m4p to mp3, loss of quality? Options

[airtas]

I have an itunes file that is 128 kbps m4p

how much quality will be lost when converting to mp3?

[j7n]

Depends on target bitrate, use generous ~192 kBit/s. A little quality loss will always be there.

M4P? Expect to be struck for Rule 9. You are not allowed to discuss or post instructions how to remove copying restrictions.

This post has been edited by j7n: Aug 6 2008, 18:29

[grommet]

Burning to audio CD (or Virtual CD) is allowed by the iTunes Store for all purchased content. Re-ripping/encoding that is completely legitimate and not a TOS #9 violation.

[sizetwo]

Regardless of the two previous posts, I think you should read up on whats known as transcoding, which is what you will be doing.

The general idea is, as you pointed out, that its not a great idea. But do some testing your self (ABX'ing) and it should answer your question.

[Dynamic]

The general idea is, as you pointed out, that its not a great idea. But do some testing your self (ABX'ing) and it should answer your question.

I've had decent-sounding results with LAME -V2 --vbr-new encodings from 17 iTunes 128 kbps and three iTunesPlus 256 kbps AAC-LC files, which I applied Replay Gain to in foobar2000 before encoding, IIRC. I used -V2 (a transparent setting) as a compromise to hopefully minimise audibility of transcoding artifacts without excessive bitrates and named my files with a !transcoded!.mp3 filename and a comment tag detailing my settings. The bitrate came out fairly low for -V2, at around 141 kbps for twenty tracks. A couple of them were from 256 kbps sources, one was classical (low bitrate expected - came to 164 kbps), one was a mono Otis Redding track from a moderately rare source (106 kbps despite some analogue hiss). Most weren't very modern heavily maximised tracks. One of the more maximised came out at 193 kbps, the highest of the bunch, from a 256 kbps source AAC. Another 256 kbps AAC modern track came out at 155 kbps MP3. Another modern 128 kbps AAC came out to 174 kbps MP3.

Example filename
11 - Wish You Were Here - Pink Floyd !transcoded! fade.mp3

(185 kbps encode from iTunes-plus .m4a, with fades applied in mp3DirectCut as it was a live take with applause that I wish to random-play)

I also found that LAME -V5 was practically indistinguishable too, and more like 123 kbps.

I think the low bitrates were partly down to applying Replay Gain before encoding and partly down to having mostly older, more dynamic, less loud recordings (and one mono recording!) in the first place.

If I could play lossless however, (e.g. FLAC or Wavpack), I'd probably use lossyWAV to transcode to lossyFLAC or lossyWV for added security against the most severe transcoding artifacts.

[Woodinville]

Best advice: If you care about quality, DON'T

http://www.aes.org/sections/pnw/ppt/other/...tual_coding.ppt

Please see slides 96 to 99.

[sizetwo]

The general idea is, as you pointed out, that its not a great idea. But do some testing your self (ABX'ing) and it should answer your question.

I've had decent-sounding results with LAME -V2 --vbr-new encodings from 17 iTunes 128 kbps and three iTunesPlus 256 kbps AAC-LC files, which I applied Replay Gain to in foobar2000 before encoding, IIRC. I used -V2 (a transparent setting) as a compromise to hopefully minimise audibility of transcoding artifacts without excessive bitrates and named my files with a !transcoded!.mp3 filename and a comment tag detailing my settings. The bitrate came out fairly low for -V2, at around 141 kbps for twenty tracks. A couple of them were from 256 kbps sources, one was classical (low bitrate expected - came to 164 kbps), one was a mono Otis Redding track from a moderately rare source (106 kbps despite some analogue hiss). Most weren't very modern heavily maximised tracks. One of the more maximised came out at 193 kbps, the highest of the bunch, from a 256 kbps source AAC. Another 256 kbps AAC modern track came out at 155 kbps MP3. Another modern 128 kbps AAC came out to 174 kbps MP3.

Example filename
11 - Wish You Were Here - Pink Floyd !transcoded! fade.mp3

(185 kbps encode from iTunes-plus .m4a, with fades applied in mp3DirectCut as it was a live take with applause that I wish to random-play)

I also found that LAME -V5 was practically indistinguishable too, and more like 123 kbps.

I think the low bitrates were partly down to applying Replay Gain before encoding and partly down to having mostly older, more dynamic, less loud recordings (and one mono recording!) in the first place.

If I could play lossless however, (e.g. FLAC or Wavpack), I'd probably use lossyWAV to transcode to lossyFLAC or lossyWV for added security against the most severe transcoding artifacts.

As long as you are happy with it and know the consequences of what takes place when you transcode, its all good.

[airtas]

not that its good to transcode but.........

will transcoding a 128 kbps file to 192 be better than doing it to 160?


does going to a higher bit rate minimize data loss?


is the algorithm basically thinking the mp3 file is the original file thus gets rid of more detail?

[kornchild2002]

will transcoding a 128 kbps file to 192 be better than doing it to 160?

Maybe, maybe not. As previously stated, the best thing that you can do is conduct a blind ABX test for yourself to determine which transcoding settings you should use.

does going to a higher bit rate minimize data loss?

Theoretically, yes. It also depends on the encoder that you use (Lame, iTunes mp3, FhG mp3, etc.) and the quality of your source files. I have run into many transcoding situations myself where going from -V 3 to -V 2 didn't yield any audible increases in quality.

is the algorithm basically thinking the mp3 file is the original file thus gets rid of more detail?

That is basically what is happening. The encoders work with lossy material in the same way as with lossless, they are both treated the same.

[krabapple]

Btw, itunes store downloads (.m4a AAC files) are no longer copy (DRM) protected, so Rule 9 should no longer apply.

This post has been edited by krabapple: Jun 6 2009, 05:37

[airtas]

will transcoding a 128 kbps file to 192 be better than doing it to 160?

Maybe, maybe not. As previously stated, the best thing that you can do is conduct a blind ABX test for yourself to determine which transcoding settings you should use.

does going to a higher bit rate minimize data loss?

Theoretically, yes. It also depends on the encoder that you use (Lame, iTunes mp3, FhG mp3, etc.) and the quality of your source files. I have run into many transcoding situations myself where going from -V 3 to -V 2 didn't yield any audible increases in quality.

is the algorithm basically thinking the mp3 file is the original file thus gets rid of more detail?

That is basically what is happening. The encoders work with lossy material in the same way as with lossless, they are both treated the same.

so if you transcode the higher the bit rate the better?

[kornchild2002]

so if you transcode the higher the bit rate the better?

I thought I answered that question, I guess not. You really should conduct your own set of blind ABX tests to determine what transcoding settings that you should use. You can go back and re-read my post if you are unclear as to why you should do this.

[greynol]

Topic closed.