| Why use mp3 sound files? | • |
|
How do I play mp3's? |
• |
|
How do I make new mp3's? |
• |
|
Where do I find mp3's on the Internet? |
• |
|
What software is out there, for Mac, Linux, Amiga, and Windows? |
• |
| Give me more technical information on mp3 compression! | •• |
| What's the difference between AIFF, WAV, SND, and SD2 audio files? | • |
| Is mp3 legal? | • |
(click on the blue button for information...)
page created by miq productions
 |
|
MIQ
ONLINE MP3 TUTORIAL
|
1. kbps compression
First maxim: the higher the bitrate, the less change to the sound, the larger the file size.
What is the appropriate bitrate? Hard to say. mp3.com, by virtue of their enormous web presence, has created a case for the 128 kbps compression ratio (which is about 12:1 compression, or 1 minute of audio for 1 Mb of file). Although some radio sites do broadcast at lower bitrates, there are no standards, for example, for spoken word (talk radio or radio drama), for sports casts, or even for audiophile music broadcasts.
Second maxim: some bitrates produce better sound in the mp2 format than the mp3 format.
What's that, you say, mp2? Yes, mp2 is known to sound substantially better at the lower bitrates than the equivalent compression in mp3. Why is this? This is not certain -- to understand better the process, we should spell out that mp2 and mp3 refer respectively to "MPEG-2 compression layer 2 and MPEG-2 compression layer 3", meaning that the latter uses an extra compression layer. Somewhere in this additional code lower bitrates become garbled. However, for 128 kbps and up, mp3 is the superior format.
As with all new, unfamiliar things, you will need to experiment with various compressions, both in mp2 and mp3, and an excellent sound system with which to test the results. Note that not all encoders are the same! You may wish to test 3-4 encoders at each bitrate!
2. stereo formats
There are 2 primary stereo formats; joint stereo and split stereo. Split stereo encodes each track of the stereo pair separately and merges the compressed audio data; joint stereo creates a stereo encoding algorithm, and as such can make changes to the stereo imaging of the sound- file. Though many people report liking the joint stereo effect, if you are very particular as to the stereo image you should use split encoding. For a while mp3.com was requiring joint encoding.
3. psy models
Psy models is short for "psychological encoding models," and there are normally 3 options: none, simple, and complex. The psy models are a filter on the encoding process which affect the perceived distortion caused by compression. Basically, the premise behind the psy models is that digital audio has an excess number of samples already, and that with proper modeling distortion can be added without perceivably changing the sound of the uncompressed file. The simple psy model is often fine for 128 kbps files, and the complex psy model can allow more compression before distortion. To many people the psy models sound like EQs, enhancing certain high frequencies and low bass. I've found that the psy models are detrimental to very ambient music (without percussion, that is) and some acoustic world music, but enhance electronic genres like trance and techno, making the mp3 sound better than the CD at times! As with compression ratios, experiment with the psy models. Even more than the compression schemes, the psy models differ widely from encoder to encoder, so you should check several applications to find the best compression / psy model pair for your music.