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

Closed Thread
 
Thread Tools Search this Thread Display Modes
Old 20th January 2009, 14:34   #7941  |  Link
nautilus7
Registered User
 
nautilus7's Avatar
 
Join Date: Jan 2006
Location: Athens, Greece
Posts: 1,518
Quote:
Originally Posted by asc28 View Post
So ArcSoft has always decoded ALL DTS-MA tracks w/ incorrect volume? Does Sonic (or anyone else) do it correctly?

The JAP Blu-Ray only has 5.1 sound, so the Nordic's audio is superior. I'll be combining it w/ video from JAP.
Did you see any similar message/warning with any other track?

Don't worry, eac3to we'll warn you when there's a problem. This track is one of the very few cases where ArcSoft decoder doesn't work ok.

Wasn't aware of the audio difference between the 2 Blu-rays of Sin City. But since ArcSoft can't decode it correctly, i don't know if it worths it.
nautilus7 is offline  
Old 20th January 2009, 14:38   #7942  |  Link
Jom
Registered User
 
Join Date: Mar 2008
Posts: 20
Regarding unregistering all the Arcsoft filters.

after installing TMT, I just opened up a command prompt, cd-ed to the directory that contained the filters and ran :

for /F %i in ('dir /B') do@regsvr32 /u /s %i

you then just regsvr32 the ax/dll's you need for DTS decoding.

as i recall, the names of the dlls are pretty obvious.
Jom is offline  
Old 20th January 2009, 15:29   #7943  |  Link
madshi
Registered Developer
 
Join Date: Sep 2006
Posts: 9,140
Quote:
Originally Posted by nautilus7 View Post
madshi, what about adding option to downmix to 6.1ch (-down7)?
Which downmix method should I use for that? Should I just combine the both back channels to one? Or should the back channels also partially be mixed into the side channels? That's the main problem with downmixing: Figuring out which mixing matrix should be used exactly. I don't want to add something new, if I'm not sure that I'm doing it the right way...

Quote:
Originally Posted by Paddy97 View Post
Is it possible when retreiving the features of a bluray disc to get more then 8?
eac3to is not limited to 8 titles. It lists 500 titles, if there are so many on the Blu-Ray disk. However, only titles with a runtime of 15 minutes or longer are listed. The reason for that is that there are quite a lot of Blu-Ray discs which have literally hundreds of titles. If I'd list all of them, the command prompt memory would run over so even if you scrolled up, you wouldn't see the movie's main title, anymore. So I *have* to set a runtime limit somewhere. Or else we'll be flooded with unneeded titles. Of course it's not nice if short films are missing in the title list. But with Blu-Ray there's no way for eac3to to know which titles are important and which aren't. So eac3to is strictly looking for a specific minimum runtime of 15 minutes.

Quote:
Originally Posted by asarian View Post
Simple. I asked eac3to to combine and extract a the main VC-1 stream, like:

eac3to 1) 2: c:\video\amg.vc1

I then remux with tsMuxeR, as I always do. Seems like each of the multipart m2ts files has its own duration index (or so it seems). As a result, the length of the entire movie is set to the length of the first part (~13 mins).
So the problem is most probably caused by tsMuxeR and not by eac3to. You can try playing the demuxed "amg.vc1" file, does it also stop at 13 min? Probably not. Or ask eac3to to mux the vc1 file to MKV. Does that MKV also stop playing at 13 min? Probably not. So it's probably a bug in tsMuxeR.

Quote:
Originally Posted by jfcarbel View Post
My followup question then, when would the ArcSoft DTS decoder be needed? Is this just when we need to re-encode the DTS to maybe a lower bitrate? And if so then I assume that task would then also require the SurCode DVD–DTS Encoder.
You need to separate decoding and encoding. You need a decoder for decoding and an encoder for encoding. The ArcSoft DTS decoder is needed for best quality DTS-HD decoding. It doesn't matter if you store the decoded data as WAV or if you reencode it to any other format.

Which encoders and decoders are recommended and which need to be supplied externally are documented on the first page of this thread.
madshi is offline  
Old 20th January 2009, 16:25   #7944  |  Link
asc28
Registered User
 
Join Date: Jan 2009
Posts: 31
Quote:
Originally Posted by nautilus7 View Post
Did you see any similar message/warning with any other track?

Don't worry, eac3to we'll warn you when there's a problem. This track is one of the very few cases where ArcSoft decoder doesn't work ok.

Wasn't aware of the audio difference between the 2 Blu-rays of Sin City. But since ArcSoft can't decode it correctly, i don't know if it worths it.
Oh, I thought that message could've been included in the new eac3to versions. Is there any alternative to decoding this track (Sonic doesn't do 7.1, right)? Would love to have Sin City in beautiful 7.1.

Also, was this problem discussed in this thread/elsewhere? What specifically is wrong w/ the decoded track? My interpretation is that the "strange channel setup" messes w/ the volume? Thanks for the help

Last edited by asc28; 20th January 2009 at 16:30.
asc28 is offline  
Old 20th January 2009, 16:31   #7945  |  Link
madshi
Registered Developer
 
Join Date: Sep 2006
Posts: 9,140
Quote:
Originally Posted by asc28 View Post
Also, was this problem discussed in this thread/elsewhere?
Well, there's an easy way to find out, you know?

madshi is offline  
Old 20th January 2009, 16:40   #7946  |  Link
asc28
Registered User
 
Join Date: Jan 2009
Posts: 31
Quote:
Originally Posted by madshi View Post
Well, there's an easy way to find out, you know?

Unfortunately, I don't think it's possible to search a specific thread, only sections of the forum. I've searched for "strange setup" and other terms, but it just returns the eac3to thread, not specific posts/pages in the 400-page thread.

Please let me know if there's any info on this problem, or a way to search for it. Thanks!
asc28 is offline  
Old 20th January 2009, 16:43   #7947  |  Link
madshi
Registered Developer
 
Join Date: Sep 2006
Posts: 9,140
Quote:
Originally Posted by asc28 View Post
Unfortunately, I don't think it's possible to search a specific thread
Come on, it's not so hard. Why don't you simply click on "search this thread"?
madshi is offline  
Old 20th January 2009, 16:44   #7948  |  Link
lexor
Registered User
 
Join Date: Jan 2004
Posts: 849
Quote:
Originally Posted by asc28 View Post
Unfortunately, I don't think it's possible to search a specific thread, only sections of the forum. I've searched for "strange setup" and other terms, but it just returns the eac3to thread, not specific posts/pages in the 400-page thread.

Please let me know if there's any info on this problem, or a way to search for it. Thanks!
Actually there is a way, scroll to the top of the page, there is a link saying "Search this Thread". Click it, a search box appears, type in search terms.

/edit: oops, madshi is too fast, but I gave more detailed instructions... probably 'couse I don't have to give them every page.
__________________
Geforce GTX 260
Windows 7, 64bit, Core i7
MPC-HC, Foobar2000
lexor is offline  
Old 20th January 2009, 17:05   #7949  |  Link
asc28
Registered User
 
Join Date: Jan 2009
Posts: 31
Quote:
Originally Posted by madshi View Post
Come on, it's not so hard. Why don't you simply click on "search this thread"?
Sorry, really retarded of me to miss that. Doh

I found this in an earlier post:
Quote:
Originally Posted by madshi
If you want, you can report this problem to the ArcSoft guys (together with a small sample). Should be easy for them to fix. Just ask them to decode the sample as 5.1 and then as 7.1. The 7.1 decoding volume will be lower, which doesn't really make any sense.
By 5.1, I'm guessing you mean the DTS core? So, if I wanted to, I could compare the volume difference between the 5.1 and 7.1 and alter the decoded PCM's volume by that amount...

Why wouldn't this process be lossless? Wouldn't adding +Xdb and then subtracting the same -Xdb value to a PCM track return the same data? Sorry if I'm relatively uninformed on this matter.

And hopefully, I can assume, every channel in the decoded ArcSoft track is too soft by the same amount?

Last edited by asc28; 20th January 2009 at 17:07.
asc28 is offline  
Old 20th January 2009, 17:13   #7950  |  Link
madshi
Registered Developer
 
Join Date: Sep 2006
Posts: 9,140
Quote:
Originally Posted by asc28 View Post
Wouldn't adding +Xdb and then subtracting the same -Xdb value to a PCM track return the same data?
No. Because volume change is done in floating point, while PCM is integer. Conversion from floating point to integer is not lossless.

Quote:
Originally Posted by asc28 View Post
And hopefully, I can assume, every channel in the decoded ArcSoft track is too soft by the same amount?
Yes. Furthermore with some tracks decoding doesn't work properly at all (distortion). If that is the case, you'll easily notice that on a quick check, though.
madshi is offline  
Old 20th January 2009, 17:26   #7951  |  Link
asc28
Registered User
 
Join Date: Jan 2009
Posts: 31
Quote:
Originally Posted by madshi View Post
No. Because volume change is done in floating point, while PCM is integer. Conversion from floating point to integer is not lossless.
Thanks for your help! I think I'll convert to floating-point PCM to apply the volume change, and then feed the floating-point directly to Nero AAC encoder. Hopefully, if I understand these concepts correctly, this should not result in much quality loss, had I just encoded an originally correct-volume PCM to AAC.

I'll also try to contact ArcSoft about the incorrect decoding.
asc28 is offline  
Old 20th January 2009, 17:30   #7952  |  Link
madshi
Registered Developer
 
Join Date: Sep 2006
Posts: 9,140
Quote:
Originally Posted by asc28 View Post
Thanks for your help! I think I'll convert to floating-point PCM to apply the volume change, and then feed the floating-point directly to Nero AAC encoder. Hopefully, if I understand these concepts correctly, this should not result in much quality loss, had I just encoded an originally correct-volume PCM to AAC.
The quality loss (compared to correct-volume PCM to AAC) should be minimal, most probably not audible for any normal person.

Quote:
Originally Posted by asc28 View Post
I'll also try to contact ArcSoft about the incorrect decoding.
If you get any useful reply, please let us know.
madshi is offline  
Old 20th January 2009, 17:54   #7953  |  Link
Boulder
Pig on the wing
 
Boulder's Avatar
 
Join Date: Mar 2002
Location: Finland
Posts: 5,718
A (yet another) feature request: ability to use the SoundTouch library for audio speedup or slowdown without changing the audio pitch.
__________________
And if the band you're in starts playing different tunes
I'll see you on the dark side of the Moon...
Boulder is offline  
Old 20th January 2009, 19:43   #7954  |  Link
asc28
Registered User
 
Join Date: Jan 2009
Posts: 31
Quote:
Originally Posted by asc28 View Post
I'll convert to floating-point PCM to apply the volume change, and then feed the floating-point directly to Nero AAC encoder. Hopefully, if I understand these concepts correctly, this should not result in much quality loss, had I just encoded an originally correct-volume PCM to AAC.
I actually might instead change the volume by applying gain on the AAC post-encode. I believe this would be a lossless process if the gain can be applied the same way mp3gain does it (by modifying the frame headers). Will eac3to do it this way (change volume losslessly by modifying ADTS frame headers on the AAC) if I do: eac3to input.aac output.aac -Xdb?

[edit] Also here are some tests I did w/ the latest WaveGain:
Quote:
wavegain sin51.C.wav sin71.C.wav sin51.LFE.wav sin71.LFE.wav sin51.L.wav sin71.L.wav sin51.R.wav sin71.R.wav

Analyzing...

Gain | Peak | Scale | New Peak |Left DC|Right DC| Track
| | | |Offset | Offset |
--------------------------------------------------------------
-1.16 dB | 32767 | 0.87 | 28671 | 0 | 0 | sin51.C.wav
+2.19 dB | 21538 | 1.29 | 27715 | 0 | 0 | sin71.C.wav
-0.00 dB | 32767 | 1.00 | 32766 | 0 | 0 | sin51.LFE.wav
+3.64 dB | 21538 | 1.52 | 32765 | 0 | 0 | sin71.LFE.wav
+0.15 dB | 32209 | 1.02 | 32766 | 0 | 0 | sin51.L.wav
+3.64 dB | 21538 | 1.52 | 32765 | 0 | 0 | sin71.L.wav
+0.02 dB | 32681 | 1.00 | 32766 | 0 | 0 | sin51.R.wav
+3.64 dB | 21538 | 1.52 | 32765 | 0 | 0 | sin71.R.wav

WaveGain Processing completed normally
Looks like the volume difference is not exactly the same for every channel (varies from 3.35 to 3.64), though this might be explained by how the extra 2 channels are mapped / DTS compression.

Last edited by asc28; 20th January 2009 at 20:19.
asc28 is offline  
Old 21st January 2009, 00:24   #7955  |  Link
asarian
Registered User
 
Join Date: May 2005
Posts: 1,462
Quote:
Originally Posted by asc28 View Post
Oh, I thought that message could've been included in the new eac3to versions. Is there any alternative to decoding this track (Sonic doesn't do 7.1, right)? Would love to have Sin City in beautiful 7.1.

Also, was this problem discussed in this thread/elsewhere? What specifically is wrong w/ the decoded track? My interpretation is that the "strange channel setup" messes w/ the volume? Thanks for the help
I have had both Sin-City Blu-Rays. The Japanese actually has TrueHD 5.1 audio (instead of just DD 5.1), but the track is simply broken: if you try to extract it, you'll find it to be corrupted, and eac3to aborted at the time, so I could only salvage the AC3 core. The Scandinavian Blu-Ray comes with an odd reddish tint in the video, but has a flawsless DTS-HD MA 7.1 track. So, I finally remuxed the video from the Japanese disc with the audio from the Scandinavian one, threw out the Japanese disc, and now all is fine again. My longwinded point being: I had no trouble whatsoever converting the DTS-HD MA 7.1 track from the Nordic disc to LPCM, and experienced no volume issue or conversion hicup of any kind. This was done with an older version of eac3to at the time (eac3to v2.78), and, of course, the Arcsoft DTS Decoder.
__________________
Gorgeous, delicious, deculture!
asarian is offline  
Old 21st January 2009, 00:47   #7956  |  Link
nautilus7
Registered User
 
nautilus7's Avatar
 
Join Date: Jan 2006
Location: Athens, Greece
Posts: 1,518
What you're saying about flawless decoding of the DTS-HD track is simply not possible. If you're talking about the same disc, then eac3to (+arcsoft) should have produced the same warning (unless that eac3to version didn't warn for this).

What do you mean the TrueHD from the JAP disc is corrupted? Do you have a log or something. (if you don't have don't mind, i have the disc so i can test in a few days.)
nautilus7 is offline  
Old 21st January 2009, 08:22   #7957  |  Link
Paddy97
Registered User
 
Join Date: Oct 2006
Posts: 70
Quote:
Originally Posted by madshi View Post
eac3to is not limited to 8 titles. It lists 500 titles, if there are so many on the Blu-Ray disk. However, only titles with a runtime of 15 minutes or longer are listed. The reason for that is that there are quite a lot of Blu-Ray discs which have literally hundreds of titles. If I'd list all of them, the command prompt memory would run over so even if you scrolled up, you wouldn't see the movie's main title, anymore. So I *have* to set a runtime limit somewhere. Or else we'll be flooded with unneeded titles.
Thanks for the update on how the logic works. I can relate to this design choice.

As allways there is a way around everything even though you need some more tools. What I do now is that I use BDInfo to pinpoint the correct playlist ( using the length as a way of picking the right one ) and then I go from there.
Paddy97 is offline  
Old 21st January 2009, 10:25   #7958  |  Link
asarian
Registered User
 
Join Date: May 2005
Posts: 1,462
Quote:
Originally Posted by nautilus7 View Post

What do you mean the TrueHD from the JAP disc is corrupted? Do you have a log or something. (if you don't have don't mind, i have the disc so i can test in a few days.)
Actually, I still have a log.

eac3to 1) 3: c:\video\sincity.pcm
M2TS, 1 video track, 4 audio tracks, 5 subtitle tracks, 2:04:01
1: Chapters, 29 chapters
2: h264/AVC, 1080p24 /1.001 (16:9)
3: TrueHD/AC3, English, 5.1 channels, 48khz, dialnorm: -27dB
(embedded: AC3, 5.1 channels, 640kbps, 48khz, dialnorm: -27dB), -13ms
4: TrueHD/AC3, Japanese, 5.1 channels, 48khz, dialnorm: -27dB
(embedded: AC3, 5.1 channels, 640kbps, 48khz, dialnorm: -27dB), -13ms
5: AC3, English, 2.0 channels, 256kbps, 48khz, dialnorm: -27dB, -13ms
6: AC3, English, 2.0 channels, 256kbps, 48khz, dialnorm: -27dB, -13ms
7: Subtitle (PGS), Japanese
8: Subtitle (PGS), Japanese
9: Subtitle (PGS), English
10: Subtitle (PGS), Japanese
11: Subtitle (PGS), Japanese
[a03] Extracting audio track number 3...
[a03] This track is not clean.
[a03] Extracting TrueHD stream...
[a03] Decoding with libav/ffmpeg...
[libav] Stream parameters not seen; skipping frame
[a03] The libav decoder output an unexpected bitdepth (1).
Aborted at file position 49152.


The Nordic disc decoded without any warning or noticeable lower volume, near as I can tell.
__________________
Gorgeous, delicious, deculture!
asarian is offline  
Old 21st January 2009, 10:46   #7959  |  Link
madshi
Registered Developer
 
Join Date: Sep 2006
Posts: 9,140
Quote:
Originally Posted by Boulder View Post
A (yet another) feature request: ability to use the SoundTouch library for audio speedup or slowdown without changing the audio pitch.
As far as I know, SoundTouch is not the preferred choice for best quality processing. Commercial competing algorithms are supposed to sound noticeably better, right? I don't really like the idea to implement something and say: eac3to can do it, but better don't use it, cause it's quality isn't really good.

Quote:
Originally Posted by asarian View Post
Actually, I still have a log.

eac3to 1) 3: c:\video\sincity.pcm
M2TS, 1 video track, 4 audio tracks, 5 subtitle tracks, 2:04:01
1: Chapters, 29 chapters
2: h264/AVC, 1080p24 /1.001 (16:9)
3: TrueHD/AC3, English, 5.1 channels, 48khz, dialnorm: -27dB
(embedded: AC3, 5.1 channels, 640kbps, 48khz, dialnorm: -27dB), -13ms
4: TrueHD/AC3, Japanese, 5.1 channels, 48khz, dialnorm: -27dB
(embedded: AC3, 5.1 channels, 640kbps, 48khz, dialnorm: -27dB), -13ms
5: AC3, English, 2.0 channels, 256kbps, 48khz, dialnorm: -27dB, -13ms
6: AC3, English, 2.0 channels, 256kbps, 48khz, dialnorm: -27dB, -13ms
7: Subtitle (PGS), Japanese
8: Subtitle (PGS), Japanese
9: Subtitle (PGS), English
10: Subtitle (PGS), Japanese
11: Subtitle (PGS), Japanese
[a03] Extracting audio track number 3...
[a03] This track is not clean.
[a03] Extracting TrueHD stream...
[a03] Decoding with libav/ffmpeg...
[libav] Stream parameters not seen; skipping frame
[a03] The libav decoder output an unexpected bitdepth (1).
Aborted at file position 49152.
This is probably with a rather old eac3to build. Try a new one.
madshi is offline  
Old 21st January 2009, 10:54   #7960  |  Link
Boulder
Pig on the wing
 
Boulder's Avatar
 
Join Date: Mar 2002
Location: Finland
Posts: 5,718
Quote:
Originally Posted by madshi View Post
As far as I know, SoundTouch is not the preferred choice for best quality processing. Commercial competing algorithms are supposed to sound noticeably better, right? I don't really like the idea to implement something and say: eac3to can do it, but better don't use it, cause it's quality isn't really good.
What program would you recommend for such operation? Preferably something that accepts 32bit float input
__________________
And if the band you're in starts playing different tunes
I'll see you on the dark side of the Moon...

Last edited by Boulder; 21st January 2009 at 10:57.
Boulder is offline  
Closed Thread

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 00:53.


Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2024, vBulletin Solutions Inc.