MPEG4 Modifier v1.3.4 Released (Updated 2005-Dec-14) - Page 17

Moitah · 13th December 2005, 17:50

Quote:

Originally Posted by eXistenZ_69

How do I identify real N-VOPs?

You will get an error message in 1.3.3. I removed the code because I don't know of any encoders that write real N-VOPs when packed bitstream is used. If someone finds an encoder that does this, I re-add support for it.

SeeMoreDigital · 13th December 2005, 17:55

Hi Moitah,

Are you going to update this forum thread title?

Cheers

Moitah · 13th December 2005, 18:11

Quote:

Originally Posted by SeeMoreDigital

Hi Moitah,

Are you going to update this forum thread title?

Cheers

I tried to a couple times, it won't completely change:

SeeMoreDigital · 13th December 2005, 19:01

How odd...

That said, I've noticed the same, when trying to amend some old posts in the Hardware Players section of the forum too!

I ended up having to make up a short title first, save it. Open it again and then write the title I actually wanted, second time around

Cheers

Moitah · 13th December 2005, 20:09

Quote:

Originally Posted by eXistenZ_69

@Moitah

Thank you so much for the MPEG4Modifier update as well as the
FindBadUnpack tool! You're awesome

Will, start using both today...

You're welcome. Can you tell me how many bad files you find? I'm just curious.

bond · 14th December 2005, 00:03

Quote:

Originally Posted by Moitah

If anyone does find a file with this, please let me know.

no need to worry, there is no encoder using packed bitstream plus real n-vops

still this doesnt mean syskin will not change that in xvid somewhen

Moitah · 14th December 2005, 00:20

Quote:

Originally Posted by bond

no need to worry, there is no encoder using packed bitstream plus real n-vops

still this doesnt mean syskin will not change that in xvid somewhen

Oh, good. I don't mind adding it back if that ever happens, but it was in my way for that last bugfix, and I didn't feel like messing with it if the code isn't even being used.

eXistenZ_69 · 14th December 2005, 16:08

Been experimenting with the new MPEG4Modifier 1.3.3 and just found a packed bitstream encode with real N-VOPs (at least the error message says so)

Quote:

55178: I-VOP (0:38:13.000)
55179: P-VOP (0:38:13.125)
B-VOP (0:38:13.041)
55180: B-VOP (0:38:13.083)
55181: N-VOP(D) (0:38:13.125)
55182: P-VOP (0:38:13.250)
B-VOP (0:38:13.166)
55183: B-VOP (0:38:13.208)
55184: N-VOP(D) (0:38:13.250)
55185: N-VOP (0:38:13.291)
55186: I-VOP (0:38:13.333)

Moitah · 14th December 2005, 16:38

Quote:

Originally Posted by eXistenZ_69

Been experimenting with the new MPEG4Modifier 1.3.3 and just found a packed bitstream encode with real N-VOPs (at least the error message says so)

Wow, I wasn't expecting that. I wonder which encoder produced this, can you tell me what the "user data" is?

Moitah · 14th December 2005, 17:43

After checking some files I had, I found one also:

Code:

  8972: P-VOP    (0:06:13.875)
        B-VOP    (0:06:13.792)
  8973: B-VOP    (0:06:13.834)
  8974: N-VOP(D) (0:06:13.875)
  8975: N-VOP    (0:06:13.917)
  8976: I-VOP    (0:06:13.959)

The user data is marked XviD0038. So it looks like I have some more work to do.

bond · 14th December 2005, 18:25

Quote:

Originally Posted by Moitah

The user data is marked XviD0038. So it looks like I have some more work to do.

is that an old version of xvid?

there was once a bug in xvid which made using the frame drop ratio option (placing real n-vops) possible together with b-frames, but it was creating borked streams, which caused syskin to fix it by simply disabling the framedropratio settings when using b-frames

so i wouldnt assume that such streams make much sense

Moitah · 14th December 2005, 18:34

Quote:

Originally Posted by bond

is that an old version of xvid?

there was once a bug in xvid which made using the frame drop ratio option (placing real n-vops) possible together with b-frames, but it was creating borked streams, which caused syskin to fix it by simply disabling the framedropratio settings when using b-frames

so i wouldnt assume that such streams make much sense

I think it's one of the 1.1 betas. I asked sysKin on IRC and he said 1.1 might insert them even when B-VOPs are used. I am doing some encoding right now to see if I can reproduce this.

celtic_druid · 14th December 2005, 18:40

1.1beta1 throught till but not including beta2 or xvid.h v1.48 and 1.49.

I would still have all my old builds from that period (16 Jan 2005 -->04 Apr 2005) if anyone wants them.

Six builds including 1.1beta1.

eXistenZ_69 · 14th December 2005, 19:03

Quote:

Originally Posted by Moitah

Wow, I wasn't expecting that. I wonder which encoder produced this, can you tell me what the "user data" is?

User Data: DivX999b000p, XviD0038

Packed bitstream: Yes
QPel: No
GMC: No
Interlaced: No
Aspect ratio: Square pixels
Quant type: H.263

I-VOPs: 693 (1,07%)
P-VOPs: 23569 (36,37%)
B-VOPs: 40540 (62,56%)
S-VOPs: 0 (0,00%)
N-VOPs: 1 (0,00%)

Max consecutive B-VOPs: 2
1 consec: 24,33%
2 consec: 75,67%

Moitah · 14th December 2005, 19:25

I was able to create a file like this using celtic_druid's latest CVS build (2005-Dec-10). It doesn't happen often, but it does (right before I-VOPs from what I've seen), and it's not a bug. I will fix MPEG4 Modifier very soon.

Moitah · 15th December 2005, 04:21

1.3.4 released. It better work

.

EDIT: UnpackCL just updated as well.

SeeMoreDigital · 15th December 2005, 10:20

There's just no stopping the man.... he's on a roll

Thanks mate

eXistenZ_69 · 15th December 2005, 10:32

Excellent, will check out the new version later today...

Thanks,

eXistenZ_69

eXistenZ_69 · 22nd December 2005, 01:32

Finally got around to testing the new 1.3.4 version, as far as I can tell it works great. Problem cases behave much better now on my pc and stand-alone player, no more visual artefacting

Sometimes a slight jerk is noticable, but that's as good as it gets I guess since the original packed files have an invalid bitstream. The file with real N-VOPs I have, got nicely unpacked as well.

So big thanks to you, Moitah

Moitah · 22nd December 2005, 04:54

@eXistenZ_69: You're welcome, thanks for testing.

13th December 2005, 17:55	#322 \| Link
SeeMoreDigital Life's clearer in 4K UHD Join Date: Jun 2003 Location: Notts, UK Posts: 12,227	Hi Moitah, Are you going to update this forum thread title? Cheers __________________ \| I've been testing hardware media playback devices and software A/V encoders and decoders since 2001 \| My Network Layout & A/V Gear \|

13th December 2005, 19:01	#324 \| Link
SeeMoreDigital Life's clearer in 4K UHD Join Date: Jun 2003 Location: Notts, UK Posts: 12,227	How odd... That said, I've noticed the same, when trying to amend some old posts in the Hardware Players section of the forum too! I ended up having to make up a short title first, save it. Open it again and then write the title I actually wanted, second time around Cheers __________________ \| I've been testing hardware media playback devices and software A/V encoders and decoders since 2001 \| My Network Layout & A/V Gear \|

14th December 2005, 17:43	#330 \| Link
Moitah Registered User Join Date: Feb 2004 Location: Virginia, USA Posts: 334	After checking some files I had, I found one also: Code: 8972: P-VOP (0:06:13.875) B-VOP (0:06:13.792) 8973: B-VOP (0:06:13.834) 8974: N-VOP(D) (0:06:13.875) 8975: N-VOP (0:06:13.917) 8976: I-VOP (0:06:13.959) The user data is marked XviD0038. So it looks like I have some more work to do. __________________ moitah.net

14th December 2005, 18:40	#333 \| Link
celtic_druid Registered User Join Date: Oct 2001 Location: Melbourne, Australia Posts: 2,171	1.1beta1 throught till but not including beta2 or xvid.h v1.48 and 1.49. I would still have all my old builds from that period (16 Jan 2005 -->04 Apr 2005) if anyone wants them. Six builds including 1.1beta1. Last edited by celtic_druid; 14th December 2005 at 18:47.

14th December 2005, 19:25	#335 \| Link
Moitah Registered User Join Date: Feb 2004 Location: Virginia, USA Posts: 334	I was able to create a file like this using celtic_druid's latest CVS build (2005-Dec-10). It doesn't happen often, but it does (right before I-VOPs from what I've seen), and it's not a bug. I will fix MPEG4 Modifier very soon. __________________ moitah.net

15th December 2005, 04:21	#336 \| Link
Moitah Registered User Join Date: Feb 2004 Location: Virginia, USA Posts: 334	1.3.4 released. It better work . EDIT: UnpackCL just updated as well. __________________ moitah.net Last edited by Moitah; 15th December 2005 at 04:28.

15th December 2005, 10:20	#337 \| Link
SeeMoreDigital Life's clearer in 4K UHD Join Date: Jun 2003 Location: Notts, UK Posts: 12,227	There's just no stopping the man.... he's on a roll Thanks mate __________________ \| I've been testing hardware media playback devices and software A/V encoders and decoders since 2001 \| My Network Layout & A/V Gear \|

15th December 2005, 10:32	#338 \| Link
eXistenZ_69 Registered User Join Date: Mar 2003 Posts: 87	Excellent, will check out the new version later today... Thanks, eXistenZ_69

22nd December 2005, 01:32	#339 \| Link
eXistenZ_69 Registered User Join Date: Mar 2003 Posts: 87	Finally got around to testing the new 1.3.4 version, as far as I can tell it works great. Problem cases behave much better now on my pc and stand-alone player, no more visual artefacting Sometimes a slight jerk is noticable, but that's as good as it gets I guess since the original packed files have an invalid bitstream. The file with real N-VOPs I have, got nicely unpacked as well. So big thanks to you, Moitah

22nd December 2005, 04:54	#340 \| Link
Moitah Registered User Join Date: Feb 2004 Location: Virginia, USA Posts: 334	@eXistenZ_69: You're welcome, thanks for testing. __________________ moitah.net