Support forum for DG Tools

1.VS mode
core.std.LoadPlugin("C:/Program Files/dgdecnv/x64 Binaries/DGDecodeNV.dll")
src = core.dgdecodenv.DGSource("G:/DGNVindex/XXX/XXX.DVD.dgi",fieldop=2,deinterlace=0,crop_t=0,crop_b=0,crop_l=0,crop_r=0,resize_w=0,resize_h=0,fulldepth=True)

2.AVS mode
core.avs.LoadPlugin("C:/Program Files/dgdecnv/x64 Binaries/DGDecodeNV.dll")
src = core.avs.DGSource("G:/DGNVindex/XXX/XXX.DVD.dgi",fieldop=2,deinterlace=0,crop_t=0,crop_b=0,crop_l=0,crop_r=0,resize_w=0,resize_h=0,fulldepth=True)

The source is a interlaced DVD which has 4 VOB files.I use DGindexNV open these 4 VOB,and save project called "XXX.DVD.dgi".
Then I try to open the dgi with VS mode.But VSedit gives me a error:
"
Error on frame 0 request:
Resize error: Resize error 3074: unrecognized transfer characteristics (0/0/0 => 0/0/0). May need to specify additional colorspace parameters.
"

AVS mode is OK.

Is this a bug?

This is likely a consequence of the recent change populating the metadata for VS.

Are you viewing this in VS Editor? If not, what are you viewing the script in? Somehow VS's zimg is getting involved, even though it is not invoked in your script. What happens if you open the script in VirtualDub2?

Please post output of ClipInfo() in VS version.

Also, please post a link to a piece of one of the VOBs that I can use to duplicate this.

Google "Resize error 3074: unrecognized transfer characteristics" for some relevant info. Still, please provide the requested information so I can determine whether anything is wrong with DGDecNV. Thank you.

admin wrote: ↑

Wed May 02, 2018 10:49 am

This is likely a consequence of the recent change populating the metadata for VS.

Are you viewing this in VS Editor? If not, what are you viewing the script in? Somehow VS's zimg is getting involved, even though it is not invoked in your script. What happens if you open the script in VirtualDub2?

Please post output of ClipInfo() in VS version.

Also, please post a link to a piece of one of the VOBs that I can use to duplicate this.

Google "Resize error 3074: unrecognized transfer characteristics" for some relevant info. Still, please provide the requested information so I can determine whether anything is wrong with DGDecNV. Thank you.

I use VapourSynth Editor r18 to write .vpy files.
In VS version,the .vpy can PIPE or Benchmark(F7),but can't Preview(F5).
I get some information by searching "Resize error 3074: unrecognized transfer characteristics".I think the reason of this problem is DGSource in VS version send a unrecognized(incorrect) info about Transfer(possibly Primary/Matrix).So when Preview,VapourSynth Editor try to use this incorrect info to convert YUV→RGB and fail.Then VapourSynth Editor send a error--"Resize error 3074: unrecognized transfer characteristics" .
If PIPE or Benchmark,there isn't the process of converting YUV→RGB.So it works well.

I did some tests.

1. 2. 3. 4. I guess "SRGB" Matrix is a incorrect info....

I tried another DVD,the problem also existed.And my network is not very fast.So I think you can use any 480i DVD to test...Sorry....

emmm,My mother tongue is not English... Forgive me...

I can reproduce with random dvd source

OK, guys. I know what the problem is and will fix it today. I also want to do something for Guest 2's autocrop issue. Thanks for the report and debugging info.

I have put dgdecnv2053_test.zip in the usual binaries area. Can you please test it with various files of known colorimetry?

* DGI file format now includes a COLORIMETRY line and deletes other appearances of matrix coefficients.
DGI file format is bumped to NV20.

* Correct problem with erroneous reporting of sRGB.

* No longer set a colorimetry based on resolution for streams that do not declare their colorimetry.
Report "Unknown" [2] for such cases.

* Increased the tolerance for autocrop to address Guest 2's stream.

* DGIndexNV user manual updated.

If all is well I'll release this as slipstream 146.

My VapourSynth Editor crashes on refreshing, but idk if that's on my side (I'm using self-built R44-test1 from github) or DGDecoder's fault, so we'll wait for another reports. My workflow is depend on last changes, so I need to use it. And downgrading is sort of hassle, so I'd like to wait for reports if any.
About error. Seems it's gone, at least I can preview, move forward/backward until I refresh the script.

What nVidia driver version and card? 397.31 is known to act like this in many apps.

Can't you collect crash logs or dumps to identify the faulting module?

Reinstalled entire system so yeah, the latest driver from Nvidia. Okay, will downgrade to something order then. I knew that the latest update is evil
About crashdumps... Application is just exiting on refresh, no error messages or so. Where I can retrieve such logs?

If it is not crashing there will be no log. You first said it was crashing. And it doesn't sound like a DGDecNV problem.

First revert to 391.35. If the problem is still there, get support for VS editor, as exiting on refresh is not its expected behavior.

The problem is solved.
Matrix(Primary and Transfer) of DVD which has no matrix(primary and transfer) info is Unknown in ClipInfo() instead of "sRGB".Preview is OK.

I have only BT.709 and BT.601(DVD) source.They all work well now...

Downgraded to 391.35 and it fixed "exit on refresh" issue. But I found another similar problem. with show=True I can load the video, but on scrolling (backward/forward) application is closing on it's own. Is that something on my side or with dgtools?

Also noticed some glitches with show=True

Managed to get some logs with VirtualDub2: https://pastebin.com/BXcYbAp4

With show=True,both Preview(F5) and Benchmark(F7) MAY cause crashing.Sometime it works well.Sometime it crashes.

If it don't crash, some glitches exist.
The source is the bluray disk of Underworld.Extended.Cut.2003.

Thanks, guys. I know what's wrong. Standby for fix...

Please redownload dgdecnv2053_test.zip and update DGDecodeNV.dll; report results. Thank you.

Things for show=True were not initialized properly when show2 was absent.

No more crashes on my side.

The problem is solved.

Thank you, gentlemen. I will release it now.