russdog
500+ Head-Fier
- Joined
- Jul 20, 2006
- Posts
- 695
- Likes
- 16
Quote:
It's lossless but it is not identical in file size. This is because the file is compressed for the purpose of reducing its size. It is "lossless compression" because no data is removed. A CD contains nothing but 1's and 0's. So, to get lossless compression, you need some way to save all the 1's and 0's without using as much space.
Think of it like this. Imagine that the CD contains a little stretch of data that includes 12 zero's in a row, or "000000000000". If you think of it in terms of characters, that's "12 character's worth of space" that is being used. However, we might convey the same information by saying "12-0" (meaning "12 zero's"), and the way I've encoded that here takes up only 4 character's worth of space. So, I just compressed that little bit of data to 33.3% of it's original size.
To uncompress it back to its original form, all I need is something to read the "12-0' code and turn it back into "000000000000", and I recover the original without losing anything. When I turned the "000000000000" into "12-0" I encoded it. When I turn the "12-0" back into "000000000000", I decode it. FLAC (and any other implementation of a compression scheme) is called a "codec", which is short for "coder-decoder". This specific example is of a very simple compression scheme, but the basic idea is the same. Make sense?
Everybody who understands what happens agrees that "lossless compression" and "uncompressed" are identical in the end. The disagreement focuses on whether stronger compression schemes that do throw out some data can be just as good as the original. The high quality versions of "lossy compression" are more than just compression schemes. They are also based on psychoacoustic models of what people can and cannot perceive. They throw out the data that the models say people cannot perceive. Depending on how much you try to compress, people either can or cannot tell, and people vary on what their threshhold is. So, at what point can you start telling that they threw some bits away? That's what the main fuss is about.
Originally Posted by J Tran See something odd? Isn't FLAC "mathematically" equal to a CD? If so why is it smaller than the CD size? And the .WAV is equal to the CD size. |
It's lossless but it is not identical in file size. This is because the file is compressed for the purpose of reducing its size. It is "lossless compression" because no data is removed. A CD contains nothing but 1's and 0's. So, to get lossless compression, you need some way to save all the 1's and 0's without using as much space.
Think of it like this. Imagine that the CD contains a little stretch of data that includes 12 zero's in a row, or "000000000000". If you think of it in terms of characters, that's "12 character's worth of space" that is being used. However, we might convey the same information by saying "12-0" (meaning "12 zero's"), and the way I've encoded that here takes up only 4 character's worth of space. So, I just compressed that little bit of data to 33.3% of it's original size.
To uncompress it back to its original form, all I need is something to read the "12-0' code and turn it back into "000000000000", and I recover the original without losing anything. When I turned the "000000000000" into "12-0" I encoded it. When I turn the "12-0" back into "000000000000", I decode it. FLAC (and any other implementation of a compression scheme) is called a "codec", which is short for "coder-decoder". This specific example is of a very simple compression scheme, but the basic idea is the same. Make sense?
Everybody who understands what happens agrees that "lossless compression" and "uncompressed" are identical in the end. The disagreement focuses on whether stronger compression schemes that do throw out some data can be just as good as the original. The high quality versions of "lossy compression" are more than just compression schemes. They are also based on psychoacoustic models of what people can and cannot perceive. They throw out the data that the models say people cannot perceive. Depending on how much you try to compress, people either can or cannot tell, and people vary on what their threshhold is. So, at what point can you start telling that they threw some bits away? That's what the main fuss is about.