Transcoding

Started by Gurn Blanston, May 28, 2015, 05:36:01 PM

Previous topic - Next topic

0 Members and 1 Guest are viewing this topic.

Gurn Blanston

I have some files which are .M4A (MP4), which I didn't rip myself. I can't say if they are lossless or lossy, all I can say is the bitrate ranges from 260-270kbps. If they were MP3 I would say they were VBR, but for all I know, M4A files are all VBR and this variation is to be expected. My FLAC rips are all in the 600-700kbps range, but there again, maybe M4A lossless shows a lower bitrate and yet has equal quality. I don't know.

I have some playback equipment which only plays MP3 and I would like to make a nice little MP3 file to put on it. I want to do this in the way which will cause the least degradation of file quality. Should I blow these out to WAV and recompress the WAV file?  Or should I directly create MP3 VBR files from the M4A's?  Something else altogether??

Input welcome. :)

8)
Visit my Haydn blog: HaydnSeek

Haydn: that genius of vulgar music who induces an inordinate thirst for beer - Mily Balakirev (1860)

Ken B

Quote from: Gurn Blanston on May 28, 2015, 05:36:01 PM
I have some files which are .M4A (MP4), which I didn't rip myself. I can't say if they are lossless or lossy, all I can say is the bitrate ranges from 260-270kbps. If they were MP3 I would say they were VBR, but for all I know, M4A files are all VBR and this variation is to be expected. My FLAC rips are all in the 600-700kbps range, but there again, maybe M4A lossless shows a lower bitrate and yet has equal quality. I don't know.

I have some playback equipment which only plays MP3 and I would like to make a nice little MP3 file to put on it. I want to do this in the way which will cause the least degradation of file quality. Should I blow these out to WAV and recompress the WAV file?  Or should I directly create MP3 VBR files from the M4A's?  Something else altogether??

Input welcome. :)

8)

GLBTFLACMP3WAV, whatever floats your boat Gurn. As long as its consensual I say go for it.

Gurn Blanston

Quote from: Ken B on May 28, 2015, 06:32:11 PM
GLBTFLACMP3WAV, whatever floats your boat Gurn. As long as its consensual I say go for it.

Surprisingly little that doesn't float my boat.  0:) It's the Age of the Acronym, Ken! :D

8)
Visit my Haydn blog: HaydnSeek

Haydn: that genius of vulgar music who induces an inordinate thirst for beer - Mily Balakirev (1860)

Pat B

Quote from: Gurn Blanston on May 28, 2015, 05:36:01 PM
I have some files which are .M4A (MP4), which I didn't rip myself. I can't say if they are lossless or lossy, all I can say is the bitrate ranges from 260-270kbps. If they were MP3 I would say they were VBR, but for all I know, M4A files are all VBR and this variation is to be expected. My FLAC rips are all in the 600-700kbps range, but there again, maybe M4A lossless shows a lower bitrate and yet has equal quality. I don't know.

I have some playback equipment which only plays MP3 and I would like to make a nice little MP3 file to put on it. I want to do this in the way which will cause the least degradation of file quality. Should I blow these out to WAV and recompress the WAV file?  Or should I directly create MP3 VBR files from the M4A's?  Something else altogether??

1. The bitrate of ~270kbps indicates it is lossy.

2. It's conceivable that a transcoder might do something smart when converting directly from one lossy format to another, but it probably doesn't make any difference. But I can't imagine that you would get any benefit from converting to WAV.

You probably already know this, but: use VBR and don't drop the bitrate much.

71 dB

Why not convert to 320 kbps mp3 ? That should mean least degration of sound quality.
Spatial distortion is a serious problem deteriorating headphone listening.
Crossfeeders reduce spatial distortion and make the sound more natural
and less tiresome in headphone listening.

My Sound Cloud page <-- NEW July 2025 "Liminal Feelings"

Peter Power Pop

Quote from: Gurn Blanston on May 28, 2015, 05:36:01 PM
I have some files which are .M4A (MP4), which I didn't rip myself. I can't say if they are lossless or lossy, all I can say is the bitrate ranges from 260-270kbps. If they were MP3 I would say they were VBR, but for all I know, M4A files are all VBR and this variation is to be expected. My FLAC rips are all in the 600-700kbps range, but there again, maybe M4A lossless shows a lower bitrate and yet has equal quality. I don't know.

I have some playback equipment which only plays MP3 and I would like to make a nice little MP3 file to put on it. I want to do this in the way which will cause the least degradation of file quality. Should I blow these out to WAV and recompress the WAV file?  Or should I directly create MP3 VBR files from the M4A's?  Something else altogether??

Input welcome. :)

8)

My suggestion would be to save yourself a whole heap o' time and convert the M4A files directly to MP3 (at 320 kbps CBR). If you convert an M4A to a WAV file, the audio data in the WAV file will be the same as the M4A anyway, so converting from WAV will just be an unnecessary step (as far as I can see).

Gurn Blanston

Quote from: Pat B on May 28, 2015, 07:16:46 PM
1. The bitrate of ~270kbps indicates it is lossy.

2. It's conceivable that a transcoder might do something smart when converting directly from one lossy format to another, but it probably doesn't make any difference. But I can't imagine that you would get any benefit from converting to WAV.

You probably already know this, but: use VBR and don't drop the bitrate much.

Hey, Pat,
Thought that might be the case, but I've been wrong applying FLAC/MP3 knowledge to other formats.  I always use VBR at the highest rate I can set (what LAME used to call insane!), so that is a given. I want to do your step 2, I was just making sure I wasn't going to do too much damage opposed to some other, unknown to me, way. :)

Quote from: 71 dB on May 28, 2015, 08:55:27 PM
Why not convert to 320 kbps mp3 ? That should mean least degration of sound quality.

VBR maxed out should give the same result with a smaller file, unless the algorithm tosses out some stuff... :-\  Since it has a max setting of 320, it shouldn't toss anything that straight 320 would have kept, I believe.

Quote from: Peter Power Pop on May 28, 2015, 09:29:07 PM
My suggestion would be to save yourself a whole heap o' time and convert the M4A files directly to MP3 (at 320 kbps CBR). If you convert an M4A to a WAV file, the audio data in the WAV file will be the same as the M4A anyway, so converting from WAV will just be an unnecessary step (as far as I can see).

Peter,
And ultimately, I believe this is exactly what I will do. What I was mainly trolling for here (and I got a couple of PM's too) was some indication that M4A might have some feature which would change my way of thinking. Apparently it doesn't, since ultimately everyone is suggesting I do what I was wanting to do anyway. :)

Thanks all for your input,
8)
Visit my Haydn blog: HaydnSeek

Haydn: that genius of vulgar music who induces an inordinate thirst for beer - Mily Balakirev (1860)

merlin

PMFJI, but I am wondering if there is any advantage to converting flac to wav before burning to cd?  My burner, Brasero, automatically converts them, but I wonder if there might be some loss of information that might be otherwise saved by doing the conversion first?

Gurn Blanston

Quote from: merlin on May 29, 2015, 10:23:22 AM
PMFJI, but I am wondering if there is any advantage to converting flac to wav before burning to cd?  My burner, Brasero, automatically converts them, but I wonder if there might be some loss of information that might be otherwise saved by doing the conversion first?

FLAC and WAV are both lossless, so there will be no data loss. A lot of audio burning software uses WAV as its native format, so this is why they decode from FLAC. No big deal for you there.  :)

8)
Visit my Haydn blog: HaydnSeek

Haydn: that genius of vulgar music who induces an inordinate thirst for beer - Mily Balakirev (1860)

merlin

Thanks, Gurn.  Much appreciated!

71 dB

Quote from: Gurn Blanston on May 29, 2015, 04:30:04 AM
VBR maxed out should give the same result with a smaller file, unless the algorithm tosses out some stuff... :-\  Since it has a max setting of 320, it shouldn't toss anything that straight 320 would have kept, I believe.

At a given bitrate, VBR is better than ABR, which is better than CBR.

Why I always use CBR is because I have had problems playing VBR mp3 files and CBR is "foolproof". The music in my portable player is coded at 192 kps CBR, because higher quality is pretty much unnecessory thanks to noisy listening environment (mostly bus). No joint stereo but always pure stereo coding! Also, I pre-crossfeed the music with Audacity. Crossfeeding is 100 times more important than high bitrate. I rather listen to 128 kbps mp3s crossfed than lossless files not crossfed. Yes, spatial distortion is that bad of a thing in headphone listening.

In fact, I believe much of the difference between bitrates is due to how spatial distortion has "survived" lossy encoding. Crossfeeding spatial distortion away makes the problems of low bitrates go away in this regard, making the negative effect of low bitrate on sound less dramatic.

Lossy encoders ALWAYS toss something, it is just a question of how much. Whenever VBR is under 320 kbps, it has worse sound quality than CBR at 320 kbps, of course.

How can you make VBR mp3 files with maximum bitrate of 320 kbps? In Audacity the "Best Quality" VBR is 220-260 kbps. ABR goes up to 320 kbps being theoretically better than CBR 320 kbps.
Spatial distortion is a serious problem deteriorating headphone listening.
Crossfeeders reduce spatial distortion and make the sound more natural
and less tiresome in headphone listening.

My Sound Cloud page <-- NEW July 2025 "Liminal Feelings"

merlin

Hi 71 db.  I cannot seem to locate the crossfeed feature in Audacity for Linux, nor in the manual.  Clearly I am missing something!

Can you please point me in the right direction?  Thanks!

Gurn Blanston

#12
Quote from: 71 dB on May 29, 2015, 01:38:34 PM
At a given bitrate, VBR is better than ABR, which is better than CBR.

Why I always use CBR is because I have had problems playing VBR mp3 files and CBR is "foolproof". The music in my portable player is coded at 192 kps CBR, because higher quality is pretty much unnecessory thanks to noisy listening environment (mostly bus). No joint stereo but always pure stereo coding! Also, I pre-crossfeed the music with Audacity. Crossfeeding is 100 times more important than high bitrate. I rather listen to 128 kbps mp3s crossfed than lossless files not crossfed. Yes, spatial distortion is that bad of a thing in headphone listening.

In fact, I believe much of the difference between bitrates is due to how spatial distortion has "survived" lossy encoding. Crossfeeding spatial distortion away makes the problems of low bitrates go away in this regard, making the negative effect of low bitrate on sound less dramatic.

Lossy encoders ALWAYS toss something, it is just a question of how much. Whenever VBR is under 320 kbps, it has worse sound quality than CBR at 320 kbps, of course.

How can you make VBR mp3 files with maximum bitrate of 320 kbps? In Audacity the "Best Quality" VBR is 220-260 kbps. ABR goes up to 320 kbps being theoretically better than CBR 320 kbps.

This is true. However, when you look at a file's details, one which is encoded in VBR, that is, the 'bitrate ' number is an average bitrate, which is why file are mainly different from each other. And they have internally preset limits. When you choose the highest setting for VBR (that is, 'extreme'), it tells you that it will show a bitrate around 250. But this isn't a constant number, it is an average of all the samples made in a rip. Depending upon the complexity of the music, dynamics, frequency etc., the top end is 320. When it doesn't need to be 320, it is less. All those moments of silence and narrow frequency and constant dynamic range etc.. So the bottom line is that the things it is tossing out are the bits you can't hear anyway. And unlike 320 CBR, which is still 320 when it is dead quiet, this is sampling at, maybe 64 when it is dead quiet.

The reason I know this is that I have been doing this stuff for a long time. I got my first LAME encoder in 1999, and back then, when VBR was virtually an unknown quantity, you didn't rely on internal presets; you could actually type in the values for high and low. I used 320 and 64. When VBR got popular, that is, when people recognized that it gave the best ratio of quality to file size, everyone started using it. (now I will be mean) In order to stop the huge flow of WTF?'s from all the ignorant people who refused to read about it and learn, the consortium who write LAME simply built the options in as presets and took away control. The net result is that the files I wrote 10 years ago are slightly larger than the files I wrote now, because I set my high and low limits to higher values than the ones which the presets are today.

Back then I used MusicMatch Jukebox as a front end for LAME. This was before Yahoo bought them out and totally destroyed them. It was unsurpassed, even by anything I have today (dBpoweramp, XRECode, Audacity). Sometimes, the more we go forward, the further behind we fall. :-\

8)

PS - Forgot to add; your complaint of some things not handling VBR is no longer valid. Everything handles VBR seamlessly nowadays. It was certainly true at one time, but not any more. Indulge yourself!  :)
Visit my Haydn blog: HaydnSeek

Haydn: that genius of vulgar music who induces an inordinate thirst for beer - Mily Balakirev (1860)

71 dB

Quote from: merlin on May 29, 2015, 05:07:43 PM
Hi 71 db.  I cannot seem to locate the crossfeed feature in Audacity for Linux, nor in the manual.  Clearly I am missing something!

Can you please point me in the right direction?  Thanks!

The crossfeed feature is a nyquist-plugin 'XFEEDER' I wrote for Audacity. I can send you the plugin if you want. You simply add the file among other .ny-plugins (on macs Applications/Audacity/plug-ins/... )and next time you start Audacity, it has been added to the effect list.
Spatial distortion is a serious problem deteriorating headphone listening.
Crossfeeders reduce spatial distortion and make the sound more natural
and less tiresome in headphone listening.

My Sound Cloud page <-- NEW July 2025 "Liminal Feelings"

71 dB

Quote from: Gurn Blanston on May 29, 2015, 06:17:15 PM
This is true. However, when you look at a file's details, one which is encoded in VBR, that is, the 'bitrate ' number is an average bitrate, which is why file are mainly different from each other. And they have internally preset limits. When you choose the highest setting for VBR (that is, 'extreme'), it tells you that it will show a bitrate around 250. But this isn't a constant number, it is an average of all the samples made in a rip. Depending upon the complexity of the music, dynamics, frequency etc., the top end is 320. When it doesn't need to be 320, it is less. All those moments of silence and narrow frequency and constant dynamic range etc.. So the bottom line is that the things it is tossing out are the bits you can't hear anyway. And unlike 320 CBR, which is still 320 when it is dead quiet, this is sampling at, maybe 64 when it is dead quiet.
This is correct. Yes, the upper limit of VBR is 320 kbps. However, lossy encoding is lossy encoding even when you don't hear the diffencence. It is there and may become audible when you edit the sound file somehow (dynamic compression, filtering, etc.)

Quote from: Gurn Blanston on May 29, 2015, 06:17:15 PMThe reason I know this is that I have been doing this stuff for a long time. I got my first LAME encoder in 1999, and back then, when VBR was virtually an unknown quantity, you didn't rely on internal presets; you could actually type in the values for high and low. I used 320 and 64. When VBR got popular, that is, when people recognized that it gave the best ratio of quality to file size, everyone started using it. (now I will be mean) In order to stop the huge flow of WTF?'s from all the ignorant people who refused to read about it and learn, the consortium who write LAME simply built the options in as presets and took away control. The net result is that the files I wrote 10 years ago are slightly larger than the files I wrote now, because I set my high and low limits to higher values than the ones which the presets are today.
This all sound very logical. This is how it always goes.  ;)

Quote from: Gurn Blanston on May 29, 2015, 06:17:15 PMBack then I used MusicMatch Jukebox as a front end for LAME. This was before Yahoo bought them out and totally destroyed them. It was unsurpassed, even by anything I have today (dBpoweramp, XRECode, Audacity). Sometimes, the more we go forward, the further behind we fall. :-\
Also true. We are living now the "Apple" era when the customer can be dumb because the engineers have thought everything for them.

Quote from: Gurn Blanston on May 29, 2015, 06:17:15 PMPS - Forgot to add; your complaint of some things not handling VBR is no longer valid. Everything handles VBR seamlessly nowadays. It was certainly true at one time, but not any more. Indulge yourself!  :)
Yes, but I learned to rely on CBR in the past and these things are hard to unlearn. At home I hardly ever listen to mp3-files. It's when I sit in a bus to work and back when I listen to mp3-files. In such a listening environment I don't feel needing better than 192 kbps CBR, but of course I could try VBR of similar average bitrate.

CBR has the advantage of constant upper frequency limit. This perhaps gives the sound a feel of stability?
Spatial distortion is a serious problem deteriorating headphone listening.
Crossfeeders reduce spatial distortion and make the sound more natural
and less tiresome in headphone listening.

My Sound Cloud page <-- NEW July 2025 "Liminal Feelings"

merlin

Quote from: 71 dB on May 30, 2015, 01:44:22 AM
The crossfeed feature is a nyquist-plugin 'XFEEDER' I wrote for Audacity. I can send you the plugin if you want. You simply add the file among other .ny-plugins (on macs Applications/Audacity/plug-ins/... )and next time you start Audacity, it has been added to the effect list.

Many thanks for your kind offer!  My email is merlin@evening-sun.com

71 dB

Quote from: merlin on May 30, 2015, 08:11:32 AM
Many thanks for your kind offer!  My email is merlin@evening-sun.com

You are welcome Merlin! Check your email.  ;)
Spatial distortion is a serious problem deteriorating headphone listening.
Crossfeeders reduce spatial distortion and make the sound more natural
and less tiresome in headphone listening.

My Sound Cloud page <-- NEW July 2025 "Liminal Feelings"