|   Sir Terrence.... | Monstrous Mike
Oct 23, 2002 7:39 AM | | ...the Terrible.
I'd like to discuss the MLP compression of music a bit more.
Using the Winzip analogy again, some files compress better than others. The reason for this is that lossless compression (and I'm not sure about MLP) uses the technique of substituting longer strings of data that repeat for a shorter representative string. This is a form of coding. Using this type of coding, data which doesn't have many repetitive strings cannot be compressed as much as data that has many repetitive strings.
In computer land, a bitmap image can be compressed by winzip to a fairly large ratio since it usually contains many repetitive elements. Jpeg is already a compressed graphics format so winzip will not compress it much further.
So getting back to MLP and the compression of music, two minutes of silence should be able to be greatly compressed, whereas two minutes of symphony music with many instruments would no be so compressible (is that a word?).
While I realize that MLP performs operations on 1s and 0s, and doesn't care what the musical content is, if the musical content has a higher variety of sounds, and perhaps a higher randomness of 1s and 0s, do some music tracks compress more than others?
In other words, like in computer land, does digital music data that has more "randomness" or variety of bit combinations compress as well as music data that has many repetitive elements?
The coding courses I have taken have indicated that the amount (or ratio) of compression is related to the content of the data. I was wondering if this applies to MLP compression for DVD-Audio. |
| Mike, you forgot bowing to Sire TT :) | KiD SmokE
Oct 23, 2002 9:28 AM | | So here is one so TT don't get offended <img src="http://www.regalweb.co.uk/loony/animated/bowingboy1.gif">
Great question and I wait for the answer :) |
| The kid saved you from being beheaded.... | Sir Terrence the Terrible 1
Oct 24, 2002 9:06 PM | | MM,
I almost ordered you beheaded for your lack of following protocol. I am going to have to knight the Kid as my International Minister of Protocol so he may show my subjects how to address royalty.
Winzip shares many characteristic of MLP. MLP does not look at the individual notes of instruments in the data, it looks at the whole stream for compression. So it doesn't matter the density or the complexity of the 0's and 1's. It looks for signal redundancies(just like lossy codecs do)reduces them(just like lossy codecs do) thereby reducing the amount of information or bits(squeezing)allowing for more data on the disc. MLP offers these other features as well as data compression;
At least 4 bits/sample of compression for both average and peak data rates
Easy transformation between fixedrate and variablerate data streams
Careful and economical handling of mixed input sample rates
For decoders, modest number crunching requirements and special stereo playback considerations
So rather than looking at the individual 0's and 1's, MLP looks at the stream of 0's and 1's, checks for redundacies, and reduces the bits accordingly.
Sir Terrence(a bow is in order)the Terrible |
| |
