Originally Posted by desertdome
There are no "embedded HDR luminance levels" when you send an SDR2020 signal to whatever is next in the signal chain. The correct setting for this is HDR off and Color Space set to BT.2020. I've been to Sony headquarters in San Diego for HDR training with Sony engineers. While Sony prefers to use their own tone mapping, they have also enabled one to use upstream tone mapping and send SDR2020 to the 5000ES. It does not remap anything when using the correct settings.
Sony says in their manual, "If the setting is not correct for the input content, the bright and dark areas of the video may appear too bright or too dark." This happens if you leave HDR to Auto, but send it SDR2020 content. Perhaps this is what you see as the problem. I think the HDR mode is triggered by the detected color space. One is either going to let the Sony do the tone mapping and send it HDR and BT.2020 or let the Lumagen (or madVR) do the tone mapping and set HDR to Off. There really isn't a use case when one wants to have the Lumagen tone map some of the time and have the Sony tone map some of the time and have the HDR Auto setting detect the HDR metadata. I just sent the 5000ES back to Sony today so I can't check into this further.
I've tried HDR off with a tone-mapped SDR2020 signal from the Radiance and it does not change the presentation of the image: Auto
for HDR are exactly the same. So this is not a "correct" setting in my experience (and the experience of others here who have compared both approaches)
As much as I theoretically agree with you that that hybrid approach doesn't make "sense," it is in actuality the better approach for Radiance and 5000 users as far as I can ascertain in my content testing to date...and after conversations with Sony, and other comments posted here about limitations with the current processing in the 5000. It seems to assume an SDR flag means "that's the color volume" it should expect in the incoming signal; yet the tone-mapped signal of course contains a larger color volume (and peak whites) - greater than expected for SDR (29 fl/100 nits?). According to Kris, other devices, such as the JVC 4500, could care less how the flag is set in the incoming signal, and processes/displays whatever is presented within the limitations of the device. Apparently the Sony cares for some reason.
Everything I understand is that color output is impacted by both 2020 (P3) as well as the higher luminance values of HDR. So if the 5000 is not processing this correctly, brightness and color range would be reduced, which is pretty much what I see.
I mean, take the Radiance out of the chain - an HDR signal from your UHD player is brighter and more vibrant on the 5000 then sending tone-mapped SDR-2020 from the Radiance to it. It just is.
I'm ALWAYS open to re-looking at a wrong setting, or just learning more - but again, a few others are also trying this hybrid approach and are much more satisfied with the results. Again, this appears not to be a Radiance issue, but how Sony is processing higher luminance values via a SDR2020 signal.