Skip to main content

Notice

Please note that most of the software linked on this forum is likely to be safe to use. If you are unsure, feel free to ask in the relevant topics, or send a private message to an administrator or moderator. To help curb the problems of false positives, or in the event that you do find actual malware, you can contribute through the article linked here.
Topic: ALAC Converter that supports 32-bit fixed (AKA 32 bit integer)? (Read 2464 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

ALAC Converter that supports 32-bit fixed (AKA 32 bit integer)?

ALAC itself supports 32-bit integer, but I'm questioning whether a converter for it exists. I've looked at dbPoweramp, CUETools, FFmpeg, and alactools, but none of them support 32-bit ALAC conversion. Does anyone know of one that supports 32-bit int?

Re: ALAC Converter that supports 32-bit fixed (AKA 32 bit integer)?

Reply #1
EZ CD Audio Converter encodes 32-bit Apple Lossless ALAC
https://www.poikosoft.com


Re: ALAC Converter that supports 32-bit fixed (AKA 32 bit integer)?

Reply #3
Hello,

You can also use refalac from qaac package.

    AiZ

Re: ALAC Converter that supports 32-bit fixed (AKA 32 bit integer)?

Reply #4
OP: from your question I take it that you mean encoding. As AiZ points out, refalac implements it as good as anything.
refalac also supports encoding by drag+drop: drag the .wav file onto refalac64.exe and drop it on the icon, and it will convert.

But there are better codecs around, and there is no reason to select ALAC unless you are on Appleware. Both of the following two reference encoders also support 32-bit and also (on Windows) support drag+drop encoding:
* WavPack has supported 32-bit (integer and float) since long, has better security against corrupted files, and for "fake" 24-bit or 32-bit signals (that are really say 16, but zero-padded to 24 or 32 bits), it compresses better. See the hi-res section of http://audiograaf.nl/downloads.html
Drag+drop can even be configurated by by renaming wavpack.exe
* FLAC supports integer 32-bit from release 1.4.0. But if you have an audio player stuck with a lower FLAC version, then you are in trouble.

Re: ALAC Converter that supports 32-bit fixed (AKA 32 bit integer)?

Reply #5
OP: from your question I take it that you mean encoding. As AiZ points out, refalac implements it as good as anything.
refalac also supports encoding by drag+drop: drag the .wav file onto refalac64.exe and drop it on the icon, and it will convert.

But there are better codecs around, and there is no reason to select ALAC unless you are on Appleware. Both of the following two reference encoders also support 32-bit and also (on Windows) support drag+drop encoding:
* WavPack has supported 32-bit (integer and float) since long, has better security against corrupted files, and for "fake" 24-bit or 32-bit signals (that are really say 16, but zero-padded to 24 or 32 bits), it compresses better. See the hi-res section of http://audiograaf.nl/downloads.html
Drag+drop can even be configurated by by renaming wavpack.exe
* FLAC supports integer 32-bit from release 1.4.0. But if you have an audio player stuck with a lower FLAC version, then you are in trouble.


Thanks for the response; I was just investigating whether a tool for this existed out of curiosity, as even macosforge's alacconverter doesn't implement ALAC's 32 bit capabilities, even though the codec itself supports it. Obviously, since ALAC isn't very popular, it was a bit of a struggle to find a tool, as the more popular (to me, at least) encoding tools like FFmpeg doesn't support 32 bit ALAC encoding.

One thing you mentioned piqued my interest, though. Would a 24-bit signal be considered "fake" even if it originally recorded in 24-bit despite not using the extra headroom (in terms of dynamic range) offered over a 16-bit signal?

Re: ALAC Converter that supports 32-bit fixed (AKA 32 bit integer)?

Reply #6
Obviously, since ALAC isn't very popular, it was a bit of a struggle to find a tool,
refalac is based on the official ... maybe the ALAC wikipedia page needs update.
https://github.com/nu774/qaac/wiki/refalac-usage

as the more popular (to me, at least) encoding tools like FFmpeg doesn't support 32 bit ALAC encoding.
You should also be aware that FFmpeg "doesn't default to correct format". Its default does often coincide with the "correct" format, but it isn't really written for lossless operations. For example, default .wav output means (down)converting to 16 bits - otherwise you have to specify it. FFmpeg will happily clip without warning when converting from float to integer. Also it does strange things like converting 24-bit FLAC to 32-bit WavPack (I don't know how to override that).

Not supporting 32-bit ALAC encoding (it supports decoding) is kinda ... there are several formats where it decodes more than it encodes. So it isn't at all out of character. It is a weird choice given that a 32-bit ALAC encoder is openly available. (ffmpeg's is different than refalac! https://hydrogenaud.io/index.php/topic,123511.0.html )

One thing you mentioned piqued my interest, though. Would a 24-bit signal be considered "fake" even if it originally recorded in 24-bit despite not using the extra headroom (in terms of dynamic range) offered over a 16-bit signal?
"fake" was something I wrote because it conveyed some kind of information. Whether it would be considered "fake" in any pejorative sense depends on the circumstances.
It sometimes happens that someone opens a file in their software and exports it, using the default format. Assuming the software doesn't add dither (some will ...), is that "fake" if it is saved as a 24-bit FLAC even if the bottom 8 bits are all zero? You could argue that "fake" is a too harsh word, it was not intended to deceive anyone and the result is just fine. But if it was done to market it as higher resolution, it is fake indeed. (Maybe not dishonest - audiophile mumbo-jumbo is often a result of wishful thinking. Look at this bloody nonsense: https://www.soundliaison.com/index.php/studio-masters/856-ray-carmen-gomes-inc Recorded in DXD (352.8/24 integer, that is a dumb format), DACed to a Studer analog tape, digitized from there to 768 kHz digital ... )

Now in some scientific traditions it would be bad practice to claim more resolution than you have. Are 2.13*105 and 2.1300*105 the same? In high school physics I was taught that making the latter statement amounts to a higher accuracy claimed, which you shouldn't do if you cannot justify it.

Bottom line: I was too lazy to clarify "fake" because I wasn't in a situation where I thought more clarity was needed.

 

Re: ALAC Converter that supports 32-bit fixed (AKA 32 bit integer)?

Reply #7
Also it does strange things like converting 24-bit FLAC to 32-bit WavPack (I don't know how to override that).
This should do

ffmpeg -i somefile.flac -bits_per_raw_sample 24 somefile.wv
Music: sounds arranged such that they construct feelings.