GH5s? New low light monster?

[Samuel H]

GH5s vs a7S II tests posted on dpreview. No, the GH5s is not the new king of low light, I'd say the a7S II is a stop and a half better (compared with GH5s ISO 12800, a7SII ISO 25600 looks better but ISO 51200 looks worse).

Still an impressive feat for such a small sensor. But no, it's not as good as some initial testers said. I think they were fooled by the very strong noise reduction that the GH5s seems to be applying (noise is low, but the image goes soft very fast as you increase the ISO, and edges become swirly).

https://www.dpreview.com/news/2333575124/panasonic-lumix-dc-gh5s-studio-scene-posted

 

[TheDingo]

GH5s vs a7S II tests posted on dpreview. No, the GH5s is not the new king of low light, I'd say the a7S II is a stop and a half better (compared with GH5s ISO 12800, a7SII ISO 25600 looks better but ISO 51200 looks worse).

Those are still-photo tests. With almost all of the actual VIDEO tests I've seen the GH5s has less noise and better color than the A7SII at ISO speeds up to 25K.

 

[TheDingo]

 

[Thomas Smet]

GH5s vs a7S II tests posted on dpreview. No, the GH5s is not the new king of low light, I'd say the a7S II is a stop and a half better (compared with GH5s ISO 12800, a7SII ISO 25600 looks better but ISO 51200 looks worse).

Still an impressive feat for such a small sensor. But no, it's not as good as some initial testers said. I think they were fooled by the very strong noise reduction that the GH5s seems to be applying (noise is low, but the image goes soft very fast as you increase the ISO, and edges become swirly).

https://www.dpreview.com/news/2333575124/panasonic-lumix-dc-gh5s-studio-scene-posted


View attachment 128829

View attachment 128830

With regards to video I think the reason why people consider it the new low light champ is due to the more frequently used ISO ranges to "get the shot". Most users will use ISO 400 to ISO 6400 as a realistic range for most situations before realizing they should probably use some kind of lights. So far the GH5s seems to do better in those ranges. The A7S can of course suck in more light but that is strictly from an engineering perspective and not really a useable use case for anybody that actually cares about quality. Fine for a scientific experiment but not for something requiring high production value.

So it is really about how clean the two cameras look at ISO200, 400, 800, 1600, 3200 and 6400 that makes people feel like the GH5s does better.

 

[Ron Evans]

This is great if you want to sync multiple Panasonic cameras together... but what about audio?

I guess I would include this in other equipment when I mentioned studio sync in my last sentence. The example people tend to use is to sync cameras for multicam and I think if the cameras are all Panasonic there is no need for this cable and with some or all using the XLR adapter one can get high quality sound recorded by the GH5 to and avoid the sync issue with sound recording separately though one has to recognize the different distances of the cameras from sound source that will easily be out of lip sync anyway even for separate audio recording depending where the mic is placed.

 

[Joshua Cadmium]

GH5s vs a7S II tests posted on dpreview. No, the GH5s is not the new king of low light, I'd say the a7S II is a stop and a half better (compared with GH5s ISO 12800, a7SII ISO 25600 looks better but ISO 51200 looks worse).

Still an impressive feat for such a small sensor. But no, it's not as good as some initial testers said. I think they were fooled by the very strong noise reduction that the GH5s seems to be applying (noise is low, but the image goes soft very fast as you increase the ISO, and edges become swirly).

https://www.dpreview.com/news/2333575124/panasonic-lumix-dc-gh5s-studio-scene-posted

I'd agree that the A7S II looks to be somewhere around 1.5 stops better when looking at that dpreview comparison. (When comparing in raw, at a two stop difference, you can clearly see that the GH5s looks better than the A7S II).

That's great news, because it does confirm what some people were saying, which is that the GH5s is better than the A7S II when thinking about equivalency and depth of field, since the A7S II already has a two stop advantage with a larger sensor. At the same quantity of light, the GH5s is going to be somewhere around half a stop better.

So, the GH5s is the "new king of low light" at least when total photons hitting the sensor is taken into consideration.

 

[TheDingo]

So, the GH5s is the "new king of low light" at least when total photons hitting the sensor is taken into consideration.

For most of us here it's all about the video image, so before we start talking about DOF equivalence or adding Speedboosters to gain 1 - 1.5 f-stops of additional light, the GH5s appears to easily handle the A7SII at ISO speeds up to 25K when shooting video.

Once you add on DOF equivalence and Speedboosters then the GH5s should consistantly outperform the A7SII in most low-light environments shooting video or not.

 

[Thomas Smet]

Comparing to other cameras on DP Review test tool the GH5s really does well against pretty much every single APS-C and most FF cameras. Any advantage some FF cameras may have over the GH5s is quickly eliminated once you do add a Speedbooster.

Lets also not forget the dual native ISO. Not sure if that works on raw stills so that DP Review tool may not be showing us exactly how good the ISO levels are above ISO 2500 when shooting video. Having the same general noise and dynamic range at ISO 2500 vs ISO 400 is pretty huge compared to any other camera out there minus the EVA1 and Varicam. Even the A7S drops to 11 stops of DR at ISO 2500 while the GH5 stays at 12 stops. The dual native ISO is really what makes the GH5s the indoor camera to beat.

 

[puredrifting]

In the splitscreen footage that Panasonic showed us, the GH5S had much better color depth and less grain and aliasing than the A7S II, it was no comparison, these were at ISO 3200, 6400 and 12,800.

 

[dcloud]

Exciting new techs. Cant wait what sony will answer with on a7s3. ��

 

[TheDingo]

Hmmm... Why does the Imaging Resource still-photo comparison look WAY different than DPR's studio tests ?

Imaging Resource Comparison : GH5s vs A7SII at 25,600 ISO

I wonder if DPR had the GH5s set to use 2500 native ISO when they did their tests ?

 

[jb.]

Hmmm... Why does the Imaging Resource still-photo comparison look WAY different than DPR's studio tests ?

Imaging Resource Comparison : GH5s vs A7SII at 25,600 ISO

I wonder if DPR had the GH5s set to use 2500 native ISO when they did their tests ?

Ugh. That is not flattering for the GH5s. It's less "noisy" but only because it's smearing quite a bit of detail. There's a whole texture to the bright red swatch of fabric (right about 3 o'clock on the pinwheel of fabric swatches) that's nearly gone on the GH5s side.

Also, what are the green blotchy artifacts between the bottle necks and then the green blotchy stuff on the two rightmost "yellow" bundles of thread above that? That's pretty gross.

 

[Thomas Smet]

Ugh. That is not flattering for the GH5s. It's less "noisy" but only because it's smearing quite a bit of detail. There's a whole texture to the bright red swatch of fabric (right about 3 o'clock on the pinwheel of fabric swatches) that's nearly gone on the GH5s side.

Also, what are the green blotchy artifacts between the bottle necks and then the green blotchy stuff on the two rightmost "yellow" bundles of thread above that? That's pretty gross.

To be fair ISO 25,600 should really be used as an absolute last resort on just about any camera. Its a Hail Mary ISO level when you need to get the shot no matter the quality hit. At more realistic ISO levels the GH5s does exceptionally well compared to the A7S.

 

[jb.]

To be fair ISO 25,600 should really be used as an absolute last resort on just about any camera. Its a Hail Mary ISO level when you need to get the shot no matter the quality hit. At more realistic ISO levels the GH5s does exceptionally well compared to the A7S.

Sure. The weird thing is, I went and looked at the actual IR field test and the 12,800 result is actually worse on that red hankie.

I also disagree with his assessment that the GH5s had better color rendering and detail at either 12,800 or 25,600. And that's coming from a guy who loves the color I get out of my GH5/VLog-L.

But earlier in the test he also puts a still grab of a GH5 (not S) at 12,800 (the one of the dog comparing it to the 1DXII and GH5s) that looks way way way better than any result I've ever gotten on the GH5 at 12,800. I feel like the GH5 starts to get messy around 1600 and by 3200 or 6400 is getting on towards unusable.

So, I'm a little dubious about their testing methodology and conclusions based on my own experience.

 

[TheDingo]

There's a whole texture to the bright red swatch of fabric (right about 3 o'clock on the pinwheel of fabric swatches) that's nearly gone on the GH5s side.

The A7SII image looks more contrasty than the GH5s image does. Looking at the green fabric swatches the A7SII image has non-green color blotches mixed in, which are not present in the GH5s image. I agree about the medium red swatch pattern almost disappearing on the GH5s, and the light pink swatch has lost almost all it's color in the GH5s image. The grey wall behind the items looks much more neutral in the GH5s image, where it has a rainbow mix of colored grain in the A7SII image.

Overall the A7SII image appears to be sharper, but both cameras are showing different color issues with the image they capture, so I would not say that one camera produces better color than the other.

Also, what are the green blotchy artifacts between the bottle necks and then the green blotchy stuff on the two rightmost "yellow" bundles of thread above that? That's pretty gross.

Yes, the GH5s is showing green blotches, where the A7SII is showing rainbow blotches. I don't think either one wins in this comparison.

 

[jb.]

The A7SII image looks more contrasty than the GH5s image does. Looking at the green fabric swatches the A7SII image has non-green color blotches mixed in, which are not present in the GH5s image. I agree about the medium red swatch pattern almost disappearing on the GH5s, and the light pink swatch has lost almost all it's color in the GH5s image. The grey wall behind the items looks much more neutral in the GH5s image, where it has a rainbow mix of colored grain in the A7SII image.

Yeah, that's chroma noise. The GH5s has an insane amount of relatively heavy-handed noise reduction applied to the image while the A7SII's NR seems to be much lighter-touch. That's why the GH5s is losing tons of detail in a lot of places like the red swatch. NR like that does work great on spots like the gray wall where there isn't a lot of detail to destroy, but there is a lot of chroma noise. That's why the GH5s looks much smoother there. If that wall was textured, you'd see the GH5s lose nearly or all of the texture to it's NR algorithms there as well... like it did with the red fabric.

Yes, the GH5s is showing green blotches, where the A7SII is showing rainbow blotches. I don't think either one wins in this comparison.

Yeah. Like I said, the A7SII is showing uncorrected chroma noise. The GH5s green blotches appear to be artifacts from it's NR work on that chroma noise. I'd agree, neither one looks "great" but I do worry a lot more about ugly artifacts than I do chroma noise. Chroma noise is going to smooth out a bit when you go from stills to video... but those artifacts could get really ugly and hard to fix when you move from stills to video. I hope they're just a pre-production glitch.

 

[Matt Gottshalk]

 

[Trek of Joy]

Comparing to other cameras on DP Review test tool the GH5s really does well against pretty much every single APS-C and most FF cameras. Any advantage some FF cameras may have over the GH5s is quickly eliminated once you do add a Speedbooster.

The gh5s is impressive on the video side for reasons mentioned, but its raw stills are behind current aps-c and the 5d3, which is pretty old at this point. I used the "comp" button to downscale larger images to the gh5s' 10 mp count. Its good for m43, but its still not as good as larger sensors. Looks to be a lot of NR at higher ISO's on the video side since raw files are so noisy. But I'm guessing few will buy this for its stills capabilities anyway, so its really no big deal.

 

[Camera Expert]

GH5S interview

 

[Joshua Cadmium]

Lets also not forget the dual native ISO. Not sure if that works on raw stills so that DP Review tool may not be showing us exactly how good the ISO levels are above ISO 2500 when shooting video. Having the same general noise and dynamic range at ISO 2500 vs ISO 400 is pretty huge compared to any other camera out there minus the EVA1 and Varicam. Even the A7S drops to 11 stops of DR at ISO 2500 while the GH5 stays at 12 stops. The dual native ISO is really what makes the GH5s the indoor camera to beat.

The dual native ISO is working on raw stills according to DPReview on the first page of their first impressions: "From a stills point of view, the two circuits are used from ISO 160 - 640 and from ISO 800 and upwards, respectively." - https://www.dpreview.com/reviews/panasonic-lumix-dc-gh5s-first-impressions-review

The article also goes on to say that the GH5s has "sophisticated noise reduction" in video. The camera also has less megapixels that it has to work with, so it can spend more time processing the image than it would if it had more data that needed to be crunched.

This video ( https://youtu.be/IJvcI8Twnlc?t=6m27s ) which was shared earlier, shows a massive difference in noise in video between the GH5 and the GH5s - one that shouldn't be there if the image processing was the same, since the difference in raw is not that different when looking at the DPReview comparison between the two (it's somewhere around half a stop to a stop).

Based on that, it appears that the biggest difference in video is the noise reduction under the hood. If we were able to get raw video out of both the GH5 and GH5s and applied Neat Video after the fact, the difference wouldn't be so great.