Vorbis 1.0 and BeSweet's 1.4RC4 gain

[mikeson]

Hi folks,

I've got problem with BeSweet 1.4RC4 and AC3->OGG transcoding. The poblem is in audio gain. When I use -ota( -g max ), final ogg has no audio gain although BeSweet found for example 7.1db gain (not using post gain). But when transcoding to MP3, everything is fine. I'm using latest Tobias DS filters when playing with WMP and Nullsoft Vorbis Decoder 1.2 when playing with Winamp.

Here is command line:

BeSweet -core( -input "j:\Matrix AC3 T01 3_2ch 384Kbps DELAY -103ms.ac3" -output "e:\DivX\Matrix (DS2 -q 6.00).ogg" -logfilea besweet.log ) -azid( -z1 -b1 -c normal -L -3db -F wav32 -s surround2 ) -ota( -d -103 -g max ) -ogg( -q 6.00 )

Here is log file:
BeSweet v1.4RC4 by DSPguru.
--------------------------
Using azid.dll v1.8 (b825) by Midas (midas@egon.gyaloglo.hu).
Using libVorbis.dll v1.0 ( Jul 19 2002 ) by John33 (www.inf.ufpr.br/~rja00).

Logging start : 07/24/02 , 02:55:07.

BeSweet -core( -input j:\Matrix AC3 T01 3_2ch 384Kbps DELAY -103ms.ac3 -output e:\DivX\Matrix (DS2 -q 6.00).ogg -logfilea besweet.log ) -azid( -z1 -b1 -c normal -L -3db -F wav32 -s surround2 ) -ota( -g max -d -103 ) -ogg( -q 6.00 )

[00:00:00:000] +------- BeSweet -----
[00:00:00:000] | Input : j:\Matrix AC3 T01 3_2ch 384Kbps DELAY -103ms.ac3
[00:00:00:000] | Output: e:\DivX\Matrix (DS2 -q 6.00).ogg
[00:00:00:000] | Floating-Point Process: Yes
[00:00:00:-103] | Overall Track Gain: 12.1dB
[00:00:00:-103] +-------- AZID -------
[00:00:00:-103] | Output Stereo mode: Dolby Surround 2 compatible
[00:00:00:-103] | Total Gain: 0.0dB, Compression: Normal
[00:00:00:-103] | LFE levels: To LR -3.0dB, To LFE 0.0dB
[00:00:00:-103] | Center mix level: BSI
[00:00:00:-103] | Surround mix level: BSI
[00:00:00:-103] | Dialog normalization: No
[00:00:00:-103] | Rear channels filtering: No
[00:00:00:-103] | Source Sample-Rate: 48.0KHz
[00:00:00:-103] +-------- OGG --------
[00:00:00:-103] | VBR Quality : 0.600
[00:00:00:-103] +---------------------
[02:10:41:337] Conversion Completed !
[02:10:41:337] Actual Avg. Bitrate : 154kbps
[00:31:20:000] <-- Transcoding Duration
Logging ends : 07/24/02 , 03:26:27.

On the other hand, using -azid( --maximize ) works, but it gives me some overflows. For example when finding maximum gain in Forrest Gump -ota( -g max ) gives me 10.4db and -azid ( --maximize ) gives me 19.2db. It is big difference, isn't it?

So your suggestions?

Mikeson

[DSPguru]

post all the 3 logfiles.
using -ota( -g max )
using -ota( -G max )
using -azid( --maximize )

[mikeson]

@DSPguru

Ok, but it will take a while because I've got to transcode it again...

So please be patient...

Mikeson

[DSPguru]

no rush :)

[mikeson]

@DSPguru

So here are log files:

BeSweet v1.4RC4 by DSPguru.
--------------------------
Using azid.dll v1.8 (b825) by Midas (midas@egon.gyaloglo.hu).
Using libVorbis.dll v1.0 ( Jul 19 2002 ) by John33 (www.inf.ufpr.br/~rja00).

Logging start : 07/27/02 , 13:04:08.

BeSweet -core( -input j:\Forrest Gump AC3 T01 3_2ch 448Kbps DELAY 0ms.ac3 -output e:\DivX\Forrest Gump (DS2 -q 6.00) -azid( --maximize ).ogg -logfilea 1.log ) -azid( -z1 -b1 -c normal -L -3db -F wav32 -s surround2 --maximize ) -ogg( -q 6.00 )

[00:00:00:000] +------- BeSweet -----
[00:00:00:000] | Input : j:\Forrest Gump AC3 T01 3_2ch 448Kbps DELAY 0ms.ac3
[00:00:00:000] | Output: e:\DivX\Forrest Gump (DS2 -q 6.00) -azid( --maximize ).ogg
[00:00:00:000] | Floating-Point Process: Yes
[00:00:00:000] +-------- AZID -------
[00:00:00:000] | Output Stereo mode: Dolby Surround 2 compatible
[00:00:00:000] | Total Gain: 19.2dB, Compression: Normal
[00:00:00:000] | LFE levels: To LR -3.0dB, To LFE 0.0dB
[00:00:00:000] | Center mix level: BSI
[00:00:00:000] | Surround mix level: BSI
[00:00:00:000] | Dialog normalization: No
[00:00:00:000] | Rear channels filtering: No
[00:00:00:000] | Source Sample-Rate: 48.0KHz
[00:00:00:000] +-------- OGG --------
[00:00:00:000] | VBR Quality : 0.600
[00:00:00:000] +---------------------
[02:16:16:992] Conversion Completed !
[02:16:16:992] Actual Avg. Bitrate : 166kbps
[00:44:48:000] <-- Transcoding Duration
Logging ends : 07/27/02 , 13:48:56.


BeSweet v1.4RC4 by DSPguru.
--------------------------
Using azid.dll v1.8 (b825) by Midas (midas@egon.gyaloglo.hu).
Using libVorbis.dll v1.0 ( Jul 19 2002 ) by John33 (www.inf.ufpr.br/~rja00).

Logging start : 07/27/02 , 13:48:56.

BeSweet -core( -input j:\Forrest Gump AC3 T01 3_2ch 448Kbps DELAY 0ms.ac3 -output e:\DivX\Forrest Gump (DS2 -q 6.00) -ota( -g max ) -logfilea 2.log ) -azid( -z1 -b1 -c normal -L -3db -F wav32 -s surround2 ) -ota( -g max ) -ogg( -q 6.00 )

[00:00:00:000] +------- BeSweet -----
[00:00:00:000] | Input : j:\Forrest Gump AC3 T01 3_2ch 448Kbps DELAY 0ms.ac3
[00:00:00:000] | Output: e:\DivX\Forrest Gump (DS2 -q 6.00) -ota( -g max )
[00:00:00:000] | Floating-Point Process: Yes
[00:00:00:000] | Overall Track Gain: 13.1dB
[00:00:00:000] +-------- AZID -------
[00:00:00:000] | Output Stereo mode: Dolby Surround 2 compatible
[00:00:00:000] | Total Gain: 0.0dB, Compression: Normal
[00:00:00:000] | LFE levels: To LR -3.0dB, To LFE 0.0dB
[00:00:00:000] | Center mix level: BSI
[00:00:00:000] | Surround mix level: BSI
[00:00:00:000] | Dialog normalization: No
[00:00:00:000] | Rear channels filtering: No
[00:00:00:000] | Source Sample-Rate: 48.0KHz
[00:00:00:000] +-------- OGG --------
[00:00:00:000] | VBR Quality : 0.600
[00:00:00:000] +---------------------
[02:16:16:992] Conversion Completed !
[02:16:16:992] Actual Avg. Bitrate : 148kbps
[00:41:16:000] <-- Transcoding Duration
Logging ends : 07/27/02 , 14:30:12.


BeSweet v1.4RC4 by DSPguru.
--------------------------
Using azid.dll v1.8 (b825) by Midas (midas@egon.gyaloglo.hu).
Using libVorbis.dll v1.0 ( Jul 19 2002 ) by John33 (www.inf.ufpr.br/~rja00).

Logging start : 07/27/02 , 14:30:12.

BeSweet -core( -input j:\Forrest Gump AC3 T01 3_2ch 448Kbps DELAY 0ms.ac3 -output e:\DivX\Forrest Gump (DS2 -q 6.00) -ota( -G max).ogg -logfilea 3.log ) -azid( -z1 -b1 -c normal -L -3db -F wav32 -s surround2 ) -ota( -G max ) -ogg( -q 6.00 )

[00:00:00:000] +------- BeSweet -----
[00:00:00:000] | Input : j:\Forrest Gump AC3 T01 3_2ch 448Kbps DELAY 0ms.ac3
[00:00:00:000] | Output: e:\DivX\Forrest Gump (DS2 -q 6.00) -ota( -G max).ogg
[00:00:00:000] | Floating-Point Process: Yes
[00:00:00:000] +-------- AZID -------
[00:00:00:000] | Output Stereo mode: Dolby Surround 2 compatible
[00:00:00:000] | Total Gain: 0.0dB, Compression: Normal
[00:00:00:000] | LFE levels: To LR -3.0dB, To LFE 0.0dB
[00:00:00:000] | Center mix level: BSI
[00:00:00:000] | Surround mix level: BSI
[00:00:00:000] | Dialog normalization: No
[00:00:00:000] | Rear channels filtering: No
[00:00:00:000] | Source Sample-Rate: 48.0KHz
[00:00:00:000] +-------- OGG --------
[00:00:00:000] | VBR Quality : 0.600
[00:00:00:000] +---------------------
[02:16:16:992] Gain of 13.1dB had been asserted to file.
[02:16:16:992] Conversion Completed !
[02:16:16:992] Actual Avg. Bitrate : 148kbps
[00:27:55:000] <-- Transcoding Duration
Logging ends : 07/27/02 , 14:58:07.


File sizes are:
Forrest Gump (DS2 -q 6.00) -ota( -G max ).ogg - 152,038,638 bytes
Forrest Gump (DS2 -q 6.00) -ota( -g max ).ogg - 152,038,870 bytes
Forrest Gump (DS2 -q 6.00) -azid( --maximize ).ogg - 169,991,354 bytes

1st and 2nd oggs are very silent, the 3rd is ok and loud enough, but I'm afraid of overflows. But transcoded to mp3 with -ota( -g max ) it is ok, so I really don't know. Am I doing something wrong? :confused:

Mikeson

[DSPguru]

1. i don't see overflows in the logfiles. can you hear overflows/clippings.. ?
2. your logfiles order are different from the summary order, what do you call '1st', '2nd' & '3rd' ogg ?

[mikeson]

@DSPGuru

1. Yes, when I converted file from ogg to mp3 to read it in Cool Edit Pro, there were some parts that where clipped.
Anyway why is -ota ( -g max ) gain 13.1dB and -azid ( --maximize ) gain 19.2dB? Which gain is correct? BeSweet's or Azid's?

2.
1st = Forrest Gump (DS2 -q 6.00) -ota( -G max ).ogg - 152,038,638 bytes (silent)
2nd = Forrest Gump (DS2 -q 6.00) -ota( -g max ).ogg - 152,038,870 bytes (silent)
3rd = Forrest Gump (DS2 -q 6.00) -azid( --maximize ).ogg - 169,991,354 bytes (loud enough but possible clipping/overflows)
Sorry for confusing...

Mikeson

[DSPguru]

Originally posted by mikeson
@DSPGuru

1. Yes, when I converted file from ogg to mp3 to read it in Cool Edit Pro, there were some parts that where clipped.let's just hope those clips weren't added by the ogg->mp3 process.
Anyway why is -ota ( -g max ) gain 13.1dB and -azid ( --maximize ) gain 19.2dB? Which gain is correct? BeSweet's or Azid's?could be a bug with DS2 azid's maxgain seeking routines, i'll check it out.
anyway, i would bet that BeSweet's gain value is the right one.
if you find that the soundtrack isn't loud enough after normalizing, you might wanna try asserting a harder dynamic compression.

Dg.

[mikeson]

@DSPGuru

let's just hope those clips weren't added by the ogg->mp3 process.

I don't think so because overal track was much louder than it should be (but this is only personal opinion).

But I can say:
-ota ( -g max ) with mp3 - perfect, no clipping/overflows
-ota ( -g max ) with ogg - find maxiumum gain, but it seems to me like keep gain untouched :confused:
-azid ( --maximize ) - the same with ogg or mp3 (find and insert maximum gain), but gain is much higher than with -ota ( -g max )

So which one do you think is correct? I think that BeSweet's is correct, but as I said, it doesn't work with ogg (for me).
So what about (but this is only workaround):
1. Find maximum gain with -ota ...
2. Insert this gain to target file with -azid( -g ...

could be a bug with DS2 azid's maxgain seeking routines, i'll check it out.

It is the same with DS downmix, not only DS2

[DSPguru]

Originally posted by mikeson
It is the same with DS downmix, not only DS2 in that case, i believe you found a bug.
will check it out, whenever possible.


thank you, mike !

[mikeson]

@DSPguru

Hi,

How's going Vorbis gain issue solution?

[DSPguru]

i haven't forgot that.. :)

[canadian_fbi]

i know this is totally unrelated to this thread (:)) but has the delay issue in the equalizer been fixed? i saw "Fixed tiny issue with equalizer" in the bug fixes, but was wondering what exactly that entailed. looking forward to trying that out along with rehuff in the new version.

p.s.: is it just me, or does that smiley face inside parentheses kind of look like a smiley wearing headphones (that are too big for its head and falling off)?

[DSPguru]

should be, indeed, fixed.

[canadian_fbi]

ok, i should give it a go a bit later this week. :) i spent a good hour and a half today trying to get right of a dead spot for bass frequencies right where i sit in front of the speakers/tv... tried moving around the speakers as much as i could given the fact that there are doors against the walls on either side of them. finally i came to the conclusion that i think it's a standing wave between the ceiling and the hardwood floors. really annoyed me today. anyway, this is all off-topic, just wanted to vent :)

[DSPguru]

why won't you share your EQ file ? (+ autocad/visio drawing with the design of your room.. ;))

[canadian_fbi]

which eq file, the lowpass one? i can attach it or upload it to my site if you want - i don't even know if i did it correctly or not. :) i believe i just set it to do -300 dB between 17500-26000 hz. which i'm sure is overkill, but i figured i'd be on the safe side.

and my room is extra complicated, since the speakers face a large opening to the dining room which brings in all sorts of extra possibilities for weird interference, and there are doors in three corners of the room that don't give me many options for moving the speakers or even rearranging where things go or face.

[DSPguru]

Originally posted by mikeson
@DSPguru

Hi,

How's going Vorbis gain issue solution? should be fixed in RC7.

[mikeson]

@DSPguru

Great! It works now! :)

Thanks

[DSPguru]

Originally posted by mikeson
@DSPguru

Great! It works now! :)

Thanks Great :)

[canadian_fbi]

hey dg, just wanted to let you know that i had a 16 ms delay again with your latest version of besweet. relevant oggdec stuff and logfiles follow (nevermind that i mistakenly encoded the wrong ac3 file :))

BeSweet v1.4RC7 by DSPguru.
--------------------------
Using azid.dll v1.8 (b825) by Midas (midas@egon.gyaloglo.hu).
Using libVorbis.dll v1.0 ( Jul 19 2002 ) by John33 (www.inf.ufpr.br/~rja00).

Logging start : 08/10/02 , 17:32:07.

C:\Video\Tools\BeSweet\BeSweet.exe -core( -input C:\Video\DVD ripping\planet of the apes AC3 T04 2_0ch 192Kbps DELAY 0ms.ac3 -output C:\Video\DVD ripping\planet of the apes q 4.99 normal.ogg -logfilea BeSweet.log ) -azid( -g max -c normal -s stereo ) -ogg( -q 0.499 ) -profile( The OggMachine v0.5 )

[00:00:00:000] +------- BeSweet -----
[00:00:00:000] | Input : C:\Video\DVD ripping\planet of the apes AC3 T04 2_0ch 192Kbps DELAY 0ms.ac3
[00:00:00:000] | Output: C:\Video\DVD ripping\planet of the apes q 4.99 normal.ogg
[00:00:00:000] | Floating-Point Process: Yes
[00:00:00:000] +-------- AZID -------
[00:00:00:000] | Output Stereo mode: Stereo
[00:00:00:000] | Total Gain: 3.3dB, Compression: Normal
[00:00:00:000] | LFE levels: To LR -INF, To LFE 0.0dB
[00:00:00:000] | Center mix level: BSI
[00:00:00:000] | Surround mix level: BSI
[00:00:00:000] | Dialog normalization: No
[00:00:00:000] | Rear channels filtering: No
[00:00:00:000] | Source Sample-Rate: 48.0KHz
[00:00:00:000] +-------- OGG --------
[00:00:00:000] | VBR Quality : 0.499
[00:00:00:000] +---------------------
[01:59:57:472] Conversion Completed !
[01:59:57:472] Actual Avg. Bitrate : 138kbps
[00:20:57:000] <-- Transcoding Duration
Logging ends : 08/10/02 , 17:53:04.

BeSweet v1.4RC7 by DSPguru.
--------------------------
Using azid.dll v1.8 (b825) by Midas (midas@egon.gyaloglo.hu).
Using Shibatch.dll v0.15 by Naoki Shibata & DSPguru (shibatch.sourceforge.net).
Using libVorbis.dll v1.0 ( Jul 19 2002 ) by John33 (www.inf.ufpr.br/~rja00).

Logging start : 08/10/02 , 18:01:29.

C:\Video\Tools\BeSweet\BeSweet.exe -core( -input C:\Video\DVD ripping\planet of the apes AC3 T04 2_0ch 192Kbps DELAY 0ms.ac3 -output C:\Video\DVD ripping\planet of the apes q 4.99 lowpass.ogg -logfilea BeSweet.log ) -azid( -g max -c normal -s stereo ) -ogg( -q 0.499 ) -shibatch( -equ C:\Video\Tools\Besweet\17500 lowpass.eq )

[00:00:00:000] +------- BeSweet -----
[00:00:00:000] | Input : C:\Video\DVD ripping\planet of the apes AC3 T04 2_0ch 192Kbps DELAY 0ms.ac3
[00:00:00:000] | Output: C:\Video\DVD ripping\planet of the apes q 4.99 lowpass.ogg
[00:00:00:000] | Floating-Point Process: Yes
[00:00:00:000] +-------- AZID -------
[00:00:00:000] | Output Stereo mode: Stereo
[00:00:00:000] | Total Gain: 3.3dB, Compression: Normal
[00:00:00:000] | LFE levels: To LR -INF, To LFE 0.0dB
[00:00:00:000] | Center mix level: BSI
[00:00:00:000] | Surround mix level: BSI
[00:00:00:000] | Dialog normalization: No
[00:00:00:000] | Rear channels filtering: No
[00:00:00:000] +------ Shibatch -----
[00:00:00:000] | Equ Preset File : C:\Video\Tools\Besweet\17500 lowpass.eq
[00:00:00:000] | Equ Quality Level : 0.8
[00:00:00:000] | Source Sample-Rate: 48.0KHz
[00:00:00:000] +-------- OGG --------
[00:00:00:000] | VBR Quality : 0.499
[00:00:00:000] +---------------------
[01:59:57:344] Conversion Completed !
[01:59:57:344] Actual Avg. Bitrate : 136kbps
[00:23:56:000] <-- Transcoding Duration
Logging ends : 08/10/02 , 18:25:25.

- snip snip -

Bitstream is 2 channel, 48000Hz
Scale = 1.0000
Decoded length: 345478656 samples = 119:57 mins.
Encoded by: Xiph.Org libVorbis I 20020717
Decoding: planet of the apes q 4.99 normal.ogg

- snip snip -

Bitstream is 2 channel, 48000Hz
Scale = 1.0000
Decoded length: 345477888 samples = 119:57 mins.
Encoded by: Xiph.Org libVorbis I 20020717
Decoding: planet of the apes q 4.99 lowpass.ogg

actually... looking at the logfile now, it looks like the logfile and the decoder differ on the difference of the length of the two files... unless i got my math all wrong.

on a different note, i used rehuff in another test, and didn't experience any of the problems people have mentioned in the past. am i to take it that the bugs that were fixed in the latest patch fixed all the known issues and that it's now working and stable, or something like it?

[DSPguru]

are you Sure the delay is still there ?
would you mind checking this by comparing the two decoded streams on a visual signal editor (cooledit, etc'..) ..

[canadian_fbi]

eek... you're right. using cooledit i can confirm that there's no delay both at the beginning and an hour in. well, that's good news then. :) between that and rehuff, you could save something like 5% on each encoded file with no noticeable quality loss.

[DSPguru]

Originally posted by canadian_fbi
eek... you're right. using cooledit i can confirm that there's no delay both at the beginning and an hour in. well, that's good news then. :) between that and rehuff, you could save something like 5% on each encoded file with no noticeable quality loss. good news indeed :D.