Help - Search - Members - Calendar
Full Version: AC-3 for stereo music-only DVD: good enough?
Hydrogenaudio Forums > Lossy Audio Compression > Other Lossy Codecs
rfa
I'd like to put all my CD music into music-only DVDs (more time per disc; better compatibility with DVD players than mp3/wma, etc.). Initially, I'd like to have lossless stereo music in the DVDs.

Following an advise got in EAC forum, I've downloaded "Audio DVD Creator" and found that it offers PCM@48kHz/16bit as default "lossless" audio format AND AC-3@192kbps/2.0ch as default lossy "high quality" audio format. AFAIK, AC-3 is as well supported by conventional DVD players as PCM@48kHz, so my question is: is it OK to use AC-3@192kbps? Is it transparent? Should I use a higher bit rate? Should I stick with PCM?

Thanks for any help.

Rosivaldo.
Leto Atreides II
If you want lossless then you need to do it with PCM. Whether or not AC-3 sounds "good enough" for you or not, only you can decide. In general I don't think AC-3 at 192kbps is considered transparent. You might also consider using some other 3rd party encoder that allows you to do AC-3 with a higher bitrate -- it can go up to 384kbps.
benski
It also depends on what AC-3 encoder you are using. The open-source ones produce very poor quality encodes compared to Dolby's commercial encoder, if I remember correctly.
AtomicGreymon
QUOTE(Leto Atreides II @ Nov 10 2007, 19:53) *

If you want lossless then you need to do it with PCM. Whether or not AC-3 sounds "good enough" for you or not, only you can decide. In general I don't think AC-3 at 192kbps is considered transparent. You might also consider using some other 3rd party encoder that allows you to do AC-3 with a higher bitrate -- it can go up to 384kbps.


Are you sure 384 is the highest? Sound Forge's AC-3 encoder gives me the option of up to 640kbps.
shadowking
Hmm.. Where are individual / public AC3 listening tests showing 2ch-192k is not transparent ?
wisodev
QUOTE(Leto Atreides II @ Nov 11 2007, 04:53) *
it can go up to 384kbps.


It can go up to 640 kbps.

QUOTE(benski @ Nov 11 2007, 05:51) *

It also depends on what AC-3 encoder you are using. The open-source ones produce very poor quality encodes compared to Dolby's commercial encoder, if I remember correctly.


ABX test please?
Leto Atreides II
QUOTE(wisodev @ Nov 11 2007, 01:31) *

QUOTE(Leto Atreides II @ Nov 11 2007, 04:53) *
it can go up to 384kbps.


It can go up to 640 kbps.


Actually, assuming we're talking about video DVDs we're both wrong:

http://en.wikipedia.org/wiki/DVD-Video#Audio_data

says that the max is 448kbps.

Other formats allow for higher bitrate AC3, but not normal DVDs.
j7n
Such a DVD should never serve as an archival copy. Therefore it would not matter that 224-256 kBit/s AC3 is a little lossy.

I wonder if standalone players generally handle 640 kBit/s streams or not. (Yeah Wikipedia says they should not.)
Bad Monkey
Going to have to upsample to 48KHz too though hmmmn?
rfa
I wanted lossless because I didn't realize the AC-3 possibility. My home theater has some limitations with mp3 and wma music that, I think, will not appear with AC-3 (despite some evidences, indeed I need to test, maybe today).

If I could have for sure that at some bit rate AC-3 will provide *really* transparent music, I would be glad to use it. So I'm asking to skillful people some help. :-)

Audio DVD Creator offers to me AC-3 to 2/5.1ch and 96-640kbps. For 5.1ch the default bit rate is 448kbps. For 2ch, it's 192kbps. Since I'm coping with CD music, I think I don't need 5.1ch. But if 192kbps isn't enough, how much would be?

Thanks any additional help.

Rosivaldo.

QUOTE(Bad Monkey @ Nov 11 2007, 03:03) *

Going to have to upsample to 48KHz too though hmmmn?


Forgive me, but I didn't understand... :-(

QUOTE(benski @ Nov 10 2007, 22:51) *

It also depends on what AC-3 encoder you are using. The open-source ones produce very poor quality encodes compared to Dolby's commercial encoder, if I remember correctly.


DVD Audio Creator docs doesn't mention the encoder used to create AC-3. May it be a serious issue?

Rosivaldo.

QUOTE(benski @ Nov 10 2007, 22:51) *

It also depends on what AC-3 encoder you are using. The open-source ones produce very poor quality encodes compared to Dolby's commercial encoder, if I remember correctly.


DVD Audio Creator doesn't mention the encoder used. May it be a serious issue?

Rosivaldo.

QUOTE(j7n @ Nov 11 2007, 02:52) *

Such a DVD should never serve as an archival copy. Therefore it would not matter that 224-256 kBit/s AC3 is a little lossy.


The lossless idea served two purposes: sound quality and archival. But archival may be done by other means. If I could have really transparent quality with AC-3, I would have the advantage or getting more time per DVD.

Best regards.

Rosivaldo.
j7n
I'd go with 256 kBit/s. It's no need to try to approach the lowest possible bitrate: a fair amount of space on that DVD will still be wasted in MPEG-2 PS overhead and video menus.

You could just make one DVD and listen to the AC-3 encodes on your computer. I'm not familiar with the way this program stores temporary files. But it should be possible to extract AC-3 streams from the created DVD and compare them to originals.

QUOTE
Forgive me, but I didn't understand... :-(

You can't store 44100 S/s on a Video DVD. If your source files are in this rate (i.e. ripped from audio CDs) then you'll have to upsample and even uncompressed PCM won't be mathematically lossless.

Even if DVD did support 44.1 kS/s, it would not be a wise choice for archival, similar how it is with redbook audio CDs. It takes great effort to properly demultiplex a DVD with many streams and there is no verification that the data is intact. I don't mean read errors, but skipped parts because of bogus de-/multiplexing.
rfa
QUOTE(j7n @ Nov 11 2007, 17:09) *

I'd go with 256 kBit/s. It's no need to try to approach the lowest possible bitrate: a fair amount of space on that DVD will still be wasted in MPEG-2 PS overhead and video menus.

(...)

You can't store 44100 S/s on a Video DVD. If your source files are in this rate (i.e. ripped from audio CDs) then you'll have to upsample and even uncompressed PCM won't be mathematically lossless.

Even if DVD did support 44.1 kS/s, it would not be a wise choice for archival, similar how it is with redbook audio CDs. It takes great effort to properly demultiplex a DVD with many streams and there is no verification that the data is intact. I don't mean read errors, but skipped parts because of bogus de-/multiplexing.


Ok, now I see that. Anyway, the upsampled music would not loose quality, would it? That is, even not being mathematically lossless, it would be really transparent, wouldn't it?

Rosivaldo.
j7n
Yes, I suppose. It is real hard or impossible to hear effects of resampling on music.
aharden
If the goal is transcoding for ease of playback, then I think AC-3 with a decent encoder at 224kbps or up should provide more than adequate results.

If the goal is archiving to protect against CD media failure, I think this is the wrong route. I'd much rather have a hard disk or DVD-R full of FLAC/CUE encodes than a 48kHz PCM DVD. With the DVD you might lose cue points, you would be resampling, and it would be tedious to reconstruct an audio CD from that data.
xmixahlx
did no one mention mp2?

using toolame or (twolame or whatever) would probably result in better quality than the free ac3 encoders.

and yes that is just speculating.
2Bdecided
Supposedly early DVD players in the USA couldn't do mp2. It certainly wasn't mandated in the spec.

It's still not required for American DVD players (AC-3 is), only for European ones - but global manufacturing etc mean that it's normally decoded in all players.


Dolby themselves recommend 256kbps for stereo AC-3 - they've tried (unsuccessfully) to persuade all the software vendors to move up from 192kbps as the default. Some HD broadcasts in the USA are using 384kbps stereo.

Cheers,
David.
This is a "lo-fi" version of our main content. To view the full version with more information, formatting and images, please click here.
Invision Power Board © 2001-2009 Invision Power Services, Inc.