A more realistic impression of the Sony A7S low light performance at ISO 12,800

[jcs]

For stills, the 5D Mark III might be better due to color science (and for sure higher resolution) even though the A7S has more DR. For video, the A7S is looking to be a worthy replacement for the 5D3, including RAW. To clarify, 14-bit RAW from the 5D3 will likely provide a better color image, however when factoring in recording time, disk space, in-camera low-light NR, and processing time, 50Mbit/s XAVC-S will be more detailed and cleaner than 5D3 RAW with a lot less work and disk usage. The A7 & A7R are clearly better for stills vs. the A7S.

 

It's looking like the A7S sensor isn't the real magic*: it appears to be the trick NR processor for video. If reports regarding lower than normal battery performance are accurate for the A7S, this could be from the hardware NR processing. This could also explain why they couldn't do 4K internally: they're at their thermal and CPU limit just doing 4K = > 1080p + NR + H.264.

 

 

 

* If the sensor was truly lower noise from (gapless) bigger pixels, still images should be cleaner on the A7S vs. the A7 & A7R. However, downsampling higher resolution to lower resolution also reduces noise (requires more computing power- an issue with video).

[Bob Wall]

For battery performance, sony lists ~340 still shots as the performance for all the A7 cameras, and all the NEX cameras too.  For anybody who has any of these cameras, what does this translate to for video shoot time?  I'm thinking of my 5D and I am sure it can do waaay more than 340 stills on a single battery but I guess I never counted.   I just looked in a folder that I left it going for a timelapse the other day and there were more than 1300 shots, and I'm sure the battery wasn't half drained.

[noone]

Why is the DR so low ???? :((((( they claimed 15.3....  and now it is less than Nikon d5200 --' 

Only up to ISO 200.   At ISO 3200 the Sony is about a stop ahead and by the time it gets to ISO 12800 the Nikon is more than 2 stops behind.    Even the 5d mk iii is two stops behind at ISO 12800.

 

13 stops at base ISO is still pretty good though maybe not quite what was expected but how many have been waiting to see it as a good light camera?     Above ISO 1600 and things look great.

[leeys]

Why is the DR so low ???? :((((( they claimed 15.3....  and now it is less than Nikon d5200 --' 

This is pretty much how most sensors tweaked for high ISO respond. Average low ISO DR but fantastic high ISO DR.

 

I am pretty sure that the 15 stop claim is not possible unless Sony had some breakthrough in sensor tech, which doesn't seem to be present.

 

As I mentioned, the JPEGs are fantastic. No one has any public tools available right now to decode the raw files though. I'd love to see what they're like.

[horshack]

For stills, the 5D Mark III might be better due to color science (and for sure higher resolution) even though the A7S has more DR. For video, the A7S is looking to be a worthy replacement for the 5D3, including RAW. To clarify, 14-bit RAW from the 5D3 will likely provide a better color image, however when factoring in recording time, disk space, in-camera low-light NR, and processing time, 50Mbit/s XAVC-S will be more detailed and cleaner than 5D3 RAW with a lot less work and disk usage. The A7 & A7R are clearly better for stills vs. the A7S.

 

It's looking like the A7S sensor isn't the real magic*: it appears to be the trick NR processor for video. If reports regarding lower than normal battery performance are accurate for the A7S, this could be from the hardware NR processing. This could also explain why they couldn't do 4K internally: they're at their thermal and CPU limit just doing 4K = > 1080p + NR + H.264.

 

 

 

* If the sensor was truly lower noise from (gapless) bigger pixels, still images should be cleaner on the A7S vs. the A7 & A7R. However, downsampling higher resolution to lower resolution also reduces noise (requires more computing power- an issue with video).

 

Not sure what you mean by Canon's color science - Canon sensors are some of the worse in terms of metamersim and color selectivity. Also not sure what you mean by a "trick" specific to video - the DxoMark A7s results measured stills performance and demonstrate a 1.7EV improvement in High ISO DR/shadow noise over the current best FF sensor on the market at ISO 102,400. This is exhibited in the High ISO A7s raws posted earlier in this thread.

[quobetah]

DXOMark must be using older firmware because this is some serious DR from "actual" A7s shoots via Matt Allard's twitter: https://twitter.com/mattallardACS/status/478914394785472512/photo/1

 

[quobetah]

And they tested at ISO 100, where the base ISO of the A7s is 3200 "Base ISO of #a7s in Slog2 is 3200!!!-Matt Allard". Base ISO = Max DR.

[leeys]

Ok, so I downloaded RawTherarpee and managed to get the RAWs working. Better than the JPEGs, so certainly looks good. Definitely not super duper THREE STOPS ZOMG good though.

 

And seriously, some of the posts I see here...

[jcs]

Not sure what you mean by Canon's color science - Canon sensors are some of the worse in terms of metamersim and color selectivity. Also not sure what you mean by a "trick" specific to video - the DxoMark A7s results measured stills performance and demonstrate a 1.7EV improvement in High ISO DR/shadow noise over the current best FF sensor on the market at ISO 102,400. This is exhibited in the High ISO A7s raws posted earlier in this thread.

 

Color science as in the reason the ARRI Alexa and Canon 5D3 (with 14-bit RAW: "Baby Alexa") are so popular: most folks find people and skintones look better with those systems (vs. RED (the new Dragon is changing this), Panasonic (new GH4 is not bad), Sony (especially bad skintones and/or hard to work with color in many cases (A7S looks better so far)), etc.). A low metamerism score means about the same as a low CRI score: not much with real world images), From: http://***URL removed***/forums/post/53084809:

 

 

The OP's idea that better metamerism should produce better skin tones is theoretically correct, but in practice it is not. For example, many Nikon models with horrible skin tones have much better metamerism than the Canon models famous for superior skin tones. You can safely ignore the DXO metamerism measurements. Either it is irrelevant in real life or the DXO protocol is too far from real life situations (e.g. AWB in artificial light).

 

Regard a "trick" feature- while I don't have an A7S to test yet, from example videos and folks comments it sounds like there is more effective NR happening in video than with stills. The GH4 does temporal NR (as does Neat Video), and I would expect Sony to be doing the same. Since the low ISO DR isn't even as high as the D800 (and certainly not Sony's claimed 15.3, at least for the DXO test), it appears the higher ISO extended DR is due to image processing (including powerful new NR) and not necessarily the sensor (depends on how and if analog gain affects sensor performance at higher ISOs). This could be a reason battery life isn't reported as being very long.

[horshack]

Color science as in the reason the ARRI Alexa and Canon 5D3 (with 14-bit RAW: "Baby Alexa") are so popular: most folks find people and skintones look better with those systems (vs. RED (the new Dragon is changing this), Panasonic (new GH4 is not bad), Sony (especially bad skintones and/or hard to work with color in many cases (A7S looks better so far)), etc.). A low metamerism score means about the same as a low CRI score: not much with real world images), From: http://***URL removed***/forums/post/53084809:

 

Regard a "trick" feature- while I don't have an A7S to test yet, from example videos and folks comments it sounds like there is more effective NR happening in video than with stills. The GH4 does temporal NR (as does Neat Video), and I would expect Sony to be doing the same. Since the low ISO DR isn't even as high as the D800 (and certainly not Sony's claimed 15.3, at least for the DXO test), it appears the higher ISO extended DR is due to image processing (including powerful new NR) and not necessarily the sensor (depends on how and if analog gain affects sensor performance at higher ISOs). This could be a reason battery life isn't reported as being very long.

 

Still not following your color comments. Professional video is almost always color graded so why would the out-of-box color defaults have significance here for this class of product? And low metarism is significant for reproducing correct colors in differentrs specta of light. As for the video NR, again the rendered JPEGs from A7s raws posted are as clean as the  A7svideos that have been posted thus far.

[jcs]

Still not following your color comments. Professional video is almost always color graded so why would the out-of-box color defaults have significance here for this class of product? And low metarism is significant for reproducing correct colors in differentrs specta of light. As for the video NR, again the rendered JPEGs from A7s raws posted are as clean as the  A7svideos that have been posted thus far.

Reading about labs tests (or doing them yourself) provides useful insight (especially metamarism and CRI), however actual real world testing is needed to understand how the devices really perform. For example, showing clients and talent skintones shot on the Sony FS700 compared to the Canon 5D Mark 3, most of them much prefer the Canon. I can spend a lot of time trying to make the FS700 look like the Canon, and while I can get very close, many times it's not worth the effort. This is not just my findings, many others have the same experience and that's why the C100/C300 cameras sell so well and are used so much even though their specs were dated even at launch. 5D3 RAW looks so good folks call it the baby Alexa. The FS700 rocks at slomo (for the price).

[animan]

its often a matter of taste, personally i would choose nikon skintones any day over canon, not a fan of the orange bias..

[leeys]

Reading about labs tests (or doing them yourself) provides useful insight (especially metamarism and CRI), however actual real world testing is needed to understand how the devices really perform. For example, showing clients and talent skintones shot on the Sony FS700 compared to the Canon 5D Mark 3, most of them much prefer the Canon. I can spend a lot of time trying to make the FS700 look like the Canon, and while I can get very close, many times it's not worth the effort. This is not just my findings, many others have the same experience and that's why the C100/C300 cameras sell so well and are used so much even though their specs were dated even at launch. 5D3 RAW looks so good folks call it the baby Alexa. The FS700 rocks at slomo (for the price).

I have to say, you're quite right on this. Many consumers love the skin tones from the Canons. This extends to the stills as well, so a lot of wedding photographers use the 5D series cameras.

[Michael Ma]

Only up to ISO 200.   At ISO 3200 the Sony is about a stop ahead and by the time it gets to ISO 12800 the Nikon is more than 2 stops behind.    Even the 5d mk iii is two stops behind at ISO 12800.

 

13 stops at base ISO is still pretty good though maybe not quite what was expected but how many have been waiting to see it as a good light camera?     Above ISO 1600 and things look great.

Maybe 15.3 stops is with S-Log2 for video after some noise filtering?

[hmcindie]

Still not following your color comments. Professional video is almost always color graded so why would the out-of-box color defaults have significance here for this class of product? And low metarism is significant for reproducing correct colors in differentrs specta of light. As for the video NR, again the rendered JPEGs from A7s raws posted are as clean as the  A7svideos that have been posted thus far.

 

Because - for example - the FS700 looks quite bad. I've never gotten a good "people" look out of it (and I've tried). On all the shoots I've used it as an A-cam and the 5d as the B-cam, I've actually liked the image from the 5d more. Trying to grade the FS700 to match the 5d is just... hard.

[Andrew Reid]

And they tested at ISO 100, where the base ISO of the A7s is 3200 "Base ISO of #a7s in Slog2 is 3200!!!-Matt Allard". Base ISO = Max DR.

 

That isn't true at all. Base ISO is always 200 on DSLR sensors. You will not get the best dynamic range at ISO 3200.

Also the tunnel shot which Matt took a pic of from a Sony TV at a show - none of you shot with the A7S on that day so none of you have any idea what the real light levels were under the tunnel. Zero proof of the dynamic range abilities of the camera at all!

[Nikkor]

That isn't true at all. Base ISO is always 200 on DSLR sensors. You will not get the best dynamic range at ISO 3200.

Also the tunnel shot which Matt took a pic of from a Sony TV at a show - none of you shot with the A7S on that day so none of you have any idea what the real light levels were under the tunnel. Zero proof of the dynamic range abilities of the camera at all!

Nah, base ISO can be 200 but isn't always there. All the actual tipical sony sensors have base iso 100. Anyway, base iso 3200 sounds unreasonable because at iso 100 you would have horrible DR where it is most needed (daylight).

[Chris Gibbs]

Andrew Reid Wrote:

 

>Base ISO is always 200 on DSLR sensors<

 

​For all sensors & from all manufacturers?  Please explain this.

[Policar]

That isn't true at all. Base ISO is always 200 on DSLR sensors. You will not get the best dynamic range at ISO 3200.

Also the tunnel shot which Matt took a pic of from a Sony TV at a show - none of you shot with the A7S on that day so none of you have any idea what the real light levels were under the tunnel. Zero proof of the dynamic range abilities of the camera at all!

 

SLOG does a digital push (like highlight tone priority) to retain highlight detail beyond what you'd get at normal gammas.

 

The F5 has an 800 ISO base in normal gammas, but 2000 ISO in SLOG2, so a stop and a half push (not-so curiously it has a stop and a half more highlight detail than the C300)....

 

If SLOG 3 on the A7S has "1300%" more dynamic range, as claimed, that's a 3.7 stop push from 200 ISO. That gets you to just under 3000 ISO. Reasonable that 3200 ISO would be the base for SLOG 3, due to the digital push, but weird. Fwiw, I do not like the F5's tonality in shadows using SLOG 2 but the dynamic range is better than anything short of the Alexa, which has more DR AND better tonality and more pleasing noise structure... and better everything.

 

The C100/C300 low light test is irrelevant and incompetently handled, fwiw. Whoever performed it approached it wrong. Both cameras should look the same first of all, secondly there's no stopping down to compensate for the ISO push. 80,000 ISO on these cameras is terrible.

[Martin BetterFly]

check out "yak films"  filming on the A7S preproduction model

it looks very nice, sometimes clean and then not so clean (due to different sensor modes?, as they say in the description they have used some vintage nikon glass on it) we can also see the famous rolling shutter but then also some choppy motion (no ND filters I guess or I don't know) 

what you think guys?