Does the filtered area of an audio file contain bitrate? Assume I encode an audio file at 320kb/s at sampling rate of 44.1kHz (22.05kHz per channel) and by doing this gives me a certain file size. Now using the same bitrate and source, I created another file by filtering to 40kHz. I then look at the spectral graph of this file and see that everything over 20kHz per channel is cut off and entirely black but the file size is about the same as the unfiltered file. So my question is: 1. Is this above 20kHz black or cuttoff area contributing to the file size or is the entire 320kb/s bitrate concentrated at the none cuttoff area (0-20kHz)? Thanks!
Filesize is entirely determined by bitrate and time (bitrate X time = filesize). Sampling frequency and compression have nothing to do with filesize. PS, Have you visited the Afterdawn Glossary? An easy way to find it is to click on an underlined term like: 'bitrate' up there^^.
I know what bitrate is. You misunderstand me. Sampling rate is not the same as filtering (I was talking about filtering). For example, if I resample a 44.1kHz wav file to a 40kHz wav file then the file size will get reduced by about 10% (the same percentage that the sampling rate was reduced) which show a bitrate of 1280kb/s instead of 1411kb/s. But if I filter to 40kHz instead of changing the sampling rate then the file size would be exactly the same. So I ask myself, how can this be? I looked at the spectral graph of this filtered file and seen that the graph per channel is cut off at 20kHz but the container level was up to 22kHz (because it had the same 44.1kHz sampling rate). That means the area from 20kHz to 22kHz was black. Here's a picture of what I'm talking about: http://www.telefonica.net/web2/prg/comparison/results_at_256_kbps.htm Look at the lame 3.97 and AAC itunes graphs and you will see a black area above the cuttoff area (this is filtering). Now those graphs look just like the wav graph I filtered. And again, this filtered wav file that I created had the same exact size as the unfiltered wav file. But when I resampled to 40kHz instead (of filtering) the entire graph was filled up (no black area) but the highest mark shown was 20kHz (for example the highest mark on those graphs in the link I gave is 22kHz). So my question is does this 10% black area (when I filtered) contain 10% of the bits (I said bitrate at first for a mistake) that contribute to 10% of the size of the file? If not then the unblack parts must have all the bits (it is concentrated there) of the file. I'm just trying to wonder why the file size stays the same when I filter.
The filesize remains the same because it has 320 Kbps to store. Any decent encoder won't waste bits on areas that contain no information, so that 320 Kbps will represent only the useful information. If you look at the WAV, which is not compressed, then that space is wasted. But when you compress 1411Kbps down to 320 Kbps the encoder shouldn't waste bits on spectrum that isn't there.
So you're saying that the black area has no bits (or info). And that all of the bitrate goes to the seen area (non-black area). Is this correct?
