GH5 Color difference between GH5 and GH5s (image)

joe12south

Veteran
difference.png


Closer match than GH4 > GH5. Most difference in blues and greens, least in oranges and reds. There appears to be a small exposure difference, as well, but until I can get a camera for myself I can't be sure how much. Looks like it will be fairly straightforward to make a LUT for matching.
 
Cool test. Thanks.

Have there been any reviewers of the GH5s been able to compare the dynamic range and noise performance of the S version compared to the regular GH5 at LOW ISO's?

Being able to go higher in ISO can be handy sometimes, but I mainly shoot between ISO 100 and 800 anyway so I am interested to hear if there has been any improvement IQ wise in those ranges. Is there fewer noise because of the larger pixels? A bit more DR? Or are they virtually the same there?
 
Cool test. Thanks.

Have there been any reviewers of the GH5s been able to compare the dynamic range and noise performance of the S version compared to the regular GH5 at LOW ISO's?

Being able to go higher in ISO can be handy sometimes, but I mainly shoot between ISO 100 and 800 anyway so I am interested to hear if there has been any improvement IQ wise in those ranges. Is there fewer noise because of the larger pixels? A bit more DR? Or are they virtually the same there?

Wander around YouTube and you can find several tests. The GH5 has noise in the shadows at every ISO that appear lessened in the GH5s. Color accuracy and dynamic range start to drop after ISO 400 on the GH5, noticably so by ISO 1600. The GH5s seems to hold onto a comparable amount up until 6400.
 
It holds the colours much better to a much higher ISO, that's for sure. But in terms of colorimetry of the sensor, it's identical.

GH5S_HLG_Sample_ISO2500_1.3.2.png


Here's a quick test I did with the HLG profile and my latest beta LUT (developed with the original GH5) which confirms it.

Cheers,

Paul :)
 
What ISO(s) do you test at?

FYI, FilmConvert tested and released a new profile ... but I've never had any luck with their profiles matching, so I have no idea what they actually do. ;-)
 
Paul, does this mean we will not need a specialized GH5s LUT?

Also paul, after reading the DPreview on GH5s, I learned a bit about Vlog L and the info it throws away. I understand this will make it easy to match with my EVA1. On the other hand, it sounds like shooting in HLG will provide more info and safer exposures?
 
This is not scientific, just a perception, not a comparison because I haven't made one, but I thought I observed with HLG an odd color shift to tungsten light reflected off of a dark gray wall, which gave me pause, not wanting to accept a risk of that. Then, I read the DPReview which remarked about a similar possibility, that VLog-L might have a more specific correlation to GH5s color than HLG. I was using the regular GH5, but I wondered about the same thing, a more direct alignment of color to GH5 with VLog-L than HLG which places more emphasis on scene gamma and a non-specific color gamut.

Like everyone, I'd like to use all the code values available within 10 bits than the truncated VLog-L permits, yet if there is a tradeoff in color accuracy, well that risk I may not want to accept.
 
Paul, does this mean we will not need a specialized GH5s LUT?

Also paul, after reading the DPreview on GH5s, I learned a bit about Vlog L and the info it throws away. I understand this will make it easy to match with my EVA1. On the other hand, it sounds like shooting in HLG will provide more info and safer exposures?

Correct :) I'm using my regular HLG LUT on the GH5S with perfect accuracy :)

V-LogL works just as well LUT-wise, and I've tested it with EVA1 footage and again it works perfectly. If I was going to match between cameras I'd shoot both V-LogL just to make post life more consistent.
 
I'm seeing a lot of confusion and misinformation about HLG vs V-LOG. It seems like a good number of people who don't know how to work with log footage hope that HLG is some magic bullet. A few notes that might be helpful:

- Both are capable of storing all of the dynamic range the GH5(s) can capture, but they bias the signal differently. V-LOG is a more balanced curve. The available bits are more evenly distributed than HLG, which compresses extra tonal range into the top 50% of the signal. HLG shadows have more precision, but less range.

- HLG does not use the entire bit bucket...it is a delivery standard that uses video levels. it does use more of it than V-LOG. In 10-bit, this amounts to a minor difference given a 12 stop max range. (HLG won't even work at 8-bit, 10-bit is part of the standard.)

- HLG adheres the the rec2020 color space, V-LOG, uses V-gamut which is inconsequentially bigger.

- Some people find HLG more approachable since it looks more like a "normal" picture, but that doesn't make it technically better. You are more likely to not create a giant mess adjusting HLG footage without a LUT. If all you are looking for is some highlight protection, HLG is easier to understand.

- You need to expose HLG differently than V-LOG. Their midpoint is different (see the point above about their curves.) You should not compare the two at the same exposure. HLG does not provide a "safer" better exposure, just different.

- Using ETTR techniques with HLG is riskier than with V-LOG or rec709 profiles. You're forcing more of the image into the range that is more compressed. (Again, see point about curve.)

- V-LOG is not noisier than HLG, it simply resolves the noise present in the shadows better since the shadows are not as "crushed" and are recorded at a level that is not compressed as much (compression acts like a poor man's noise reduction.)

In short, V-LOG is a "scene" profile that best balances the trade-offs required to capture as much of the scene information as possible. It MUST be transformed to an appropriate "display" profile. HLG is a display profile which balances the trade-offs required to show an HDR image on an SDR display. It only requires a minimal transformation to be viewable, but it is not as editable.
 
Back
Top