Sony A7S III

[MeanRevert]

On 8/12/2020 at 12:27 PM, Trek of Joy said:

More SOOC footage at 50/100p from the "No Limits" launch video by Jacques Crafford. 

https://drive.google.com/drive/folders/1kYjaH84X9871d8-Rmuj9DOXfmtAH6V6i?fbclid=IwAR3LPZ7njD6UadrWtejYeGQHq5hfTCNOKXBrdliOZ4w4cvrja5IrzGgRnhY

What lens was used to shoot this?  And anyone understand the comment made about using Gamma Assist to set exposure and WB?

[Rinad Amir]

4 hours ago, MeanRevert said:

What lens was used to shoot this?  And anyone understand the comment made about using Gamma Assist to set exposure and WB?

24-70gm s-gammut3cin slog3

[hmcindie]

9 hours ago, Hanriverprod said:

Just because you get a group of people doesn’t make them right. The human race is a group of people too. 

ACES workflow is pretty much for matching CGI shots and composites to shot material. It works great. Except maybe in AE where it can break some effects.

The point is that if you can match cameras to one another with ACES (and other workflows) then your point about "different looks that can't be changed ever" is kinda wrong.

[Hanriverprod]

27 minutes ago, hmcindie said:

ACES workflow is pretty much for matching CGI shots and composites to shot material. It works great. Except maybe in AE where it can break some effects.

The point is that if you can match cameras to one another with ACES (and other workflows) then your point about "different looks that can't be changed ever" is kinda wrong.

I'm aware you can match shots but most content you see from this camera is just not right. Anyway I have a Sony camera and kind of tired of always working to not make it look sickly. I was hoping this camera would be Sony's answer but it doesn't look that great either. It looks better but it still looks wrong especially indoors or with mixed lighting and any somewhat challenging outdoor shots. Almost everything from the R5 and R6 looks nice but those things are obviously broken.

[Platto]

1 hour ago, Hanriverprod said:

I'm aware you can match shots but most content you see from this camera is just not right. Anyway I have a Sony camera and kind of tired of always working to not make it look sickly. I was hoping this camera would be Sony's answer but it doesn't look that great either. It looks better but it still looks wrong especially indoors or with mixed lighting and any somewhat challenging outdoor shots. Almost everything from the R5 and R6 looks nice but those things are obviously broken.

is that bad? i mean Sony has improved its color science since a7s ii also take look at footage that Mr Jacques Crafford has provided 

 

[Rinad Amir]

42 minutes ago, Platto said:

is that bad? i mean Sony has improved its color science since a7s ii also take look at footage that Mr Jacques Crafford has provided 

 

Just downloaded and played in Resolve used Aces and the colors are just amazing i have to say tho h265 is really chocking my system.

[Django]

Just downloaded a few clips and did some grading in Resolve. Skin tones are still bust in this camera, typical Sony zombie green/yellow tint. That being said the huge codec means it's an easy fix. Definitely the nicest quality files out of a Sony mirrorless, by far.

[aaa123jc]

I downloaded the footage by Jacques Crafford. First, thank you Jacques Crafford for sharing these footage so I can test them out. 

So I put it in Resolve, tweaked a little bit and used a color space transform. It took me about a minute or two. Here is the quick result.

Of course this is far from a graded image, only a starting point. 

I have to say this is quite an improvement over the A7S and A7S2. People argue otherwise likely never use an A7S before because its color science is really really bad. I will not say this new color from A7S3 can rival Arri Alexa, but this is definitely a usable image, especially if you consider the codec is pretty solid so you can do quite a lot of adjustments without breaking it. 

Surely, color is subject so everyone can have their own opinion, but I feel like claiming these images are horrible may be a little bit exaggerated. 

Personally for me, although the color from this A7S3 may not be the best, it is totally usable and if one cannot make a good picture of it, he is the one to be blamed. If you think my quick result is very bad, probably it is my fault too. 

[Kino]

This is some of the best footage I have seen from the camera. A7SIII colors and DR rival 16 bit linear RAW from F5/F55 from a few years ago, all of that in a 10 bit 4:2:2 codec:

In the hands of a capable shooter, the lack of internal RAW is no limitation at all.

[Video Hummus]

8 minutes ago, Kino said:

This is some of the best footage I have seen from the camera. A7SIII colors and DR rival 16 bit linear RAW from F5/F55 from a few years ago, all of that in a 10 bit 4:2:2 codec:

In the hands of a capable shooter, the lack of internal RAW is no limitation at all.

Brandon Li’s work. He is super pragmatic. Says the camera is an upgrade in almost every way for him (he is not a stills shooter). He didn’t use a PP either.

I think the main highlight in this video though is the autofocus. It’s amazing!

[Llaasseerr]

22 hours ago, Video Hummus said:

That leaves DR, highlight roll of, compression artifacts, and the mystical motion cadence as the biggest differentiators. You can tweak the highlight roll off a tiny bit, but not much. That is mostly down to how good the sensor is, ie 16-bit linear off an ARRI versus 12-bit or 14-bit log off of something like this A7Siii.

 

It's a common myth that there is "highlight rolloff" from a sensor. It reads linear light values and hard clips at a certain point - there is no highlight compression, unlike with film negative. Highlight rolloff is a product of how much dynamic range a sensor captures combined with the filmic s-curve applied either in-camera or in post.

If you want more clarification on this, here are some links:

https://www.fie.us/2013/09/09/basic-color-science-for-cinema/

http://yedlin.net/OnColorScience/

 

[androidlad]

52 minutes ago, Kino said:

This is some of the best footage I have seen from the camera. A7SIII colors and DR rival 16 bit linear RAW from F5/F55 from a few years ago, all of that in a 10 bit 4:2:2 codec:

In the hands of a capable shooter, the lack of internal RAW is no limitation at all.

It has good cinematography, but it doesn't show a lot of DR, especially in highlights (F5/F55 don't have much highlight DR either, around 6EV).

Pay close attention to the way the sun and the surrounding clouds were blown out,

[androidlad]

1 minute ago, Llaasseerr said:

It's a common myth that there is "highlight rolloff" from a sensor. It reads linear light values and hard clips at a certain point - there is no highlight compression, unlike with film negative. Highlight rolloff is a product of how much dynamic range a sensor captures combined with the filmic s-curve applied either in-camera or in post.

If you want more clarification on this, here are some links:

https://www.fie.us/2013/09/09/basic-color-science-for-cinema/

http://yedlin.net/OnColorScience/

 

There's absolutely highlight roll-off on a digital sensor with dual/multiple gain readout. The roll-off is applied when merging the high gain and low gain readout path, which is how ARRI does it.

[Llaasseerr]

Just now, androidlad said:

There's absolutely highlight roll-off on a digital sensor with dual/multiple gain readout. The roll-off is applied when merging the high gain and low gain readout path, which is how ARRI does it.

I considered the dual gain architecture of the ALEV III. If you can point to a link of how the in-sensor highlight rolloff works, I'd be interested to understand. But it makes more sense to me that the high gain and low gains readouts are merged into a single high dynamic range linear image before being converted to log.

[androidlad]

Just now, Llaasseerr said:

I considered the dual gain architecture of the ALEV III. If you can point to a link of how the in-sensor highlight rolloff works, I'd be interested to understand. But it makes more sense to me that the high gain and low gains readouts are merged into a single high dynamic range linear image before being converted to log.

It's not done "in-sensor". The two readout paths are both derived from 14bit ADCs, they are processed and merged by a 16bit ISP. Roll-off as well as other ARRI colour science processing happen when doing this 28bit > 16bit conversion.

[Simon Young]

13 minutes ago, androidlad said:

It has good cinematography, but it doesn't show a lot of DR, especially in highlights (F5/F55 don't have much highlight DR either, around 6EV).

Pay close attention to the way the sun and the surrounding clouds were blown out,

In the behind the scenes video also published today Brandon Li specifically states he shot the whole project in the standard picture profile - hence not the greatest display of the camera's dynamic range capabilities.

[Llaasseerr]

4 minutes ago, androidlad said:

It's not done "in-sensor". The two readout paths are both derived from 14bit ADCs, they are processed and merged by a 16bit ISP. Roll-off as well as other ARRI colour science processing happen when doing this 28bit > 16bit conversion.

Yes, that's exactly what I'm saying.

[Llaasseerr]

20 minutes ago, Simon Young said:

In the behind the scenes video also published today Brandon Li specifically states he shot the whole project in the standard picture profile - hence not the greatest display of the camera's dynamic range capabilities.

Makes sense.

[noone]

Question from the clueless!

Is the A7siii 10 bit only in log (like the Canon R5 which is otherwise 8 bit 4:2:0 when not in C log) or is it 10 bit in standard picture profiles too? 

Does it make a difference anyway?

 

 

[Geoff CB]

1 minute ago, noone said:

Question from the clueless!

Is the A7siii 10 bit only in log (like the Canon R5 which is otherwise 8 bit 4:2:0 when not in C log) or is it 10 bit in standard picture profiles too? 

Does it make a difference anyway?

 

 

10-bit in all profiles.