0

Reading this answer here and questions like:

What lossy audio formats are more (or less) "efficient"?

Are there any perceptible differences between the sound quality of 192 versus 320 kbps MP3 files?

FFMPEG: Convert m4a files to mp3 without significant loss of information / quality. Automated bitrate

...has made me really wonder about whether I'm handling my music in the most efficient way; whether I'm converting my music collection to too high a bitrate.

At the moment, my strategy for my music collection is pretty crude; I download the songs from my YouTube playlists (using youtube-dl) in the best quality M4A versions available, and then mass-transcode those to a bitrate of 192kbps using Format Factory. If any of the downloaded files are already 192kbps or below, then I leave them as is.

I'm wondering whether forcing all my M4As to 192kpbs or below in this way is the most efficient way to go. Is it possible that 192kbps is wasteful/inefficient - that M4A files can tolerate much lower bitrates while retaining the same quality - or is this entirely dependent on the bitrate of the file and/or other factors? I'm wondering how everyone else approaches this situation. Also, while I'm at it, if the encoding/transcoding experts can spot any other point of my working process where I might be sabotaging the quality of my files, I'd really appreciate that being pointed out to me.

Community
  • 1
Hashim Aziz
  • 13,835

1 Answers1

1

I'm going to give a very high level answer

The concept you are asking about is called transparency. http://wiki.hydrogenaud.io/index.php?title=Transparency

For lossy formats, many people find it hard to tell the difference between audio encoded at a low bit rate vs. a high bit rate. Some people's ears are more sensitive than others'; also the speakers and the physical space where the music is being played does affect results.

The best way to verify this is to try ABX testing; you compare lossless format vs. a lossy audio file and see if you can tell the difference. http://wiki.hydrogenaud.io/index.php?title=ABX

Generally transcoding between two lossy formats is discouraged because it results in bad artefacts. Even if your ears can't tell the difference, transcoded files generally have more artefacts than a simply lossy file. It's like making a xerox of a xerox. You amplify distortions. Ideally you should be transcoding from a lossless format like .flac or .wav to a lossy format like m4a, ogg or mp3.

My ears are not very discriminating (also I'm 51!), but I find 192kpbs vbr 0 is more than adequate for mp3, and perhaps lower is sufficient for m4a. Personally I find it difficult to differentiate between lossless files and mp3 128 kpbs. (Don't mock me!)

Another thing to keep in mind is portability and file size. I'm willing to live with a lower bit rate to minimize file size.

If you are interested in permanent storage, you should keep .flac or some other lossless format, then use an audio conversion tool to convert lossless-to-lossy.

One more thing. I've been very impressed by the specs of the new opus format and I have been encoding all my audio files in opus format. http://wiki.hydrogenaud.io/index.php?title=Opus Not all the players support it, but support has been more widespread -- to the point where it is my audio file format of choice.

If you are using a program to scrape mp3 files from youtube, I would try to get it in one format and not do any transcoding after that.

rjnagle
  • 294