August 23, 2014

You Might Be Surprised by the Dynamic Range of the New $550 Sony A5100

Earlier in the week, we were introduced to a brand new bite-sized camera from Sony, the A5100. We already knew that Sony's new camera, which comes in at $550 for the body, would have the ability to output uncompressed 8 bit 4:2:2 via the mini HDMI port and be able to record to the XAVC-S codec, both firsts for a camera of this size and price. However, we had no idea what kind of performance (in terms of dynamic range, rolling shutter, and overall image quality) would be possible with the camera's CMOS sensor. Luckily, just like they did with the A7s, the folks at Cinema5D put the A5100 to the test, and their results are fairly exciting.

[UPDATE: We just got word from Cinema5D that the test results for dynamic range performance were skewed because of a software bug. Here's their official statement:

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.

Below is the chart with the updated dynamic range results. Sadly, the A5100 doesn't fare nearly as well this time around.]

Test-Scores_DR_sonya5100_corr

First up is dynamic range. Just like in their previous tests with the A7s, they used a DSC Labs Xyla 21 dynamic range chart. Here's how they tested the A5100, which doesn't have a specific native ISO (or at least it's not clear what that ISO is):

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 measured stops with Creative style “Portrait (-3, 0, -3)” and 12 with Creative Style “Standard (-3, 0, -3)”.

Here are the dynamic range results that Cinema5D labs came up with:

Test-Scores_DR_sonya5100

Cinema5D later updated their post with the following clarification about their results:

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 cameras, 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.

Even considering that Cinema5D doesn't think that the A5100 really has 13 full stops of dynamic range, this is nevertheless an incredibly impressive result considering both the price and size of the camera. What's even more impressive is that even though the A5100 doesn't have any significant control over gamma interpretation (i.e. advanced picture and log profiles), it's still able to pull at least 12 stops of dynamic range. It's fun to think of what kinds of results this test would have yielded if a logarithmic curve, maybe even S-log, could be applied to the footage in camera.

At this point, there's not much reason to think that Sony would spend the money to incorporate advanced gamma profile control into the A5100, especially considering that this is a basically a lower-end prosumer camera. For that reason, it seems even less likely that an S-log update might happen in the future. With that said, at the $550 price point, the A5100 has the potential to be a really killer B-cam, C-cam, or crash cam.

The folks at Cinema5D Labs also tested the rolling shutter of the A5100 (where it performed relatively well), and they tested the overall image quality and sharpness of the A5100 against the A7s. To see those tests and get the full rundown on the results, head on over to Cinema 5D.

Link: Lab Review - Sony A5100 - Mirrorless Dynamic Range Power for $550 -- Cinema5D

Your Comment

25 Comments

The rolling shutter does not change between 25 and and 50 fps nor its quality, interesting results. Add a log curve and 422 out put and I will have to put this camera on my to do list.

August 23, 2014 at 9:03PM, Edited September 4, 11:56AM

0
Reply
edgar

well, there is a 4:2:2 output via HDMI, but I agree. Add s-log and we have a winner.

August 23, 2014 at 9:27PM, Edited September 4, 11:56AM

0
Reply
David A

8-bit doesn't count.

August 24, 2014 at 11:08AM, Edited September 4, 11:56AM

4
Reply
Jangee Slardoon...

For a 550$ camera does.

August 24, 2014 at 4:33PM, Edited September 4, 11:56AM

0
Reply
edgar

There are rumors that GH4 will add a Log type feature with its next firmware update, which may be announced at Photokina or earlier.

August 23, 2014 at 9:33PM, Edited September 4, 11:56AM

0
Reply
DLD

Why they didnt publish any videos, if they have the camera?

August 23, 2014 at 9:57PM, Edited September 4, 11:56AM

0
Reply
Laurel

They are publishing videos very soon.

August 23, 2014 at 11:31PM, Edited September 4, 11:56AM

0
Reply
avatar
Joe Marine
Camera Department

GH4 at 10.9??? That's a laugh, I own the GH4 and CineLikeV might have that but CineLikeD is way more than 10.9.

August 23, 2014 at 10:57PM, Edited September 4, 11:56AM

0
Reply
Dave

Thinking about getting one of these for a two year foreign excursion. Want something small and light for photography (leaving the BMPCC at home), but also want something that will be useful afterwards… I’m really waiting for a mid-range f2.8 zoom e-mount. Any news on that? I’m allergic to kit lenses.

I primarily do film work so the 60fps 1080p looks really nice, along with all the other features. This is a service trip though so I’ll be sticking to one lens and keeping it low profile.. That’s why I’m curious about the lens

August 24, 2014 at 12:04AM, Edited September 4, 11:56AM

0
Reply
Kent

Get the A6000. A5100 may have slightly better video but you will need the viewfinder if you are just using the camera and lens.

August 24, 2014 at 12:21AM, Edited September 4, 11:56AM

17
Reply
Geoff C. Bassett

Bad advoce, the a6000 cant even compete with an old nex5. Its a really bad camera, I had to return it.

August 24, 2014 at 3:17AM, Edited September 4, 11:56AM

0
Reply
MrL

What a load of crap - I have an NEX5N and an A6000, and my A6000 absolutely kills my NEX5N.

August 24, 2014 at 4:30AM, Edited September 4, 11:56AM

12
Reply
Johnny

No way no how.

August 24, 2014 at 5:35AM, Edited September 4, 11:56AM

0
Reply
MrL

I also had both at the same time.

August 24, 2014 at 5:36AM, Edited September 4, 11:56AM

0
Reply
MrL

Indeed, the A6000 was a big leap ahead:

http://www.eoshd.com/content/12428/surprise-sony-alpha-a6000-video-mode-...

And the A5100 built upon and extended the improvements from the A6000.

Glad to see A5100 getting on NoFilmSchool some of the recognition which it deserves! :-)

It really is one of the top hybrid cameras on the market aside from the obvious two of the GH4 and A7s (which are at an entirely different price point).

August 24, 2014 at 6:40AM, Edited September 4, 11:56AM

0
Reply

Hmm,

GH4 raw stills give 12.8 stops according to DXOMark and natural settings give over 11

http://www.eoshd.com/content/12969/dxomark-rates-gh4-13-stops-dynamic-range-1-stop-better-5d-mark-iii

August 24, 2014 at 12:13AM, Edited September 4, 11:56AM

0
Reply

You have a point over there. But the GH4 is not recording raw video. Which brings me to the next question. How much dynamic range would the DXOMark give to the A5100?

August 24, 2014 at 2:09AM, Edited September 4, 11:56AM

8
Reply
edgar

Probably the same as A6000.

August 28, 2014 at 9:07PM, Edited September 4, 11:56AM

2
Reply
deep

yeah i love cinema5d, but every since they started doing tests, some of their results are really iffy. The GH4 compressed video stills have more dynamic range than canon 5d mark3 raw,,

August 24, 2014 at 4:56AM, Edited September 4, 11:56AM

4
Reply
Cjay

Agree, I started taking their tests with a grain of salt after they said the a6000 had less moire than the nex5t, my test shows the exact oposite.

August 24, 2014 at 5:38AM, Edited September 4, 11:56AM

0
Reply
MrL

Why I´m truly exited about those little cameras is, that multi-angle shooting becomes more and more affordable.
Some might use it as an A-cam, but I´m just looking forward never to use a GoPro again (well, rumors say they might branch out, but GoPro's as they are ATM I mean).
I have a very, very small travel-slider, that ended up bing to small even for a GH3, I have a carbon travel jib, that just moans when loaded with a C300 - for those 5% of specialty shots I end up using in the final edit, a 5100 will be very welcomed...

August 24, 2014 at 6:25AM, Edited September 4, 11:56AM

0
Reply
BearWithMe

I really wished that they would test the BMPCC and the Red Dragon for dynamic range.

August 24, 2014 at 7:22AM, Edited September 4, 11:56AM

4
Reply

I Agree

August 24, 2014 at 8:25AM, Edited September 4, 11:56AM

0
Reply
Ghostdog

C5D updated the results to a realistic 10.5 stops DR

Would be nice if they drew as much attention to the false claims as they did to the original #'s. I guess corrections don't drive hits as much. Just more misinformation people will run with an not check back for correct numbers.

I wonder if this software bug effected the other cameras or just the xavc-s cameras or just the a5100. A little more detail would be nice from a 'testing lab'

August 25, 2014 at 3:17PM, Edited September 4, 11:56AM

0
Reply
AdRath

5d more dr than the c300??? who is doing these tests and how exactly do they conclude these silly results. i shoot with the 5d and c100 and i can assure you that the 5d's sensor doesnt have 11.8 stops, more like 10.5 and the gh4 about the same at best.

August 28, 2014 at 6:19AM, Edited September 4, 11:56AM

13
Reply
marko