DGMPGDec 1.2.0 RC5 - Page 2

Guest · 7th February 2005, 05:35

It'll come back when I get the bugs out. I don't like to release buggy stuff if I can avoid it.

Having said that, why do you think the quality is better with demuxed video compared to frame-served video? I can't think of any reason for that to be true.

clone83 · 7th February 2005, 05:42

im not sure why it is like that but it is....

Guest · 7th February 2005, 05:56

Well, can you at least tell me what the two processes are that cause the noticeable difference.

clone83 · 7th February 2005, 06:00

all i do is demux the vid and then work on it in vdm with some filters and then i mux the audio.

p.s im always interested in new way of encoding if u have any suggestions....? and do u have msn cause i would like to talk to you in private about some deals that i could possibly offer you....?

Guest · 7th February 2005, 06:10

There's no difference between demuxing and opening in VDubMod versus using DGDecode to serve via the D2V file. Except that VDubMod does not honor pulldown flags and so is technically an incorrect decoder. And of course DGDecode allows you to use Avisynth scripting before sending the video to VDub. You may be getting the "poor man's IVTC" effect.

Please describe the quality difference that you observe.

You can PM me with any deals you have in mind. Thank you.

Cyberia · 7th February 2005, 06:11

Quote:

Originally posted by neuron2
Thanks for the update. I'll sleep better tonight. I pulled a 36-hour straight programming session to make RC2 and I still haven't slept yet after 48 hours.

Oh My God! Thats dedication. Much appreciated.

Let's all give neuron2 a Thank You!

clone83 · 7th February 2005, 07:48

Thank You Neuron2 and i will get around to sending u a pm asap....

Leak · 7th February 2005, 08:02

Quote:

Originally posted by neuron2
Thanks for the update. I'll sleep better tonight. I pulled a 36-hour straight programming session to make RC2 and I still haven't slept yet after 48 hours.

Dude! Off to bed! NOW!

If I tried to pull that off, I could stay at the hospital instead of working there...

alfixdvd · 7th February 2005, 08:10

Excuse me neuron2, I want to say :" I can't play any movies from avs scripts, if the movie is Bottom Field First.

This night, I send you two mpeg BFF.

best regards
alfixdvd

Backwoods · 7th February 2005, 10:16

Quote:

Originally posted by neuron2
Tell me the DGIndex version and put the file on my ftp server. That's the only way I can help you. Try setting both the PIDs to 2 before loading the file.

Both were set to 2 and it still would freeze. I'm using DGIndex 1.2.0 RC2.

If I run the M2T through HDTV2MPEG the resulting MPG2 file saves and demuxes fine in DGIndex 1.2.0 RC2. I'll upload a snippet of the M2T for you tomorrow.

len0x · 7th February 2005, 12:31

Wow, D2V structure changed drastically! Quick question: previously 2,3,12,13 values meant TFF, what's the equivalent now?

Guest · 7th February 2005, 14:57

@alfixdvd & Backwoods

Standing by for the failing streams...

@len0x

Please see the D2V format document included in the zip file. If the last digit is in this set { 2, 3, 6, 7, a, b, e, f }, the tff bit is set.

In the next release I'll push the frame type to the upper nibble. Then the old 0,1,2,3 will still be visible. The end marker will then be ff.

tritical · 7th February 2005, 16:07

Reading the D2V_format text I am confused about the following statement:

Quote:

One important issue still to be addressed is that upsampling based on this flag should be done *before* pulldown is applied. Currently it appears that it is not being done that way.

Assuming this is referring to dgdecode, unless something has changed drastically, the upsampling is done before pulldown. It is done right after post-processing and before showQ, all of which are done on the decoded pictures before they are copied out to the other storage buffers from which pulldown is performed using the copyeven/copyodd functions. Or am I missing something?

loni_blues · 7th February 2005, 16:51

@neuron2

Thanks for the continuous and hard work!
And sorry but, with RC2 I cannot open the d2v file in Gordian Knot, while it still can be done with the d2v made with RC1. The d2v file seems complete, though. Is it a bug in RC2?
Regards.

len0x · 7th February 2005, 16:54

Quote:

Originally posted by neuron2
In the next release I'll push the frame type to the upper nibble. Then the old 0,1,2,3 will still be visible. The end marker will then be ff.

By next version you mean another RC of 1.2.0 or 1.3?

len0x · 7th February 2005, 16:56

Quote:

Originally posted by loni_blues

And sorry but, with RC2 I cannot open the d2v file in Gordian Knot, while it still can be done with the d2v made with RC1. The d2v file seems complete, though. Is it a bug in RC2?

I suggest that you usually ask about such problems in GK forum, where you can also find a new version of GK bundled with 1.2.0RC2.

alfixdvd · 7th February 2005, 16:59

Hi neuron2.

I will not to send you any file, because with version 1.2.O RC2 all works fine.

But with versions 1.2.0.b5 to 1.2.0.RC1, I can't play any movies from avs scripts, if the movie is Bottom Field First.

What's wrong ?

best regards.
alfixdvd

Guest · 7th February 2005, 17:23

Quote:

Originally posted by len0x
By next version you mean another RC of 1.2.0 or 1.3?

I mean the very next thing I put out, 1.2.0 RC3 or final. I already have it coded and tested.

Guest · 7th February 2005, 17:26

Quote:

Originally posted by tritical
Reading the D2V_format text I am confused about the following statement:

Assuming this is referring to dgdecode, unless something has changed drastically, the upsampling is done before pulldown. It is done right after post-processing and before showQ, all of which are done on the decoded pictures before they are copied out to the other storage buffers from which pulldown is performed using the copyeven/copyodd functions. Or am I missing something?

You're right. I didn't actually look at the code.

BTW, all of the contortions you went through in making the Info screen (for re-ordering and random access) were completely finessed by letting DGIndex put flags in the D2V. When he indexes he plays through linearly and so it is much easier for him to do all that.

Guest · 7th February 2005, 17:27

Quote:

Originally posted by alfixdvd
Hi neuron2.

I will not to send you any file, because with version 1.2.O RC2 all works fine.

But with versions 1.2.0.b5 to 1.2.0.RC1, I can't play any movies from avs scripts, if the movie is Bottom Field First.

I wrote earlier that the Second_Field variable was not initialized. Glad RC2 is working for you!

7th February 2005, 06:10	#25 \| Link
Guest Guest Join Date: Jan 2002 Posts: 21,901	There's no difference between demuxing and opening in VDubMod versus using DGDecode to serve via the D2V file. Except that VDubMod does not honor pulldown flags and so is technically an incorrect decoder. And of course DGDecode allows you to use Avisynth scripting before sending the video to VDub. You may be getting the "poor man's IVTC" effect. Please describe the quality difference that you observe. You can PM me with any deals you have in mind. Thank you. Last edited by Guest; 7th February 2005 at 06:12.

7th February 2005, 12:31	#31 \| Link
len0x I'm afraid we've to stop Join Date: Mar 2003 Location: Amongst mad people Posts: 5,398	Wow, D2V structure changed drastically! Quick question: previously 2,3,12,13 values meant TFF, what's the equivalent now? __________________ Gordian Knot Family: Gordian Knot: website, download Auto Gordian Knot: Website and download, tutorial, FAQ

7th February 2005, 14:57	#32 \| Link
Guest Guest Join Date: Jan 2002 Posts: 21,901	@alfixdvd & Backwoods Standing by for the failing streams... @len0x Please see the D2V format document included in the zip file. If the last digit is in this set { 2, 3, 6, 7, a, b, e, f }, the tff bit is set. In the next release I'll push the frame type to the upper nibble. Then the old 0,1,2,3 will still be visible. The end marker will then be ff. Last edited by Guest; 7th February 2005 at 15:23.

7th February 2005, 16:59	#37 \| Link
alfixdvd Registered User Join Date: Mar 2004 Posts: 243	Hi neuron2. I will not to send you any file, because with version 1.2.O RC2 all works fine. But with versions 1.2.0.b5 to 1.2.0.RC1, I can't play any movies from avs scripts, if the movie is Bottom Field First. What's wrong ? best regards. alfixdvd Last edited by alfixdvd; 7th February 2005 at 17:01.

7th February 2005, 05:35	#21 \| Link
Guest Guest Join Date: Jan 2002 Posts: 21,901	It'll come back when I get the bugs out. I don't like to release buggy stuff if I can avoid it. Having said that, why do you think the quality is better with demuxed video compared to frame-served video? I can't think of any reason for that to be true.

7th February 2005, 05:42	#22 \| Link
clone83 Registered User Join Date: Jun 2004 Posts: 100	im not sure why it is like that but it is....

7th February 2005, 05:56	#23 \| Link
Guest Guest Join Date: Jan 2002 Posts: 21,901	Well, can you at least tell me what the two processes are that cause the noticeable difference.

7th February 2005, 06:00	#24 \| Link
clone83 Registered User Join Date: Jun 2004 Posts: 100	all i do is demux the vid and then work on it in vdm with some filters and then i mux the audio. p.s im always interested in new way of encoding if u have any suggestions....? and do u have msn cause i would like to talk to you in private about some deals that i could possibly offer you....?

7th February 2005, 07:48	#27 \| Link
clone83 Registered User Join Date: Jun 2004 Posts: 100	Thank You Neuron2 and i will get around to sending u a pm asap....

7th February 2005, 08:10	#29 \| Link
alfixdvd Registered User Join Date: Mar 2004 Posts: 243	Excuse me neuron2, I want to say :" I can't play any movies from avs scripts, if the movie is Bottom Field First. This night, I send you two mpeg BFF. best regards alfixdvd

7th February 2005, 16:51	#34 \| Link
loni_blues Registered User Join Date: Jan 2003 Location: Argentina Posts: 125	@neuron2 Thanks for the continuous and hard work! And sorry but, with RC2 I cannot open the d2v file in Gordian Knot, while it still can be done with the d2v made with RC1. The d2v file seems complete, though. Is it a bug in RC2? Regards.