[RESOLVED] not able to feed the encoder

Support forum for DGDecNV
Post Reply
mparade
Posts: 13
Joined: Mon Oct 13, 2014 7:45 am

[RESOLVED] not able to feed the encoder

Post by mparade » Sat Sep 09, 2017 9:53 am

Hello there,

I would like to feed x265 encoder using DGDecNV via Vapoursynth.
The source is HDR10. My card is: Nvidia GTX 1060 6GB with latest driver package installed.
DGDecNV version is from 08.2017.
The error code I always got includes:
Script evaluation failed:
Python exception: name 'true' is not defined
Traceback (most recent call last):
  File "src\cython\vapoursynth.pyx", line 1821, in vapoursynth.vpy_evaluateScript (src\cython\vapoursynth.c:36464)
  File "C:\TEMP\encodingTempSynthSkript_15_59_51_0810.vpy", line 7, in <module>
    clip = core.avs.DGSource("C:/TEMP/15_59_51_0810.dgi", fulldepth=true)
NameError: name 'true' is not defined
Can you help on this? Please let me know if you need some additional info to support.

Thank you very much.

User avatar
admin
Site Admin
Posts: 3885
Joined: Thu Sep 09, 2010 3:08 pm

Re: not able to feed the encoder

Post by admin » Sat Sep 09, 2017 11:21 am

Try True instead of true. If that doesn't work then you'll have to ask the Vapoursynth guys.

User avatar
hydra3333
Distinguished Aussie Member
Distinguished Aussie Member
Posts: 115
Joined: Wed Oct 06, 2010 3:34 am
Contact:

Re: not able to feed the encoder

Post by hydra3333 » Sat Sep 09, 2017 9:30 pm

Hey mparade, when you find an answer could you post it here ?

mparade
Posts: 13
Joined: Mon Oct 13, 2014 7:45 am

Re: not able to feed the encoder

Post by mparade » Sun Sep 10, 2017 3:50 am

Still no success.

Using True instead of true I got an error message from the encoder:
x265 [error]: unable to open input file
Programmer said: "Guessing that 'True' wasn't the right way to go.
Seems like this isn't possible atm."

Sorry for asking it but before asking Vapoursynth can we exclude a bug in DGDecNV (I am using DGDecNV since 2011 without any problem)?

Thank you for your help!

User avatar
admin
Site Admin
Posts: 3885
Joined: Thu Sep 09, 2010 3:08 pm

Re: not able to feed the encoder

Post by admin » Sun Sep 10, 2017 6:08 am

True is the correct constant in Python. So that is not the issue.

There is no known bug in DGDecNV. fulldepth works fine with Avisynth+.

Ask the Vapoursynth guys about it.

User avatar
admin
Site Admin
Posts: 3885
Joined: Thu Sep 09, 2010 3:08 pm

Re: not able to feed the encoder

Post by admin » Sun Sep 10, 2017 11:43 am

I'm following your thread at another forum but I do not post there, so I give some information here.

This Avisynth+ script works fine when opening in VirtualDub 64:

loadplugin("d:\don\Programming\C++\dgdecnv\DGDecodeNV\x64\Release\dgdecodenv.dll")
dgsource("H:\Streams\HEVC\10 Bit\ES\Philips_Ultra-HD_Beach_Life_supershop-demo-3.dgi",fulldepth=true)
convertbits(8)

If you have an application that accepts 16 bit you can omit the convertbits call. Or you can change it to convertbits(10) if your application accepts 10 bit.

It looks to me like Vapoursynth simply does not support HDR from Avisynth source filters. Or maybe it doesn't support 16 bit and you need to convert it to 10. Ask the author about that.

User avatar
admin
Site Admin
Posts: 3885
Joined: Thu Sep 09, 2010 3:08 pm

Re: not able to feed the encoder

Post by admin » Mon Sep 11, 2017 12:10 pm

Any news, mparade?

mparade
Posts: 13
Joined: Mon Oct 13, 2014 7:45 am

Re: not able to feed the encoder

Post by mparade » Tue Sep 12, 2017 4:06 pm

It seems the guys at Vapoursynth cannot/don't want to solve this issue (got no usable reply yet).

gonca
Distinguished Member
Distinguished Member
Posts: 563
Joined: Sun Apr 08, 2012 6:12 pm

Re: not able to feed the encoder

Post by gonca » Tue Sep 12, 2017 5:31 pm

Looks like they have no intention to solve the bug in Vapoursynth

User avatar
admin
Site Admin
Posts: 3885
Joined: Thu Sep 09, 2010 3:08 pm

Re: not able to feed the encoder

Post by admin » Tue Sep 12, 2017 7:57 pm

A couple points...

The Vapoursynth author could reproduce it with a different Avisynth source filter that delivers 16-bit (vi.pixel_type = VideoInfo::CS_YUV420P16). Or maybe even BlankClip() can generate that color space. That might be something for you to pursue, mparade.

DGDecodeNV(fulldepth=true) demonstrably works fine with Avisynth+.

Loading my Avisynth+ script into VirtualDub without the convertbits(8) call gives the expected error popup about an unsupported color space. VirtualDub does not crash. Conversely, stax76 reports that the Vapoursynth script crashes VirtualDub. Draw your own conclusions.

Marking resolved as no problem in DGDecodeNV has been established. But feel free to post again if there is pertinent new info.

Post Reply