Ogg Vorbis is an open method of encoding, compressing, and streaming digital audio. Ogg Vorbis .ogg files are similiar to mp3 in function and like mp3 can also be used for streaming audio feeds.
The people behind Ogg Vorbis make a big deal out of the fact that it's completely free, open, and unpatented. Free is good but what I think is cool about Ogg Vorbis is that usually one-eight the bit-rate provides quality superior to mp3. An ogg file at 45kb/s frequently sounds better than an mp3 at 320kb/s. This is particularly true if the original source is clean and has delicate high frequency content. I've encoded various materials at 45kbs with ogg and 320k mp3, and about 95% of the time the Ogg encoded material sounds better, crisper cleaner more detailed highs, solid lows, acoustic ambience, and general clarity.
With mp3 encoding quality at 320kb/s is generally near CD quality but even at 320k high frequency details get lost. Symbols sound like paper scratches instead of metallic crashes. At lower bit rates the first thing I notice about mp3's is that high frequency details tend to get mushy and bass tends to get muddy as well. At 160k, mp3 sounds roughly equivalent to FM broadcast. At around 128k, the highs take on a swishy quality and complex sounds become blurred. Individual instruments in a band tend to lose their distinction at 128k and below. At even lower rates the swishyness becomes very annoying and instruments begin to sound like they're off-key and broken up like a cheap cassette recorder with bad wow and flutter and tape with drop-outs.
Ogg generally sounds very good even at 45kb/s, however, it seems to exaggerate mp3 artifacts, so re-encoding mp3's as Ogg files can be ugly. It's much better to start with clean source. When I first started to play with Ogg Vorbis I thought it did icky things to music when I started re-encoding 320k mp3's. However, I later came to realize that if the source is clean, Ogg at even 45k sounds better than mp3 at 320k, but for some reason Ogg tends to make mp3 artifacts worse, so re-encoding should be avoided, start with clean uncompressed source.
While 56k mp3 encoding is often used for Internet audio broadcasting, the quality is horrid. Internet audio broadcasts using mp3 at 128k are just passable and rarely are higher bit rates used because of the cost of bandwidth. Generally 56k net broadcasts use a sample rate of 22khz thus limiting high end response to roughly 10khz. With mp3 it not possible to provide a high quality streaming audio program over a dialup connection.
Ogg Vorbis encoding at 45kb/s from a clean source sounds better than mp3 encoding at 320kb/s, near CD quality, significantly better than FM broadcast quality. This means a very acceptable quality streaming Internet audio broadcast with 44khz sampling is possible at dialup speeds using Ogg Vorbis encoding. It also means the song you encoded in Ogg Vorbis format will transmit over a 56k dialup connection faster than the same song recorded in mp3 format would transmit over a 256kb/s DSL broadband connection making it practical to trade music over a dialup connection.
It is tempting to re-encode your collection of MP3's as Ogg Vorbis to save space but experience has taught me this is not a good idea. For some reason expanding a mp3 to a wave file, then re-encoding in Ogg Vorbis tends to exaggerate the mp3 artifacts resulting in very hurt music. My advice would be to obtain uncompressed source that has never been compressed, a direct rip from a CD, an analog recording encoded in wave format, and encode directly to Ogg Vorbis. You can often find wave files with peer-to-peer programs like Shareaza but beware, some of these are expanded mp3 files, often of very poor quality.
I use OggdropXPd to encode wave files into Ogg Vorbis format. Don't let the comment about it being for experienced users scare you. It's a drag-n-drop application that is extremely simple to operate.
If you share music, be considerate of the file sharing community. Before you share any of your Ogg Vorbis files, be sure to LISTEN to them first! Don't share poor quality files! Make sure the song is ALL there. Be sure there are no skips from a poor CD rip. Be sure there is no swishiness or muddyness that would result from encoding a wav file that was expanded from an mp3. In short, make sure the recording is complete, intact, and pristine before you make it available. If it is of inferior quality, at the very least make the filename indicate that so that someone doesn't waste their time downloading something they're going to delete the first time they hear it.
When you make your own digital recordings of music there is something to be very aware of. Music has an average level and it has very short transient peaks, the peaks can be more than 20 decibels above the average level. Be sure to set the recording level low enough to allow the highest peaks to be recorded without clipping, otherwise your recording will sound flat and distorted.
In the old days of analog recording, most recording devices did not clip hard. Magnetic media in particular has a linear region which can reproduce the signal faithfully. After that analog region is exceeded there is a non-linear region where the signal is distorted, but it isn't hard clipped. This made analog recording much more forgiving of recordings made at too high of a level. Because of the soft nature of clipping on these analog devices, the distortion that is produced by the occasional clipped peaks tended to be mainly even order harmonic distortion.
Digital recording is less than forgiving. A digital signal can't be less than all zero's or more than all one's, so a digital recording is said to clip hard. Hard clipping produces mainly odd order harmonic distortion which humans tend to find aesthetically more objectionable than even order harmonic distortion prodcuced by soft clipping.
The moral of the story is that when making digital records of your music, allow a great deal of overhead, at least 20db, possibily more.