Hi all, if I was creating say a 16 bit stereo wav file and rolled off certain frequencies (say anything below 45hz and anything above I dunno, 15k?), does this leave more room in the 16 bit file to describe the rest of the song? Also, with volume, if the song has a more linear volume level does the same apply?
Would you get a clearer 16 bit file because the entire 16 bits is being used to describe a smaller frequency and sonic spectrum?
Does creating a wav file or similar work this way?
Cool question! I am not sure about this myself but will post what I think is the correct answer and see if I am correct... [img]images/icons/grin.gif[/img]
Sampling frequency describes the sonic spectrum you can capture and the bit depth describes how many steps the dynamics of the sound is being quantised into... Higher sampling frequencies means greater sonic spectrum and higher bit depth means the volume of the sound is split into smaller quantised steps... At this point, I presume that they are both independent...
If I have a 1bit dynamic range, it will either be 0 or 1 (where 1 is the maximum volume of the sound). So no matter what sampling frequency I use, a sine wave will turn out to be a triangular wave... So now if I increase the bit depth to 16 bit, then depending on the volume of the sound, the tip of the sine wave may be \"clipped\" due to this bit depth, where say for example the voltage difference between the top most point of the sine wave and the second top most point in the actual sound is 0.5V and the quantisation step is 1V, then the top, which is supposed to be at 1V will now be 0.5V, making it flat at the top... Does this make sense???
Sorry, I totally lost my point but I will leave this post here anyway for anybody who finds it funny... [img]images/icons/wink.gif[/img]
Scott, rolling off the frequencies won\'t help at all with wave files. The files work in the time domain, and the EQ is in the frequency domain. At each sample (which is like a snapshot) you get a particular level - but the individual sample has no idea of the frequency.
The best way to get your 16-bit files to work well is to keep the levels high, but not so high that they clip. And, if you can, work with 24-bits and only go down to 16-bits at the last step.
The one place where the EQ can help is if you are compressing to MP3, Real, Quicktime, WMA or some other compressed format. The low-end rolloff isn\'t that crucial, but the high-end is. By getting rid of the highest frequencies, you let the compression work the main bands more effectively. Really high tuned cymbal rides and brushed snares can start to sound like random noise when over-compressed, so dumping the 16k and over stuff can often improve the sound.
However, it is true that stuff in the rumble range eats up \"VU\" (in quotes), so it\'s not a bad idea at all to roll it off at 45 or even higher if there\'s nothing useful going on down there. But if anything you\'re always fighting for more air in your recordings, not less, so forget about lowpassing it at 15k. That\'s not going to make any difference anyway.