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 25th November 2009, 07:13   #9581  |  Link
AnryV
Registered User
 
Join Date: Jun 2007
Posts: 42
Quote:
Originally Posted by Thunderbolt8 View Post
afaik its a bug of the arcsoft decoder, we've already had this example that it cannot handle dts-hd ma 1.0 files
What about Sonic decoder?

Last edited by AnryV; 25th November 2009 at 14:55.
AnryV is offline  
Old 25th November 2009, 07:20   #9582  |  Link
asc28
Registered User
 
Join Date: Jan 2009
Posts: 31
Is there any way to recalculate and add back dialog normalization to AC3 tracks, after eac3to has stripped them?

[edit] realized that many studios use an arbitrary figure, so might not be able to recalculate. But how about using a ReplayGain like technique to find the optimal dB to be added and then adding it to an already encoded AC3 track?

Also, I know this has been asked a couple times before, but no one seems to know how exactly to go about reinserting dialnorm metadata without re-encoding.

Last edited by asc28; 25th November 2009 at 07:41.
asc28 is offline  
Old 25th November 2009, 09:59   #9583  |  Link
Esurnir
Registered User
 
Join Date: Jun 2008
Posts: 143
I'm going to sound dumb, but which product from arcsoft still include the DTS-HD filter decoder? Does TotalMedia Theater 3 work ? Any special editions? (platinum?)
Esurnir is offline  
Old 25th November 2009, 19:30   #9584  |  Link
Thunderbolt8
Registered User
 
Join Date: Sep 2006
Posts: 2,197
Quote:
Originally Posted by AnryV View Post
What about Sonic decoder?
dont know to be frank.
Thunderbolt8 is offline  
Old 27th November 2009, 05:44   #9585  |  Link
IanD
Registered User
 
Join Date: Jun 2003
Location: Australia
Posts: 190
I've been converting a DTS HD MA and DD2.0 soundtrack to wav with Eac3to, using -core and -down2 for the first, in order to align them.

However neither of the resulting wav files are directly readable by Goldwave, Cooledit or CDwave as it seems they are "raw". No matter what bit depth, signature or sample rate I choose, I only get noise. Strangely I am able to play the wav files with Directshow no problem.

How do I make the wav files readable in the sound editor applications?

It's frankly annoying that the output isn't wav standard.
IanD is offline  
Old 27th November 2009, 09:39   #9586  |  Link
kypec
User of free A/V tools
 
kypec's Avatar
 
Join Date: Jul 2006
Location: SK
Posts: 826
Try to add switch -simple and see if it helps your applications to recognize such stereo WAV files. If yes then you should blame those apps that they don't support standard WAV headers.
kypec is offline  
Old 29th November 2009, 01:01   #9587  |  Link
tebasuna51
Moderator
 
tebasuna51's Avatar
 
Join Date: Feb 2005
Location: Spain
Posts: 6,890
GoldWave (from v5.02, 2004? ) read WAVE_FORMAT_EXTENSIBLE headers without problems.
__________________
BeHappy, AviSynth audio transcoder.
tebasuna51 is offline  
Old 30th November 2009, 04:40   #9588  |  Link
73ChargerFan
Registered User
 
73ChargerFan's Avatar
 
Join Date: Dec 2006
Posts: 523
Quote:
Originally Posted by IanD View Post
I've been converting a DTS HD MA and DD2.0 soundtrack to wav with Eac3to, using -core and -down2 for the first, in order to align them.

However neither of the resulting wav files are directly readable
Perhaps the DTS core is being converted to a dts-wav file, instead of being decoded?

eac3to (DTS or DD file) filename.wavs should decode, one channel per file.
73ChargerFan is offline  
Old 30th November 2009, 12:35   #9589  |  Link
IanD
Registered User
 
Join Date: Jun 2003
Location: Australia
Posts: 190
Using -simple seems to have done the trick and both converted DTS HD MA and DD2.0 source files are now recognised as stereo wav.

However, I'm experiencing problems trying to synchronise the tracks: the DTS HD MA soundtrack is identified by eac3to as 24fps (not 23.976fps), whilst the commentary track I'm trying to synchronise to it is from a PAL source. Using the switches -25.000 and -ChangeTo24.000 results in the commentary being about 500ms longer than the soundtrack when the beginnings are synchronised.

I thought it might be possible to use the source switch -24.99x to fine tune the commentary conversion, in case the PAL source is not quite 25fps, but eac3to refuses to accept anything other than -25.000

Any thoughts?

Unfortunately I don't have access to an NTSC commentary, only a PAL one. It's quite close at the moment but not close enough because the commentary also has parts of the soundtrack that need to be in-sync with the on-screen action.
IanD is offline  
Old 30th November 2009, 20:53   #9590  |  Link
Vick
Registered User
 
Join Date: Nov 2009
Posts: 6
Quote:
Originally Posted by madshi View Post
Many thanks to Christopher Key for providing the HDCD decoder!
Hi madshi

Whether Chris has given you the HDCD decoder source code?
He does not respond to email (cjk32 cam ac uk)

Thnx
Vick
p.s. sorry for my poor english
Vick is offline  
Old 30th November 2009, 22:40   #9591  |  Link
pcordes
Registered User
 
Join Date: Mar 2008
Posts: 47
hard to use under Wine on GNU/Linux

I'm an old-school GNU/Linux command line junkie. Running eac3to inside
screen, the output comes out like this:

Code:
$ alias eac3to='wine /usr/local/.../eac3to.exe'
$ eac3to foo.eac3
fixme:mixer:ALSA_MixerInit No master control found on HDA ATI HDMI, disabling mixer
fixme:reg:GetNativeSystemInfo (0x32fc80) using GetSyE-AC3, 5.1 channels, 0:31:59, 1536kbps, 48khz
# or, discarding wine's messages: eac3to 2>/dev/null
$ eac3to 2(1) Haali Matroska Muxer
I spent a while looking for --help output, until I eventually realized that its output was designed for a Windows command window, not a VT100 terminal emulator. It happens to work outside of screen, in just gnome-terminal, and in native xterm, though.

eac3to outputs a bunch of ^h (ascii 0x08) characters at the start of every line, and maybe under screen, reverse-line-wrap happens, and the cursor backspaces up to the previous line, so every line of text overwrites the one before. If I redirect the output to a file,
Code:
$ eac3to > eac3to.txt  2>/dev/null
$ file eac3to.txt 
eac3to.txt: ASCII English text, with CRLF line terminators, with overstriking
$ hexdump -C -v eac3to.txt
00000000  08 08 08 08 08 08 08 08  08 08 08 08 08 08 08 08  |................|
00000010  08 08 08 08 08 08 08 08  08 08 08 08 08 08 08 08  |................|
00000020  08 08 08 08 08 08 08 08  08 08 08 08 08 08 08 08  |................|
00000030  08 08 08 08 08 08 08 08  08 08 08 08 08 08 08 08  |................|
00000040  08 08 08 08 08 08 08 08  08 08 08 08 08 08 08 08  |................|
00000050  20 20 20 20 20 20 20 20  20 20 20 20 20 20 0d 0a  |              ..|
00000060  08 08 08 08 08 08 08 08  08 08 08 08 08 08 08 08  |................|
00000070  08 08 08 08 08 08 08 08  08 08 08 08 08 08 08 08  |................|
00000080  08 08 08 08 08 08 08 08  08 08 08 08 08 08 08 08  |................|
00000090  08 08 08 08 08 08 08 08  08 08 08 08 08 08 08 08  |................|
000000a0  08 08 08 08 08 08 08 08  08 08 08 08 08 08 08 08  |................|
000000b0  20 20 20 20 20 20 20 20  20 20 20 20 20 20 20 20  |                |
000000c0  20 20 20 20 20 20 20 20  20 20 20 20 20 20 20 20  |                |
000000d0  20 20 20 20 20 20 20 20  0d 0a 08 08 08 08 08 08  |        ........|
000000e0  08 08 08 08 08 08 08 08  08 08 08 08 08 08 08 08  |................|
000000f0  08 08 08 08 08 08 08 08  08 08 08 08 08 08 08 08  |................|
00000100  08 08 08 08 08 08 08 08  08 08 08 08 08 08 08 08  |................|
00000110  08 08 08 08 08 08 08 08  08 08 08 08 08 08 08 08  |................|
00000120  08 08 08 08 08 08 08 08  08 08 65 61 63 33 74 6f  |..........eac3to|
00000130  20 76 33 2e 31 37 2c 20  66 72 65 65 77 61 72 65  | v3.17, freeware|
00000140  20 62 79 20 6d 61 64 73  68 69 2e 6e 65 74 20 20  | by madshi.net  |
00000150  20 20 20 20 20 20 20 20  20 20 20 20 20 20 20 20  |                |
00000160  20 20 20 20 20 20 20 20  20 20 20 20 20 20 20 20  |                |
00000170  20 20 20 20 20 20 20 20  20 0d 0a 08 08 08 08 08  |         .......|
00000180  08 08 08 08 08 08 08 08  08 08 08 08 08 08 08 08  |................|
00000190  08 08 08 08 08 08 08 08  08 08 08 08 08 08 08 08  |................|
000001a0  08 08 08 08 08 08 08 08  08 08 08 08 08 08 08 08  |................|
000001b0  08 08 08 08 08 08 08 08  08 08 08 08 08 08 08 08  |................|
000001c0  08 08 08 08 08 08 08 08  08 08 08 20 20 20 20 20  |...........     |
000001d0  20 20 20 20 20 20 20 20  20 20 20 20 20 20 20 20  |                |
000001e0  20 20 20 20 20 20 20 20  20 20 20 20 20 20 20 20  |                |
000001f0  20 20 20 20 20 20 20 20  20 20 20 20 20 20 20 20  |                |
00000200  20 20 20 20 20 20 20 20  20 20 20 20 20 20 20 20  |                |
00000210  20 20 20 20 20 20 20 20  20 20 0d 0a 08 08 08 08  |          ......|
...
But the overstriking doesn't have any effect anywhere: xterm or gnome-terminal. (it visually looks the same as regular text). In VT100, ^h is cursor-movement, not highlighting. If eac3to is doing that itself on purpose, please don't, at least under wine. The CRLF line endings are not a problem.

I'd also love to have an option to not play sound. I haven't tried just renaming the .wav files. Maybe you could add a -unix option, or check an environment variable, to turn off sound and not print funky text. It's harder to deal with options at the end, instead of as the first arguments (e.g. an alias doesn't work), but a wrapper script like
Code:
#!/bin/sh
wine /.../eac3to.exe "$@" -unix 2>/dev/null
# or use a wine cmdline arg to filter out the messages it always prints, instead of redirecting to /dev/null
would do the trick.

apparently, wine devs
don't want apps to detect wine, so the -unix option would be a nice feature. Or have a -nosound option, or a -quiet option. (OTOH, -quiet usually means no messages, not literally quiet.) I'm worried that -nosound would cause confusion, though, since someone might think it means exclude audio tracks. But it's -nosound, not -noaudio...

Anyway, that's my feature request, please and thank you.
pcordes is offline  
Old 1st December 2009, 02:09   #9592  |  Link
samepaul
Registered User
 
samepaul's Avatar
 
Join Date: May 2004
Posts: 25
I encounter problem. During THD -> AAC encoding the warning "Clipping detected, a 2nd pass will be necessary." is written and then eac3to indeed performs 2nd pass, making final audio very quiet and practically useless. Is there way to ignore clippings and avoid "2nd pass"?
Also eac3to writes "Reducing depth from 64 to 32 bits". What does it mean? Looks like volume related issue as well... if so, is it possible to prevent this downgrading too?
__________________
samepaul is offline  
Old 1st December 2009, 03:10   #9593  |  Link
tebasuna51
Moderator
 
tebasuna51's Avatar
 
Join Date: Feb 2005
Location: Spain
Posts: 6,890
Quote:
Originally Posted by samepaul View Post
I encounter problem. During THD -> AAC encoding the warning "Clipping detected, a 2nd pass will be necessary." is written and then eac3to indeed performs 2nd pass, making final audio very quiet and practically useless. Is there way to ignore clippings and avoid "2nd pass"?
You can add the parameter:
-no2ndpass
but is a very little diferent volume.

Quote:
Also eac3to writes "Reducing depth from 64 to 32 bits". What does it mean? Looks like volume related issue as well... if so, is it possible to prevent this downgrading too?
Is the size of the audio sample, trust in eac3to, NeroAacEnc don't support 64 bit float audio samples.
__________________
BeHappy, AviSynth audio transcoder.
tebasuna51 is offline  
Old 1st December 2009, 04:35   #9594  |  Link
samepaul
Registered User
 
samepaul's Avatar
 
Join Date: May 2004
Posts: 25
Quote:
Originally Posted by tebasuna51 View Post
You can add the parameter:
-no2ndpass
but is a very little diferent volume.
Thank you! It did the magic - volume remained good. Seems anti-clipping algorithm requires some improvements
__________________
samepaul is offline  
Old 1st December 2009, 21:26   #9595  |  Link
tebasuna51
Moderator
 
tebasuna51's Avatar
 
Join Date: Feb 2005
Location: Spain
Posts: 6,890
Quote:
Originally Posted by samepaul View Post
... Seems anti-clipping algorithm requires some improvements
Then there are big volume difference ...
Please, can you upload a sample?
__________________
BeHappy, AviSynth audio transcoder.
tebasuna51 is offline  
Old 1st December 2009, 22:33   #9596  |  Link
samepaul
Registered User
 
samepaul's Avatar
 
Join Date: May 2004
Posts: 25
Seems you're right. I tried to notch the place where clipping occurs, cut it out and encoded with and without 2nd pass.
There was difference in volume, but not so dramatic as it was when whole track was encoded. So I don't think that these samples are good for debugging.

I think it's not completely fault of encoder. I guess the track has places with really hard overload, but when eac3to performs 2nd pass it lowers volume of whole track evenly. Of course, after this previously clipped parts (blows, shootings etc) sound better, but "normal" parts where people talk become indiscernible. If I have tool to detect exact places where eac3to detects maximal clipping I would prepare sample for you. But without this I'd have to submit whole 700mb track , which I don't know how to do.
Btw, I've mistaken when I wrote "TrueHD". It was actually AC3.

But nevertheless, regarding to volume optimization, I think if eac3to performs 2 passes it can normalize levels non-linearly.
__________________
samepaul is offline  
Old 2nd December 2009, 01:06   #9597  |  Link
honai
Guest
 
Posts: n/a
Quote:
Originally Posted by samepaul View Post
I think if eac3to performs 2 passes it can normalize levels non-linearly.
... which is called Dynamic Range Compression (DRC) and is ignored by eac3to by default. DRC means you lose fidelity, and that is certainly not desirable if you want to preserve the quality of the original audio track.
 
Old 2nd December 2009, 01:38   #9598  |  Link
samepaul
Registered User
 
samepaul's Avatar
 
Join Date: May 2004
Posts: 25
Quote:
Originally Posted by honai View Post
... which is called Dynamic Range Compression (DRC) and is ignored by eac3to by default. DRC means you lose fidelity, and that is certainly not desirable if you want to preserve the quality of the original audio track.
Uhum... and making audio quiet so much that you hadly can distinguish what people say - this is what you call "high fidelity"? If so, I'd stick with normal fidelity

1st of all, DRC does not lower fidelity more than any volume adjustment, which is done by eac3to anyway. And really negligible comparing to quality loss due to audiocompression.
2nd, what I propose is not really DRC. Range compression only performed around overloaded zones, leaving ~90% (depending on content) of audiotrack intact.
__________________
samepaul is offline  
Old 2nd December 2009, 02:14   #9599  |  Link
IanD
Registered User
 
Join Date: Jun 2003
Location: Australia
Posts: 190
The switches -25.000 (source) and -ChangeTo24.000 imply that they are floating point values that "may" accept any floating point value, but in fact seem to be hardcoded to -23.976, -24.000 and -25.000 and nothing else is accepted.

Is there any way to define variable floating point values to finetune these parameters?

Is there an alternate way to perform non-standard resampling and time expansion/compression?
IanD is offline  
Old 2nd December 2009, 03:35   #9600  |  Link
tebasuna51
Moderator
 
tebasuna51's Avatar
 
Join Date: Feb 2005
Location: Spain
Posts: 6,890
Quote:
Originally Posted by IanD View Post
...
Is there any way to define variable floating point values to finetune these parameters?
To make the job eac3to use SSRC libraries to change the audio samplerate.
And not all conversion are possible. You can know available values here: http://avisynth.org/mediawiki/SSRC

Quote:
Is there an alternate way to perform non-standard resampling and time expansion/compression?
There are commercial software like TimeFactory

And also AviSynth methods like TimeStretch
__________________
BeHappy, AviSynth audio transcoder.
tebasuna51 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 09:35.


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