Transcoding MP3 Files - Notes and Hints

What is Transcoding?

Transcoding is the process of taking an audio file encoded with a lossy audio encoder (such as an MP3, WMA, or Ogg Vorbis file) and re-encoding, typically for the purpose of lowering the bitrate. For example, you may want to fit more files in your portable MP3 player's memory. Transcoding generally reduces the audio quality. However, if you are going to listen to the files in a non-audiophile environment (e.g. a car or plane), the quality loss may not be noticable. When transcoding, the MP3 must be decompressed (if only temporarily) to a decoded format (e.g. raw PCM) before re-encoding. Because of quality concerns, transcoding is not recommended, but sometimes it is unavoidable. If at all possible, re-rip the audio from the original source and re-encode from that instead of transcoding. But, sometimes the original source is not available, or the original recording is actually the MP3 itself. In such cases you may be forced to transcode.

Is transcoding really necessary?

If you are transcoding simply to extract a small portion of a large file, you may not need to transcode at all. Instead, use a utility like MP3 Direct Cut which can work on the MP3 audio directly, without transcoding.

Is there any point to Transcoding upward (i.e. increasing the bitrate)?

Transcoding upwards offers no quality benefit whatsoever, and results in a bigger file, so most of the time it is completely pointless. There is no point to transcoding upward (e.g. from 128Kbps to 256Kbps) except in the extremely unusual situation of a hardware player not supporting certain bitrates. For example, early MP3 players sometimes didn't support very low bitrates (like 32Kbps). To play MP3 files on those players, you may be forced to transcode upward. Similarly, there is no point to transcoding to a higher sampling rate. There is no point in transcoding a 32KHz file to 44.1KHz file. Just keep it at 32KHz when transcoding.

What makes transcoding sound bad?

Very often transcoding sounds fine. There is a slight quality loss, but for most non-audiophile purposes (e.g. casual listening) it will work out fine. In my experience, a well-encoded 192Kbps or higher bitrate MP3 at 44.1KHz frequency can be brought down to 128Kbps for listening in a car or other noisy environment, or using poor headphones in a portable MP3 player. In general, however, transcoding sounds bad when you are changing a file's sampling rate. This is done more often with low-bitrate spoken word audio. In general, for best results when transcoding, you should not change the file's sampling rate. Or, if you must change the sampling rate, try to keep it an even multiple (e.g. go from 32KHz to 16KHz). In other words, if you have a 64Kbps MP3 file at 32KHz, it may sound better to transcode it to a 16KHz file than a 22KHz file. Your MP3 encoder should offer these settings (although they may be buried in an Advanced Settings dialog box). Also, transcoding to a different encoder entirely (e.g. MP3 to Ogg Vorbis or WMA to MP3) may sound bad, due to the way the encoder algorithms interact.

Does any of this apply to lossless file formats?

Lossless file formats like FLAC, SHN, and regular PCM WAV or AIFF files do not suffer from any transcoding problems, as long as you maintain the same sampling rate and bits per sample between files. Lossless file formats are designed to maintain 100% of the information of the original audio. No information is lost when encoding, so as long as you stick to lossless file formats, transcoding is not an issue. In fact, one could argue that you are not actually "trans"coding when converting between lossless file formats, because there is no encoding going on to begin with.

How can I make my files smaller and have them still sound OK?

One way that you may not have considered is to transcode from stereo to mono. For spoken word audio, stereo rarely matters anyway. If your original files are stereo, you can transcode to mono and save space.

Summary: Things to Avoid

• Avoid transcoding in general. If possible, re-rip from the original source.
• Never transcode to a higher bitrate (e.g. 128Kbps to 192Kbps). It is pointless and wasteful.
• Never transcode to a higher sampling rate (e.g. 32KHz to 44.1KHz).
• Never transcode from mono to stereo.
• Avoid transcoding between different encoders (e.g. WMA to MP3)
• Avoid changing the sampling rate when transcoding (if you must, it is best use a even multiple like 32KHz to 16KHz)

Useful Audio Utilities

• LAME is the default MP3 encoder. A must-have utility for high-quality MP3 encoding.
• MP3 Direct Cut is a must-have utility for editing MP3s. You can remove parts, change the volume, split files or copy regions to new files without decompressing into PCM. This may eliminate your need to transcode.
• ThisFolder utilities may also be of some use to you when manipulating MP3s.