Welcome to Doom9's Forum, THE in-place to be for everyone interested in DVD conversion.

Before you start posting please read the forum rules. By posting to this forum you agree to abide by the rules.

 

Go Back   Doom9's Forum > General > Audio encoding

Reply
 
Thread Tools Search this Thread Display Modes
Old 23rd June 2007, 17:18   #221  |  Link
xkodi
Registered User
 
Join Date: Aug 2002
Posts: 221
Quote:
Originally Posted by madshi View Post
If you compare the original 16bit WAV source to the 24bit MLP decoder output, do you see a pattern? Are the 2 most significant bytes always 100% correct and the least significant byte contains "random" data? Or are the 2 most significant bytes not fully correct, either? If only the least significant byte is wrong and the 2 most significant bytes are correct, I could add code to eac3to to simply zero out the least significant byte and we'd be ready to go!

Thanks!
yes, the most significant bytes of the decoded 16bit MLP are always identical to the source, so if the least significant bytes are stripped off from the decoded file, then it is identical to the source. the wrong least significant bytes usually are small numbers like 0x01, 0x03, etc.

so, yes, code to eac3to to remove the least significant byte, when the source TrueHD/MLP is 16bit, seems to be the most proper solution based on the info we now have and we can easily judge, according to the previous test with the compression ratios, which TrueHD track is 24bit and which is 16bit, but still the thing that most least significant bytes with 16bit TrueHD stream are not zero bugs me, because with 16bit MLP most least significant bytes are zero and very few of them are not zero, but my MLP test samples are small, around 100Kb, so maybe then it is normal.

Last edited by xkodi; 23rd June 2007 at 17:28.
xkodi is offline   Reply With Quote
Old 23rd June 2007, 17:27   #222  |  Link
madshi
Registered Developer
 
Join Date: Sep 2006
Posts: 9,137
I've just checked your files. "164851_nero_decode" and "164851_merge.wav" are 100% identical - if I strip the least significant byte from the "164851_nero_decode" file. And I've done another check: The least significant byte is *ALWAYS* zero in "164851_nero_decode".

Now I'm confused...

Where did you find a non-zero least significant byte?
madshi is offline   Reply With Quote
Old 23rd June 2007, 19:06   #223  |  Link
xkodi
Registered User
 
Join Date: Aug 2002
Posts: 221
Quote:
Originally Posted by madshi View Post
I've just checked your files. "164851_nero_decode" and "164851_merge.wav" are 100% identical - if I strip the least significant byte from the "164851_nero_decode" file. And I've done another check: The least significant byte is *ALWAYS* zero in "164851_nero_decode".

Now I'm confused...

Where did you find a non-zero least significant byte?
you are right, i've done so many tests and now, after i've double checked the files, i'm also confused where, when and if i really saw that or made a mistake in the hurry. that is why i made the archive, so someone else to verify my tests and conclusions, because maybe i've made mistakes, for the moment seems that the correct conclusion is that Nero always uses 24bit output and always correctly decodes MLP, so we are almost back to the very beginning with the TrueHD tracks ...
xkodi is offline   Reply With Quote
Old 23rd June 2007, 19:12   #224  |  Link
madshi
Registered Developer
 
Join Date: Sep 2006
Posts: 9,137
Quote:
Originally Posted by xkodi View Post
you are right, i've done so many tests and now, after i've double checked the files, i'm also confused where, when and if i really saw that or made a mistake in the hurry. that is why i made the archive, so someone else to verify my tests and conclusions, because maybe i've made mistakes, for the moment seems that the correct conclusion is that Nero always uses 24bit output and always correctly decodes MLP, so we are almost back to the very beginning with the TrueHD tracks ...
Agreed. Seems that MLP decoding is perfect, but TrueHD decoding probably not. Too bad that you don't have a TrueHD encoder...
madshi is offline   Reply With Quote
Old 23rd June 2007, 20:21   #225  |  Link
xkodi
Registered User
 
Join Date: Aug 2002
Posts: 221
Quote:
Originally Posted by madshi View Post
Agreed. Seems that MLP decoding is perfect, but TrueHD decoding probably not. Too bad that you don't have a TrueHD encoder...
maybe i should look if there are Blu-ray discs, that have both PCM track and TrueHD track, then i can put part of the TrueHD track into EVOB, the same way i put MLP into EVOB, decode and compare it with the PCM track, if i'm lucky the TrueHD should be created from the same PCM.

p.s. Ghost Rider Blu-ray disc seems to have both PCM and TrueHD tracks, so maybe in several weeks, it will be possible for me to do such test, sounds like good test.
xkodi is offline   Reply With Quote
Old 23rd June 2007, 20:40   #226  |  Link
madshi
Registered Developer
 
Join Date: Sep 2006
Posts: 9,137
Quote:
Originally Posted by xkodi View Post
maybe i should look if there are Blu-ray discs, that have both PCM track and TrueHD track, then i can put part of the TrueHD track into EVOB, the same way i put MLP into EVOB, decode and compare it with the PCM track, if i'm lucky the TrueHD should be created from the same PCM.

p.s. Ghost Rider Blu-ray disc seems to have both PCM and TrueHD tracks, so maybe in several weeks, it will be possible for me to do such test, sounds like good test.
I think those Blu-Rays with both PCM and TrueHD have 16bit PCM and 20bit TrueHD.
madshi is offline   Reply With Quote
Old 24th June 2007, 18:46   #227  |  Link
Beastie Boy
Registered User
 
Beastie Boy's Avatar
 
Join Date: Mar 2002
Posts: 353
I'm currently trying to go from eAC3 to multichannel WMA whilst preserving as much quality as possible, but something isn't right.
I used eAC3to to go from my ec3 file to wav. The resulting wave file would not play using MPC, it simply crashed. Maybe I need another filter?
Anyway, undeterred I used Tranzcode to produce 6 mono wave files. These files will play but the sound is garbled, just a series of squarks. I know eAC3to works OK on my machine, because the AC3 files it produces are OK.

Any ideas where I am going wrong?

Cheers, Beastie.

PS. If this problem is not related to eAC3to I'll start a new thread so as not to go off-topic.
Beastie Boy is offline   Reply With Quote
Old 25th June 2007, 08:08   #228  |  Link
The_Keymaker
Registered User
 
Join Date: Feb 2007
Location: U.S.A.
Posts: 135
Latest version of EAC3toGUI (version 0.97)

http://www.sendspace.com/file/zdfq3x

Changelog:
- Fixed bug that would not allow *.wav files as input.

Apologize for any inconvenience this may have caused.

The_Keymaker
The_Keymaker is offline   Reply With Quote
Old 25th June 2007, 10:06   #229  |  Link
ACrowley
Registered User
 
Join Date: Apr 2006
Posts: 1,008
Thx keymaker

Great work!


Maybe its better to rename "Orbitlee" to "Sonic" ?
I mean Orbitlee DTS/AC3/DD+ Source is only the Source Filter ,the Decoder is SonicCinevisionAudioDecoder4.2

So you can select Nero=NeroAudioDecoder2 or Sonic=SonicCinevisionAudioDecoder4.2

And a .raw input extension would be good

Last edited by ACrowley; 25th June 2007 at 10:58.
ACrowley is offline   Reply With Quote
Old 25th June 2007, 12:51   #230  |  Link
madshi
Registered Developer
 
Join Date: Sep 2006
Posts: 9,137
Quote:
Originally Posted by Beastie Boy View Post
I'm currently trying to go from eAC3 to multichannel WMA whilst preserving as much quality as possible, but something isn't right.
I used eAC3to to go from my ec3 file to wav. The resulting wave file would not play using MPC, it simply crashed. Maybe I need another filter?
Anyway, undeterred I used Tranzcode to produce 6 mono wave files. These files will play but the sound is garbled, just a series of squarks. I know eAC3to works OK on my machine, because the AC3 files it produces are OK.

Any ideas where I am going wrong?

Cheers, Beastie.

PS. If this problem is not related to eAC3to I'll start a new thread so as not to go off-topic.
Not sure what the cause is if the AC3 files are correct. Maybe you just need to wait until I've added mono wave output support to eac3to.
madshi is offline   Reply With Quote
Old 25th June 2007, 15:19   #231  |  Link
Beastie Boy
Registered User
 
Beastie Boy's Avatar
 
Join Date: Mar 2002
Posts: 353
Quote:
Originally Posted by madshi View Post
Maybe you just need to wait until I've added mono wave output support to eac3to.
I was hoping you would say that

Cheers, Beastie.
Beastie Boy is offline   Reply With Quote
Old 25th June 2007, 15:40   #232  |  Link
ACrowley
Registered User
 
Join Date: Apr 2006
Posts: 1,008
Damn. i tried everything
I cant get anything working from a 24bit lpcm from Bluray

I tried to demux with SonicHDDemuxer, also with xport.

I use this Sox commadline and tried it as little endian or big endian, with -2 and -3
sox.exe -l -r48000 -t .raw -c 6 -3 -s e:\test.raw e:\test.wav

nothing! When i use eac3to do it autom. its the same Problem

The wave File makes only a heavy loud noise, 6 mono channels ( created with wavewizard or tranzcode) to.

The full lpcm 24bit is 7Gb big.
The encoded flac from eac3to is "6Gb" and playback fails.
Its wrong cause the flac must be around 3Gb.
And AC3/ wave output only the loud noise too.
A 24-16bit conversion create the same Result

When i try it on a 100Mb sample its the same

It happens with "all" 24bit lpcm Tracks from Bluray.
I really tried everything, checked it it again and again , with no succes

LPCM 16 bit works without Problems.

I really dont know where the Problem ? Any advice ?

I could only image its a Problem caused by demuxing.

Heres the same Problem, but no Solution so far i can see ?
http://forum.doom9.org/showthread.php?t=124080&page=2

Last edited by ACrowley; 25th June 2007 at 16:24.
ACrowley is offline   Reply With Quote
Old 25th June 2007, 15:42   #233  |  Link
infernix
the funk phenomena
 
infernix's Avatar
 
Join Date: Feb 2007
Posts: 21
I have a feature request.

Currently I'm running eac3to on a windows virtual machine in which it stores everything (source + raw + wav + ac3/flac) to a network share. For multiple reasons, I'd like to proceed on linux after the raw file has been dumped.

Seeing that most the tools used (all but the directsource stuff?) are available there, I'd like a simple parameter (like -n for make) that spits out the commands it would run after the directshow dump to raw, so that I can run sox and aften in linux. Thats basically everything after:
Code:
E-AC3, 5.1 channels, 1:59:02, 1536kbit/s, 48khz
Muxing eac3 file to raw. Please wait...
Audio Decoder output was 24 bit this time.

Yes, I could just dump to raw myself, then do strings eac3to.exe and figure out the right commands, but i'm lazy, and it's probably really simple to add
infernix is offline   Reply With Quote
Old 25th June 2007, 17:01   #234  |  Link
madshi
Registered Developer
 
Join Date: Sep 2006
Posts: 9,137
Quote:
Originally Posted by ACrowley View Post
Damn. i tried everything
I cant get anything working from a 24bit lpcm from Bluray
With the SonicHDDemuxer you'll probably get little endian instead of big endian. You may need to adjust the parameters of the tools you're using. E.g. when using eac3to, add "-little" as a test.
madshi is offline   Reply With Quote
Old 25th June 2007, 17:03   #235  |  Link
madshi
Registered Developer
 
Join Date: Sep 2006
Posts: 9,137
Quote:
Originally Posted by infernix View Post
I'd like a simple parameter (like -n for make) that spits out the commands it would run after the directshow dump to raw
Ok, you're the 2nd person asking me about this. So I guess I should implement it.
madshi is offline   Reply With Quote
Old 26th June 2007, 09:10   #236  |  Link
ACrowley
Registered User
 
Join Date: Apr 2006
Posts: 1,008
Quote:
Originally Posted by madshi View Post
With the SonicHDDemuxer you'll probably get little endian instead of big endian. You may need to adjust the parameters of the tools you're using. E.g. when using eac3to, add "-little" as a test.


EDIT :
I have it! Very strange cause i tried all theye combinations before...mhh

Problem : even the 24bit lpcm was demxued with graphedit, it works only with Big Endian Switch in Sox

1 .sox.exe -B -r48000 -t .raw -c 6 -3 -s e:\test.raw e:\test.wav
2. wavewizard 0 1 2 5 3 4
3. encode to whatever

@Madshi
Ofcourse i tried all Combinations mayn Times before.

I can say :eac3to fails (for me)on these 24bit lpcm , But it works manually .
Even with Big or Little Endian switch, nothing! No working wave, no flac,no ac3. outputs is a loud noise.

eac3to e:\test.pcm e:\lpcm.wav -blu-ray -24 ( i tried with -B / -L all combinations)

Heres the sample. Its LPCM 24Bit 5.1 from BluRay "Apocalypto"

http://rapidshare.com/files/39409043...24Bit.rar.html

Last edited by ACrowley; 26th June 2007 at 16:55.
ACrowley is offline   Reply With Quote
Old 26th June 2007, 16:15   #237  |  Link
kweek
Registered User
 
kweek's Avatar
 
Join Date: Oct 2006
Posts: 9
First thanks for this great prog

I don't know if this is a bug or if I'm doing something wrong but I discovered this because some one said that you can create smaller flacs with the -8 option (default is -5 see http://flac.sourceforge.net/document...coding_options).

When I use the following command
"eac3to.exe input.ddp output.flac -nero"
and I want to compress the flac some more with the command
"flac output.flac -o output.reenc.flac -f -8".

I get the following error
"output.flac: ERROR: FLAC input has STREAMINFO with unknown total samples which is not supported"

With "metaflac.exe output.flac --list" I get this.
Code:
METADATA block #0
  type: 0 (STREAMINFO)
  is last: false
  length: 34
  minimum blocksize: 4096 samples
  maximum blocksize: 4096 samples
  minimum framesize: 0 bytes
  maximum framesize: 0 bytes
  sample_rate: 48000 Hz
  channels: 6
  bits-per-sample: 24
  total samples: 0
  MD5 signature: 00000000000000000000000000000000
METADATA block #1
  type: 4 (VORBIS_COMMENT)
  is last: false
  length: 40
  vendor string: reference libFLAC 1.1.4 20070213
  comments: 0
METADATA block #2
  type: 1 (PADDING)
  is last: true
  length: 8192
Notice that streaminfo is corrupt. No minimum/maximum framesize, no total samples and no MD5 signature.

Now that someone who told me about the -8 option also gave me some pointers to fix this. Remember the sample_rate, channels and bits-per-sample.

Create a raw file from the corrupt flac.
"flac -d output.flac -o temp.raw --force-raw-format --endian=little --sign=signed"

Next create the flac from the raw with the information from metaflac (with the -8 option if you want)
"flac -8 --force-raw-format --endian=little --channels=6 --bps=24 --sample-rate=48000 --sign=signed -o "output.fixed.flac" - < temp.raw"

Now you can see that when you use metaflac on the new file it has the correct streaminfo, but I think eac3to should set the correct streaminfo to begin with .
Code:
METADATA block #0
  type: 0 (STREAMINFO)
  is last: false
  length: 34
  minimum blocksize: 4096 samples
  maximum blocksize: 4096 samples
  minimum framesize: 417 bytes
  maximum framesize: 42568 bytes
  sample_rate: 48000 Hz
  channels: 6
  bits-per-sample: 24
  total samples: 310497536
  MD5 signature: c5fd63920d96db81ec780bacb00ca4eb
METADATA block #1
  type: 4 (VORBIS_COMMENT)
  is last: false
  length: 40
  vendor string: reference libFLAC 1.1.4 20070213
  comments: 0
METADATA block #2
  type: 1 (PADDING)
  is last: true
  length: 8192

Last edited by kweek; 26th June 2007 at 16:21.
kweek is offline   Reply With Quote
Old 26th June 2007, 17:02   #238  |  Link
ACrowley
Registered User
 
Join Date: Apr 2006
Posts: 1,008
Quote:
Originally Posted by kweek View Post
First thanks for this great prog

I don't know if this is a bug or if I'm doing something wrong but I discovered this because some one said that you can create smaller flacs with the -8 option (default is -5 see http://flac.sourceforge.net/document...coding_options).

[/CODE]
-8 is maximum quality with maximum compression ,but slower encoding.
-5 should output a liottle bit bigger File with less compression and lower Quality

eac3to use -8 too, i would only use 8 for "HD Audio"

Last edited by ACrowley; 26th June 2007 at 17:21.
ACrowley is offline   Reply With Quote
Old 26th June 2007, 19:35   #239  |  Link
kweek
Registered User
 
kweek's Avatar
 
Join Date: Oct 2006
Posts: 9
Quote:
Originally Posted by ACrowley View Post
-8 is maximum quality with maximum compression ,but slower encoding.
-5 should output a liottle bit bigger File with less compression and lower Quality

eac3to use -8 too, i would only use 8 for "HD Audio"
Hi, thanks for you response. But the post was more about that eac3to doesn't set the streaminfo right.

Do you mean that eac3to uses the -8 option for flac? I'm using eac3to mainly for highdef audio.
kweek is offline   Reply With Quote
Old 26th June 2007, 20:30   #240  |  Link
ACrowley
Registered User
 
Join Date: Apr 2006
Posts: 1,008
Quote:
Originally Posted by kweek View Post
Hi, thanks for you response. But the post was more about that eac3to doesn't set the streaminfo right.

Do you mean that eac3to uses the -8 option for flac? I'm using eac3to mainly for highdef audio.

yes eac3to should use -8
The flac output has the same Size compared to a media coder -8 output.

################


ABout Sonic Cinevision 2 beta :

There a a few new Decoder Versions. Located in
C:\xxSonic/Cinevision/Filters

1. Video Decoder (I think based on Elecard/Moonlight) all are .ax Files and you can reg/use them all
-Sonic AVC/H264 Decoder
-Sonic VC1 Decoder
-Sonic MPEG Demuxer
-no AudioDecoder

2. The Cinemaster HD 4.3 Decoder ,all are .dll Files
-SonicHDDemxuer 4.3
-SonicCinemsterVideoDecoder 4.3
-SonicCinemster┴udioDecoder 4.3

The .ax Filter are working very good on BluRay H264/VC1 Video.
No Artefacts on H264 PAFF etc.

More interesting for this Thread is ,that the 4.3 CinemasterAudioDecoder connects to TrueHD!
The TrueHD is playing ,but you can hear a noise in the backround. But basicly it should work now.
I Tried it on Matrix HJDDVD TrueHD 24bit

The other Thing is the Cinevision 2 has no installation fo the Cinemaster 4.2 Decoder. When i try to reg the manually it fails with a Error Message.
You can overwrite the 4.2 dlls with the new 4.3. But Avisynth wont connect anymore.
So, it would be nice if somebody can find out a workaround how to properly reg. the Cinemaster HD 4.3 Decoder

Last edited by ACrowley; 26th June 2007 at 20:46.
ACrowley is offline   Reply With Quote
Reply

Tags
eac3to

Thread Tools Search this Thread
Search this Thread:

Advanced Search
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off

Forum Jump


All times are GMT +1. The time now is 09:14.


Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2019, vBulletin Solutions Inc.