Sorry, I meant DGHDRtoSDR.dll v.DGHDRtoSDR_1.10.rar
As to
@dennis
Are you going to ignore my request for your source file, etc.?
I'm missing something.
Nevertheless, my source is Gladiator 2000 Extended 2160p UHD BluRay REMUX HDR HEVC DTS-X-EPSiLON.
Do you want a 10/20 sec cut from it or what?
Please give me the cut that you made, you know, the one you say has 240 frames but for which DGSource() says 236.
You say DGHDRtoSDR is not working properly, so please give me the source file (if different from the one above), and the exact Avisynth script that you used. I expect the issue is related to your other tools and process, as DGHDRtoSDR cannot create non-existing frames. It simply returns frames requested by the application that is processing the Avisynth script.
Thank you for the source file. The index file sees 238 frame starts but playing the script in Vdub2 delivers only 237 (0-236). That is likely due to the way you have cut the file, i.e., not a clean cut at the end. So I don't see any problem. I won't speculate on what your avs2yuv64 and 16-bit hack stuff may be doing, and I do not provide support for them.
You haven't shown any issue with DGHDRtoSDR. Your ConvertBits(8) call is not needed as DGHDRtoSDR delivers 8-bit by default. Also, if you resize in DGSource() it will run much faster. Or resize your way but before the DGHDRtoSDR call.
would it be possible to consider using .zip or something ?
would it be possible to keep older versions of stuff online for a month or so to provide buffer time for vsrepo to catch up (I haven't looked into vsrepo's mechanism for keeping up to date with new versions, will do so shortly).
If a product is non-licensed, maybe could github and tagged releases be a possible alternative, since vsrepo seems to use those OK ?
Thank you for the source file. The index file sees 238 frame starts but playing the script in Vdub2 delivers only 237 (0-236). That is likely due to the way you have cut the file, i.e., not a clean cut at the end. So I don't see any problem. I won't speculate on what your avs2yuv64 and 16-bit hack stuff may be doing, and I do not provide support for them.
You haven't shown any issue with DGHDRtoSDR. Your ConvertBits(8) call is not needed as DGHDRtoSDR delivers 8-bit by default. Also, if you resize in DGSource() it will run much faster. Or resize your way but before the DGHDRtoSDR call.
Bottom line: no problems found.
Thank you very much for your help.
I would like to point out the John Hable's post at http://filmicworlds.com/blog/filmic-ton ... er-curves/.
His approach is to give more and easy control over the Filmic Tonemapping curves.
He says:
There are several specific issues I’m hoping to address from the Uncharted 2 curve:
(implemented in Hable tonemap operator)
Simple intuitive controls:
Direct control over dynamic range:
Well behaved curves:
Controls in engine:
Fast, closed form: The curve should be simple and fast to evaluate.
Lerpable parameters:Simple inverse:
Convolve with output gamma:
Code on Github: Makes integration simple.
Anyone working on HDR to SDR already is familiar with the Hable stuff. It's all subjective, and personally I have no interest in a "filmic look" or "crisper blacks".
hue=10.0 Hue adjustment used to correct hue shift due to the gamut mapping.
Values above 0.0 shift toward red. Values below 0.0 shift toward green. The
default value is 10.0.
Shouldn't it default to 0? Just wondering this change is all.
The gamut mapping was creating a small noticable green bias for some streams; the default hue setting corrects for that. You can set it to 0 if you prefer.
The gamut mapping was creating a small noticable green bias for some streams; the default hue setting corrects for that. You can set it to 0 if you prefer.
Alright. Yes I do prefer zero.
Seems like the versions before 1.11 has "0" as the default even tho there is no hue option.. Just have to remember to add the hue line to scripts now.
I slipstreamed 1.11 to reduce default hue to 5. Hopefully, you can live with that. I do think this hue correction is a good thing. If you can show otherwise, please do.
It's an Avisynth+ thing that you can google. There are so many variables affecting it that your best bet is to use Avsmeter to time your script and empirically determine the best value. For some scripts, any prefetch reduces performance.
All versions of DGDecodeNV, for the last great while at least, have the fulldepth instruction
From the manual
fulldepth: true/false (default: false)
When fulldepth=true and the encoded video is HEVC 10-bit or 12-bit, then DGSource() delivers 16-bit data to Avisynth with the unused lower bits zeroed. The reported pixel format is CS_YUV420P16. If either of the two conditions are not met, then DGSource() delivers 8-bit YV12 or I420 data, as determined by the i420 parameter. When fulldepth=false and the video is HEVC 10-bit or 12-bit, then the video is dithered down to 8-bit for delivery. If you need a reduced color space (less than 16 bits) for your high-bit-depth processing, you can use ConvertBits() as needed after your DGSource() call.
Have you ever considered making an HDR to SDR app for the Nvidia Shield? It has an Nvidia GPU in it and one thing it lacks is an HDR to SDR app that can hook onto Plex or Kodi or some other popular video apps.
Could even charge for it an probably make some decent bank.