|
Originally Posted by patrickvo
|
|
Please ignore my ignorance, but what do VBR and joint stereo do? Thanks.
|
They are both techniques designed to compress audio more efficiently.
Variable Bit Rate (VBR) takes advantage of the fact that audio can and does change greatly over time in both frequency and dynamics. More simply put, a song can have simple parts (a single piano playing) and complex parts (a whole orchestra playing). Encoding requires a lower bitrate to represent simple sounds with little reduction in quality than more complex sounds, so with VBR encoded audio the bitrate varies over time to accomodate these changes. A 192k VBR file means the average bitrate is 192 kilobits per second but it can fluctuate between 8k and 320k depending on the demands of the incoming sound. In fact, all mp3s are - to some extent - encoded at a Variable Bit Rate, it's just that the variability is considerably lower with Constant Bit Rate encodings.
Joint Stereo in MP3 terminology is a name applied to one of two slightly different mathematical techniques (Mid/Side Stereo and Intensity Stereo) which can more efficiently encode the stereo information contained in the original recording. Intensity stereo is only used at bitrates <=96k, which is unsuitable for most music so I will ignore it as this post is becoming too complicated already. Mid/Side stereo does not store Left and Right channels separately but as an average of both channels and the difference between both channels. This is advantageous because the majority of the available bits can be used on the Mid channel which contains most of the audible information.The Mid/Side encoding technique is also used (for slightly different reasons) in both FM Radio transmission and Dolby Surround.
@simon-ph: As some posters have pointed out different rates are suitable for different sources so the decision can be difficult. However, it is accepted by many that the best balance between filesize and quality can be obtained by using the LAME 3.90.3 encoder with the alt-preset-standard setting.
A common error with MP3 encoding is assuming that quality increases in a linear fashion, IT DOESN'T! For example the audible difference between a 64k MP3 and a 192k MP3 is huge and the bitrate differs by 128k, whereas the audible difference between a 192k MP3 and a 320k MP3 is comparitively tiny with the same 128k difference!
WMA, Ogg, AAC etc. should provide better quality at lower bitrates compared to MP3 as they address some of its shortcomings and apply additional techniques that MP3 does not support. Bear in mind though that all encoders are sensitive to their source material, some songs will sound better than others regardless of what format you use (excluding lossless techniques), there is no such thing as a perfect lossy audio encoder.