LAB Review – Sony A5100 [UPDATED!]

sony_a51003 days ago Sony announced the new Sony Alpha 5100 mirrorless camera that has made some headlines on the film blogs as it brings the powerful new XAVC S codec as well as 1080p at 60 frames per second.

At cinema5D we already had this new camera in our test labs today and we have some interesting things to share. We tested dynamic range, rolling shutter and observed sharpness and aliasing.

c5d_lab_logopsd2What is the cinema5D test lab?
At cinema5D’s new test lab we accurately measure and evaluate the performance of cameras in a controlled testing environment. We are using precise imaging tools, techniques and software to measure each camera’s performance.

At $550 the Sony A5100 is indeed a “video camera” that looks quite interesting. After Sony’s success with the Sony A7S everybody’s wondering if the same intriguing video functionality can be found in the much smaller sized mirrorless successor of the Alpha 5000. Well let’s take a look.

Dynamic Range

Testing the dynamic range on this camera wasn’t easy. We didn’t have an official native ISO and looked at all ISO combinations ranging from ISO100 up to ISO1600 and measured each with different “Creative Styles” that would provide the best rendering.

Clearly the strongest ISO values are ISO200 and ISO800, while ISO800 provides slightly more dynamic range reaching 13 [UPDATED!] 10.5 measured stops with Creative style “Portrait (-3, 0, -3)”.

Here is a chart that shows you dynamic range in comparison to other important cameras:

[UPDATE Aug. 22 ’14:] These test results are the exact numbers the software IMATEST provided in our test at a signal to noise ratio of 1/0.5 in the camera’s respective resolution and compression. Many factors influence these numbers and each sensor has its own characteristics. At this point we want to emphasise that these numbers differ from the subjective opinion we have about the camera, which for us at cinema5D is a very big point as we want to give you an ideal understanding of what the cameras can actually do for you. So we decided to mention it here. Subjectively, in comparison to the other cameras the maximum rating we would give to the Sony A5100 is 12 stops.

[UPDATE 2 Aug. 25 ’14:] The software manufacturer informed us that their software had a bug that misinterpreted the Sony A5100’s blacklevels (that seem to inherit green noise) and strongly affected the results for the Sony A5100 dynamic range test. We apologise for any inconvenience this might have caused. At cinema5D we will no longer rely on software results should they ever again differ from subjective evaluation. The software we use has been updated and the bug has been resolved.

sonya5100_zeissInterpretation
Here we tested usable dynamic range of the given cameras. [UPDATED:] The 10.5 stops of usable dynamic range the Sony A5100 provides is good considering it’s just a $550 mirrorless camera.
As you can see it performs slightly worse than the GH4.

We will go into more detail concerning overall handling and usage in the field in our upcoming video review. But one should point out at this point that the camera feels very much like a consumer camera (which it is). In that sense you will not find the log-curve picture profiles like on the A7s or other prosumer models. Instead the A5100 offers “creative styles” where as mentioned Portrait performed best. We also tested “Picture Effects” like “Soft High-key” which may seem to improve dynamic range, but in our tests we found it doesn’t.

On all our tests we use the same ultrasharp Zeiss 50mm T/2.1 Macro lens and the DSC Labs XYLA 21 step dynamic range chart. If you want to know how we test, we explain that in this article: Sony A7S dynamic range test.

Rolling Shutter

On the rolling shutter chart we can see how the Sony 5100 is also a winner. The Sony A7s while providing amazing images has a really bad rolling shutter performance, meaning the images tend to wobble on fast motion. The Sony A5100 however reads out the image even faster than the very well performing Panasonic GH4.

Test-Scores_RS_SonyA5100

It is interesting to see that rolling shutter performance is identical in 25p (PAL model) mode as well as in 50p. On other cameras rolling shutter is usually half as severe in slow motion. This tells us about how the sensor reads out the image in slow motion mode.

Sharpness / Detail / Aliasing

SonyA5100_SUB_iso800_portraitWe just talked about how the sensor reads out the picture similarly in slow motion as in normal mode.
This can also be observed when we take a subjective look at the image which looks identical in both modes as opposed to how the Sony A7s behaves. (Meaning 5100 is better in this regard)

SonyA7s_SUB_iso3200_ffsony_a5100_crop1On the left you can find full frame grabs from the Sony A5100 and the Sony A7s for comparison.
Each frame is an i-Frame extracted from the original files. They were recorded under identical lighting conditions. The A7s was set to ISO 3200 (F/11) which produced a very dark image, as opposed to the A5100 shot at ISO 800 (F5.6). In both cases measured with a light meter.

On the left we’re displaying some 1×1 crops of the original images in which the Sony A7s image was brightened up with a curve where highlights were retained.

sony_a5100_crop2You can see that in terms of sharpness and detail the Sony A7s outperforms the A5100.

In the image with the sector stars (kindly provided by Danes Picta) it becomes apparent that there is some aliasing going on in the A5100 and that it can’t hold up to the crispy clean image of the A7s.

Here’s another crop of the dark cat that shows us the soft rendering of shadow areas on the A7s where I must say the A5100 actually doesn’t perform so bad. We’ve seen much worse and this is most probably due to the good dynamic range and noise performance on all of Sony’s latest cameras including the A5100.

sony_a5100_crop4The overall pink tint of the A5100 is a weird shift in the Portrait Creative Style, but the color you can see in the dark fur, that seems to be moiré.

Where moiré can more clearly be seen on the A5100 is in the tie. It must be noted that most other DSLR kind of cameras have had worse moiré and aliasing performance. The 5100 will look good next to a T3i or most Nikon cameras for example.

sony_a5100_crop5Highlight rendering on the A5100 is good. But the A7s really shines with its Slog2 profile that has a beautiful, organic highlight rolloff.

You can clearly see how nicely the Sony A7s renders all the highlight details while they seem less organic and slightly overblown on the A5100.

You might think the Sony A5100 was exposed much brighter, but then again, they were exposed with the same exposure values and the Sony A7s’ Slog2 simply has a different way of storing the data inside the luma range which isn’t an option on the consumer A5100 (yet).

Conclusion [UPDATED]

The Sony A5100 clearly comes in behind the Sony A7s in terms of image quality, with a softer, less clean image. However for $550 (what do you expect) it is still a good performer among DSLR cameras that currently shoot video on an APS-C sized sensor. On top the slow motion mode gives you the same image quality as normal shooting mode.

In terms of dynamic range the Sony A5100 surprises with the measured results and beats most other important cameras we tested, even more expensive ones. [UPDATE:] As mentioned above the subjective impression we have about dynamic range on the A5100 is lower than the software results and can’t compete with the Sony A7s.

It also shines in rolling shutter performance. Overall it seems Sony has built its latest processing technology into this little mirrorless camera, providing good lowlight noise levels as well, which however we see nowhere close the A7s (not tested scientifically yet).

Add to that the internal XAVC S codec as well as the uncompressed 8-bit 4:2:2 output via hdmi in up to 60p and you’ve got a very interesting entry-level large sensor video camera in a tiny form factor.

We will take the Sony A5100 into the field and see how it performs in terms of usability. Watch out for our upcoming video review.

 $398 exc. VATMore Coming Soon
 
BUY IT

Leave a Comment

You are not subscribed to this post. Follow new comments

Login to comment

 Dmitri Tsitelauri Reply
Dmitri Tsitelauri August 22, 2014

So I guess this falls under good “B” camera for a7s? Photokina should be interesting. I would wait and see…

Tim Pek Reply
Tim August 22, 2014

Noise test would be awesome.

Luiz Guilherme Leste Reply
Luiz Guilherme Leste August 22, 2014

Renata S. Zamboni pensa seriamente nesse brinquedinho aqui. 13 stops de latitude.

Renata S. Zamboni Reply
Renata S. Zamboni August 22, 2014

A pocket tbm tem 13 stops… huum camera reserva mesmo? rsrsrs

Reply
Clayton Moore August 22, 2014

It seems to me that this is one more camera that’s making it more awkward for Canon to stay lackluster with video features at the “affordable by the masses” price point.

Reply
Agachart Sukchouy August 22, 2014

a5100 @ HD 0r FHD on test
,but GH4 on 4K that low reso. make few rolling
effect
please test all FHD(1080) ,it’ will look good.

Reply
Radek Wierzbicki August 23, 2014

Thank you for quick and informative preview. Regarding image detail, I’m under the impression that dialing it to -3 gives a strongly softened look. If it’s possible would you please consider testing different settings of sharpness in your video review?

Sebastian Wöber Reply
Sebastian Wöber August 24, 2014

We found that the results are better when sharpening is not done inside the camera. -3 is equal to “no internal sharpening”. With a good codec, sharpening in post should give more control and better results.

Reply
Radek Wierzbicki August 25, 2014

This was exactly the case with GH3 although on certain shots internal sharpening provided surprisingly good results. It is amazing how far Sony came with their mirrorless cameras over last few months. I’m looking forward to seeing some real world footage from a5100. Again thank you for your informative and well conducted work!

Reply
Suguru Boco August 23, 2014

Great review. I just bought a sony a6000 but I have the oportunity to return it and buy a sony a5100. My question is a5100 is better tan a6000 respect video?
Sony say that a5100 readout full the sensor and dont make a blinning, why is there aliasing in the test?
Thanks

Sebastian Wöber Reply
Sebastian Wöber August 24, 2014

We cannot tell you that as we didn’t have the a6000 in the lab. For now I would not return the a5100 based on these results as the a6000 could be very similar.

Reply
Suguru Boco August 25, 2014

Do you know if have a6000 a 8bits 4:2:2 HDMI output?

Reply
ScottyG August 24, 2014

Thank you for this series of tests. Would it be possible to include the Black Magic cameras with these? That would be great much appreciated.
Thanks.

Sebastian Wöber Reply
Sebastian Wöber August 24, 2014

Yes, we will also include the Blackmagic and any other cameras we can get a hold of. For now we’re slowly building this system so it is eventually solid and provides consistent, true results. It requires a lot of work and research for each individual test, so it takes time, but we’re on it.

Reply
Suguru Boco August 25, 2014

Have a6000 a 8bits 4:2:2 HDMI output?

Sebastian Wöber Reply
Sebastian Wöber August 25, 2014

We’re sorry to report that there was a bug in the software we used which has now been addressed, but leads to very different dynamic range results. Please see the update in the article. We apologise if we have caused any inconvenience. We relied too much on the software, which we will not do again.

Reply
AdRath August 25, 2014

Does this bug affect all of the other tests or is it just with Xavc-s or just the a5100 itself?

I’m trying to figure out how this bug relates to the other tests performed with the software.

Reply
Dave Meyer August 30, 2014

Thanks for the review.
Did you find out if the A5100 suffers from the same overheating problems as the NEX7 etc?
Also – is the maximum automatic shut-off 30mins?
Thanks.

Reply
Don Dagg September 11, 2014

Would the A7s and 5100 have similar dynamic range and image quality if S log 2 was not used. I think in the real world S log 2 will not be used that much outdoors during the day.

Sebastian Wöber Reply
Sebastian Wöber September 11, 2014

In a word: No

Reply
Don Dagg September 11, 2014

Is it the full sensor read out that makes the biggest difference? If you do not shoot in S log you will probably lose 3 to 4 stops of dynamic range which puts it in the A5100’s ballpark. I just want to understand before I make a decision.

Chris Quevedo Reply
chris September 11, 2014

these 2 cameras, the a7S and the a5100 are completely different. you can’t really compare one to the other. if you are thinking you can get even remotely the same dynamic range you are sorely mistake. i currently own the a6000 which sony claims is the same sensor as the a5100 i believe, and while its good, it is NOT a counterpart to the a7S. no contest. don’t buy a wrench and expect a hammer. this is NOT going to give you a7S performance. for that, get an a7S.

Reply
Don Dagg September 11, 2014

What other Picture Profile on the A7s offers great dynamic range but will let you shoot at low ISO’s. This is the real test for the A7s.

Chris Quevedo Reply
chris September 11, 2014

yes there are other profiles, but even in these the image is better on the a7S. cleaner with less noise, and there is worlds more flexibility with adjusting these profiles from everything to color gamut to black gamma, where the a5100 doesn’t have picture profiles at all.

Reply
Don Dagg September 11, 2014

Phillip Bloom said that the A6000 produced the best HD image he has seen from a stills camera (apart from the Canon 1DC). His only complaint when he reviewed it was the AVCHD codec. This was before the A7s. But this says alot about the A5100 and the new codec since it has the same sensor as the A6000.

Chris Quevedo Reply
chris September 11, 2014

hi again,
no i’m not disputing its a good camera. it is. but like you said, this was before the a7S. so the a5100 is more than likely an amazing camera. i’m only saying don’t expect a7S performance out of it even at the non Slog2 profiles, because, as i said, the a5100 and a6000 don’t have picture profiles AT ALL. so you can’t compare the a7S and the a5100 on an apples to apples basis for one thing. the a5100 most likely only lets you adjust contrast, saturation and sharpness, and thats it. there is no other settings you can change video picture wise. the a7S lets you control absolutely everything.

if its in your budget go ahead and get the a5100. its a good and adequate camera and you should be happy with it as long as you won’t miss the EVF. but its not an a7S competitor by any means

Reply
Don Dagg September 11, 2014

Thanks Chris. I appreciate you taking the time to give me your opinion.

Chris Quevedo Reply
Chris September 11, 2014

no problem. I own both the a7S and a6000 so if you have other specific questions let me know. since my a6000 is more or less identical to the a5100 in function I might be able to help

Reply
Chad Johnson September 14, 2014

Hi, I’m very interested in this camera for aerial video. I have the NEX-5r now, and it gets nice footage, but even with sharpening and contrast turned down, I’m getting aliasing and moire in rooftops and power lines. Would you say the a5100 un the 50mbps mode performs better in this respect?

Reply
Ievgenii September 24, 2014

I believe, you have done right when using Portrait profile at A5100.
But you would be even more right, if you would turn the DRO level 5 on.
This way you might get some better results in the Dynamic Range area.

I have also heard that A5100 does use so-called full matrix scan (no line skipping). And is interested if it is true or not and if true, then at which modes it is. I believe, we might expect it working in the 25-24p only.
Theoretically in this mode the sensitivity (snr) should be better by approx. one stop. That would be interesting to compare srn (noise level) of different video modes at high ISO. and compare it with A6000 as well

Sebastian Wöber Reply
Sebastian Wöber September 24, 2014

Hi levgenii,

To my knowledge and according to Google none of the things you describe exist.

Maybe you can elaborate.

The downsampling the camera does (or not) is not related to framerates. Quality in 30p or 25p is identical.

Reply
Ievgenii September 24, 2014

Dear Sebastian,
unfortunately I have no camera, but only the predecessor nex-5n.
in nex-5n, the Dinamic Range Optimizer exists, and DRO-Level-5 is the mostly used setting for the videoshooting.

while searching the internet for the A5100 and DRO I found this description:
http://www.bhphotovideo.com/c/product/1076400-REG/sony_ilce_5100l_b_alpha_a5100_mirrorless_digital.html

•Dynamic Range Optimizer (DRO) works to improve images featuring backlit subjects or scenes with high contrast where details can be lost in the shadows or highlights. This mode can be controlled automatically or fine-tuned using five settings.

So according to the article, the 5-level DRO exists unchanged.

Sebastian Wöber Reply
Sebastian Wöber September 24, 2014

I think I would have found it. I tried all possible settings and combinations. Are you sure it is available in video mode?

Chris Quevedo Reply
Chris September 24, 2014

it can do it.

http://Instagram.com/p/tVb48tu_c2/

Sebastian Wöber Reply
Sebastian Wöber September 24, 2014

I think that’s the A6000 there

Chris Quevedo Reply
Chris September 24, 2014

I was under the impression that minus the evf they were identical. sorry

Reply
Ievgenii September 24, 2014

Dear Sebastian,

small addition.
for Sony A6000 there is a measurement of the DRO here
http://www.dpreview.com/reviews/sony-alpha-a6000/11
you can try different DRO levels there and see how the graph changes.
Seems that DRO level 5 gives about 1 stop of extra DR.

And IMO, the DRO 5 – is something which is quite similar to the Panasonic Cinema mode or Sony Slog2.

Sebastian Wöber Reply
Sebastian Wöber September 24, 2014

Ok thank you, I will look for it when we get the camera again, or maybe someone else can share if the mode exists on A5100.

Reply
Chongnuo Ji October 19, 2014

Not sure about A5100.
But all the Sony NEX and new A series mirroless have the DRO setthing when recording video.And It really works.

I have tested:NEX-5N,NEX-6,A3000,ILCE-7.

Reply
Chad Johnson October 20, 2014

I was not aware of DRO. I have it on my NEX-5r, and it was on AUTO. I’m going to try it at level 5 and see what I get. I do aerial video, and this will help with shadows.

Reply
Chongnuo Ji October 20, 2014

Yes, manual level bring up more shadow detail than AUTO.I mostly use level 1.

Reply
kevin January 24, 2015

The mode does exist on the a5100. It’s in the 5th tab of the “Camera Settings” menu labeled as “DRO/Auto HDR”

Lorenzo Canina Reply
Lorenzo Canina November 9, 2014

Do you think it will out perform the LX100 in low light, moire, aliasing? I am deciding what to buy and it would mean the world to me! Thanks!

Reply
teddoman March 3, 2015

Can anyone confirm if the A5100 really has full sensor readout? Would love to get comparison footage against the A6000 (which doesn’t).

Ian McKay Reply
Ian McKay June 14, 2015

I’m still seeing green noise what must I do to eliminate this problem?

Sebastian Wöber Reply
Sebastian Wöber June 14, 2015

Hi Ian, I think this is a “problem” specific to the camera. You will have to raise your black levels in post to get rid of it.