(This also applies to anyone submitting recordings to IT Conversations.)
Solid-state portable audio recorders are relatively new in the consumer market, and many people are now using them to record podcasts and conference presentations. While these devices are undoubtedly convenient and can store a lot of audio in MP3 format, they’ve introduced a new problem.
Like JPEG, MP3 is a lossy compression scheme, and the effect of repeated encode/decode operations creates some nasty sounding artifacts. Furthermore MP3 is designed as a final encoding technique: The only decode operation should be playback and rendering as sound. You should never encode audio as MP3 unless and until it’s in its final form.
A problem arises when you want to edit an MP3 recording, even just to trim off a few seconds at the head or tail of the file. If you use any computer-based software like Audacity or SoundForge to edit, the MP3 will first be decoded to an internal uncompressed format whe you open the file, but this will not recover the audio lost in that first encoding operation. When you’re done editing, you’ll have to re-encode, and the sound quality will be significantly worse than what your started with.
If you have one of these portable recorders and want to be able to edit your files, or if you plan to submit your audio to IT Conversations, here are some guidelines:
1. If possible record and save in uncompressed in WAV format. I use 24 bits and 48,000 samples per second in the studio, but you can use 16/48,000, 16/44,100 or even 16/22,050 and get pretty good results if you keep your levels up.
2. If you must record or send a compressed file, encode using a much higher bit rate than the final audio. For example, IT Conversations MP3s are encoded at 64kbps. If you send me an MP3 that you’ve recorded and/or encoded at 128kbps, the improvement in quality will be substantial. In particular, the artifacts that you get when you encode at 64kbps then decode/re-encode at the same rate will be far less annoying if I start with your 128kbps version. If you encode at 192kbps or higher, the artifacts will essentially be gone.
If you can only send me a 64kbps MP3, the final 64kbps will sound about the same as a 32kbps original, and there’s a good chance I won’t be able to use it.
3 thoughts on “Those MP3 Recorders”
There is this small program that should allow limited editing on mp3 files without reencoding:
I have used it a few times, but never really tested the difference in quality (in theory there should be none). It can also do more things, like changing the gain on the mp3.
Hope this helps.
Thanks, Radu. Yes, there are also applications called Joiners and Splitters that will perform operations on MP3 files without the decode/re-encode problems. As you suggest, there are no changes in quality. One restriction is that edits resolve to the nearest MP3-frame boundaries — at least I think so; I could be wrong on this — which means the granularity of edits is limited. Does anyone know for sure?
I use a quick little MP3 editor called MP3DirectCut every day. It edits the MP3 file directly, without any transcoding. I recommend it highly — I’ve never noted the slightest degradation in audio quality from using it, though it has had some difficulty re-syncing on write some 24 kbps MP3 files generated by Inet Stream Archiver for Mac OS X. Those I re-encode through Audacity into higher-bitrate files before editing.
Download MP3DirectCut at http://www.mpesch3.de/
Also be on the lookout for MP3Gain, a handy utility for adjusting the replay gain of an MP3 file without recoding it. It’s great for setting the gain level on a batch of files you want to play as part of a single playlist.
Download MP3Gain at http://mp3gain.sourceforge.net/