jaapv Posted January 2 Share #121  Posted January 2 Advertisement (gone after registration) On 12/30/2024 at 5:33 AM, SrMi said: The NR mark (triangle down) does not say anything about how much NR has been applied and whether it has any relevant influence on visible noise. This is how I make my conclusions. The effect of NR is visible by improved DR. If the PDR improves more than linear, then the difference to linearity is the amount that NR contributed. The shadow improvement chart shows the ISO invariance of the latest Sony sensors. If that chart shows that ISO invariance ceased at certain ISOs, that can be attributed to NR. As long as the graph remains invariant (flat), the baked in NR has no effect. A good test would be to see images that include the full chain, including lifting shadows and applying relevant NR in post. Judging a camera's output solely on measurement graphs seems wrong to me. If you look carefully there is a slight deviation at the lower end of the chart. Applying NR in post is somewhat doubtful as a judging point, as there are AI components in quite a some NR methods, including LR. 1 Link to post Share on other sites More sharing options...
Advertisement Posted January 2 Posted January 2 Hi jaapv, Take a look here Trading SL2-S for SL3. I'm sure you'll find what you were looking for!
hdmesa Posted January 2 Share #122  Posted January 2 (edited) 6 hours ago, jaapv said: If you look carefully there is a slight deviation at the lower end of the chart.... The abrupt change at the low end of the ISO range is from the dual gain sensor technology. From P2P site "Notes" on the shadow improvement chart: Dips at intermediate ISO values result from two-stage amplification. It's likely the jump above 12,800 is similar and has nothing to do with the strength of the NR being applied. Also from P2P: Changes at the "Hi" ISO boundary are common. Sometimes up and sometimes down. Sharp changes are a mystery at this time. 6 hours ago, jaapv said: ...Applying NR in post is somewhat doubtful as a judging point, as there are AI components in quite a some NR methods, including LR. AI for NR is not used in the regular noise panel in LR. AI is only used with the Denoise function. The baked in NR in cameras is likely very similar to standard NR in LR/C1. Edited January 2 by hdmesa Link to post Share on other sites More sharing options...
jaapv Posted January 2 Share #123  Posted January 2 Denoise is a LR function, isn't it? The same inn ACR, not in the legacy panel, but in the newer Enhance section. Link to post Share on other sites More sharing options...
hdmesa Posted January 3 Share #124  Posted January 3 (edited) 8 hours ago, jaapv said: Denoise is a LR function, isn't it? The same inn ACR, not in the legacy panel, but in the newer Enhance section. Yes, in LRC you go to the Photo menu, choose Enhance, then in the pop-up dialog box is the Denoise function. You can also access it through a button in the Noise Reduction panel, but it's clearly labeled as AI and separated from the manual noise reduction controls. Welcome, dear visitor! As registered member you'd see an image here… Simply register for free here – We are always happy to welcome new members! It's these manual controls I recommend applying to high ISO files from the SL2-S if you want parity with the SL3 at ISO 8000 and above. How much and at what ISO is the best guess of the tester – I would say apply what looks good without it looking obvious that NR was applied, then continue to judge and apply the NR as needed at each ISO jump thereafter, also taking into account what the SL3 file looks like. Might also be worth noting that the default manual NR amounts in LRC seem adaptive to not just the ISO but the amount of noise detected in an image. For example, I have two images of the same scene both shot at ISO 1250 file but at two different exposures – the darker image with less detail has zero default NR applied by LRC, and the other image that was exposed more brightly which reveals more noise in the shadow detail will have a default value of say 10. Capture One isn't like this, it applies the same strength of NR to M11, SL3, and SL2-S files regardless of ISO or the noise level of a given image. So if one is testing SL3 versus SL2-S at high ISO by processing them in LRC, the SL3 file may have more default NR applied. Since that could be the case, the files from both cameras being compared need to either be equalized to the higher value or both reduced to zero. Then after that, if you're at or over ISO 8000, I think it's fair to apply a conservative amount of NR to the SL2-S file to compete with the SL3's baked in NR as shown by P2P charts.  Edited January 3 by hdmesa Link to post Share on other sites Simply register for free here – We are always happy to welcome new members! It's these manual controls I recommend applying to high ISO files from the SL2-S if you want parity with the SL3 at ISO 8000 and above. How much and at what ISO is the best guess of the tester – I would say apply what looks good without it looking obvious that NR was applied, then continue to judge and apply the NR as needed at each ISO jump thereafter, also taking into account what the SL3 file looks like. Might also be worth noting that the default manual NR amounts in LRC seem adaptive to not just the ISO but the amount of noise detected in an image. For example, I have two images of the same scene both shot at ISO 1250 file but at two different exposures – the darker image with less detail has zero default NR applied by LRC, and the other image that was exposed more brightly which reveals more noise in the shadow detail will have a default value of say 10. Capture One isn't like this, it applies the same strength of NR to M11, SL3, and SL2-S files regardless of ISO or the noise level of a given image. So if one is testing SL3 versus SL2-S at high ISO by processing them in LRC, the SL3 file may have more default NR applied. Since that could be the case, the files from both cameras being compared need to either be equalized to the higher value or both reduced to zero. Then after that, if you're at or over ISO 8000, I think it's fair to apply a conservative amount of NR to the SL2-S file to compete with the SL3's baked in NR as shown by P2P charts.  ' data-webShareUrl='https://www.l-camera-forum.com/topic/390831-trading-sl2-s-for-sl3/?do=findComment&comment=5733120'>More sharing options...
Recommended Posts
Create an account or sign in to comment
You need to be a member in order to leave a comment
Create an account
Sign up for a new account in our community. It's easy!
Register a new accountSign in
Already have an account? Sign in here.
Sign In Now