameras than you’ve had birthdays. I’ve never met a perfect camera and the C300 mark II is certainly no different in that respect. So “offending me and my camera” is really something you’d have a hard time doing….except when you assume you know anything about me or my experience…which by the way…seems to be a nasty tendency of yours when engaging people who disagree with you. (You should get that checked.)

Steve Oakley Reply
Steve Oakley February 21, 2016

Please see image on this link for mk2 noise – clear horizontal and vertical pattern visible. graded image http://stevenoakley.com/images/c300mk2_noise.png

Sebastian Wöber Reply
Sebastian Wöber February 22, 2016

Hi Steve, Thanks for sharing. While I’d say the C300II has more noise than many other cameras I’ve tested, what you linked looks like too much.
Please give us some info: What ISO did you shoot this on and in what way did you manipulate that shot? Was NR enabled? How far was it pushed? Thanks

Barry Goyette Reply
Barry Goyette February 22, 2016

I’ll see your noise, and raise you.

4k C300II. Canon Log with Canon LUT applied in FCPX. No noise reduction. Graded and sharpened (which would have emphasized any noise if it were there.) Screenshot is not full resolution, but there is no visible noise in the 4k file.

https://www.flickr.com/photos/barrygoyette/24813260609/in/dateposted-public/

As Sebastian said, what you’re showing is not typical of this camera, unless you’ve underexposed the image or are shooting at 25,600 ISO or above.

Steve — As has been said many times here and other smart places, log gammas are NOT designed to be pushed in post…EVER. Log gamma’s (especially CLog2) are designed to be pulled. So the question “how far was it pushed?” isn’t really relevant. If you pushed at all, then it was underexposed. Are you using your waveforms when you shoot? CLog2 at base ISO has a clear clipping point at 92 IRE. Bring your exposure up so your highlights kiss that line and then back it off a smidge.

Steve Oakley Reply
Steve Oakley February 22, 2016

@ Barry, this wasn’t “pushed” by any means. why are you assuming that ? Shot you see shows blacks pulled down a bit from original raw shot. Waveform is peaks @ 80, most of the image at 40-20 with some of it down it 0. Another words a decent exposure.

settings where : Clog2, Rec 2020, 4K, Matrix neutral, sharpness +2, coring +5, NR 2,Color Matrix Gain 10+. G-R -2, G-B -2. As for ISO, best guess is 1000. Shot you see shows blacks pulled down a bit. Waveform is peaks @ 80, most of the image at 40-20 with some of it down it 0. Another words a decent exposure. The noise isn’t in the stop or so right above black, its 2-3 stops above black. Had into canon for them to look at and they told me to turn on the NR to 2 – which I had on this shot.

what it comes down to is this : cLog2 should ONLY be used in the most extreme of circumstances like interior shots against windows where you are trying to hold onto the exterior detail. other than that, cLog 1 or one of the cine gammas will be a better choice to keep the camera out of the noise floor.

Reply
Patrick Zadrobilek February 22, 2016

With Canon Log 2 I would never ever set any other sharpness than -10, because it introduces all kinds of artifacts like double edges, raises noise dramatically and in camera noise reduction reduces resolution very fast. I will have my review ready soon were I test a lot of things on gadgetflux.at

CLog2 is beautifull when exposed correctly without any major noise in it when graded correctly, at least not much more noise than WideDR. The older clog btw, is much more noise free if you want to go for a perfectly “Alexa” clean image.

Steve Oakley Reply
Steve Oakley February 19, 2016

yAs an update since I had my Mk2 @ canon. They officially recommend setting noise reduction to 2 in cLog2. Looking at roberts test on noise reduction, you could easily go to 3 w/o any impact of significance on the image. If this fixes it, canon should consider updating the preset in the next firmware update. If Arri is basically doing the same – doing noise reduction in the lower parts of the image, so be it. FWIW the C300 Mk2 has selective noise reduction feature. you could enable that and set it up to only kick in below say, 20 IRE and solve the problem.

Barry Goyette Reply
Barry Goyette February 22, 2016

Sorry for the assumption, but that’s just what it looks like without the information you’ve now provided. If the numbers you’ve quoted are correct, there is something wrong with your camera..

Steve Oakley Reply
Steve Oakley February 22, 2016

thats what I thought. had canon look at it and they say it matches their reference camera in the shop.. in which case I say both cameras have a problem :(

@ patrik Z ; as for sharpness being at +2, thats nothing on a scale that goes to 50, and coring at +5 takes any edges issues away. however I have NEVER seen that problem with either mk1 or 2. a small adjustment like that will not drive the noise through the roof either. sharpening is much maligned I find, but if you expect to actually get resolution from the camera factory defaults on most cameras are well placed. I can rarely say I’ve ever used it in a negative value because the image just gets too soft. there is some math behind it that every camera needs to so some sharpening to hit its specs.

I also followed canon’s recommendation of NR @ 2. realistically the sharpening + NR cancel each other out. Also FWIW I have seen some recent charts shot with NR at 0,2 and 12. While at 12 it does flatten out the finest detail, I’d hardly call it as dramatic as you are making it out to be – and thats on charts. The in camera NR is good and if it cLog requires moving it to 3 or 4, so be it if thats what it takes to get rid of the pattern noise.

probably what cLog wants is the selective NR engaged… and Arri apparently doing exactly this to keep the blacks / shadows looking clean.

Reply
Patrick Zadrobilek February 22, 2016

Steve,

are you really talking about the C300 Mark II?
Because sharpness range is from -10 to +10, not 50.
I’ve done sharpness tests and as soon as you hit sharpness -7 you see increasing “edge hardening” and it gets worst on every step. Best artifact free sharpness is from -10 to -8.
I will write about this in my Review coming out today, so keep an eye at http://gadgetflux.net

Steve Oakley Reply
Steve Oakley March 2, 2016

I just looked. sharpness range is indeed 0-50, NR off,1-12, coring 0-50. so the minor adjustment to sharpening here is small.

Reply
Patrick Zadrobilek March 2, 2016

To clarify: in which menu point do you see your sharpen settings?
I’m talking about image sharpness in picture profiles at Menu->CP->OtherSettings->Sharpness->Level.

Steve Oakley Reply
Steve Oakley March 2, 2016

Yup same place. Ran it all the way up to check

 Andy Krucsai Reply
Andy Krucsai March 11, 2016

Cinematographer Geoff Boyle confirms, that is was a very good decision for us to buy the Canon C300 Mark II. to be our 4 K cinema camera.
https://vimeo.com/158017639

Reply
Jürgen Gutowski November 5, 2016

Danke Sebastian!

Dein Test hat mich vor einem ärgerlichen Fehlkauf bewahrt. Ich habe mich eh schon gewundert, das die Kamera bereits um 1/3 im Preis gefallen ist.