well, i heard that if i want to have a better quality i better need to make my songs using 24 bits (or even 32 bits, i'm not sure) and then resample/dither the final mix to 16 bits and the quality will be better than if i worked with 16 bits from the beginning.
Is that right? Why is that? Do i need to work with 48 KHz or should i stick to 41100 Hz because anyways I can record only 16 bit 41100 Hz sound on CD?
If all that is true, how do i resample and/or dither the final mixes to keep the best sound quality?
I hope that the answers to these questions will help not only me, so keep em comming
Try to imagine a ladder with 32 steps, a 32 bit recording. A audio compact disc has a resolution of 16bit only, so 50 percent of your steps in the ladder get lost and you get holes in your ladder. Dithering smears the holes and bring you a good result.
Some say it is better to work in 16bit only some say work in the best bit resoluation you can get.
Anyway dithering is a must when you have a higer bit resolution and you want a final file in a lower resulution.
I always work on 32bit and dither down to 16bit. It is better to work in a high bit resolution because you have a bigger headroom.
The headroom is that room that provides space for the dynamic of your production. In 32bit it is harder to reach the limit.
<font size=-1>[ This Message was edited by: ChrisWerner on 2003-05-09 12:38 ]</font>
higher bit depth=more headroom for processing.(reverbs etc.)the dithering process really does little violence to the original source unlike downsampling,which involves asking the computer to guess what the waveform looked like at moments which were never sampled.(changing bit depth involves looking at moments that WERE sampled)
32 bit actually means 2^32 power which gives you 4,294,967,296 steps. 16 bit will give you 65,536 steps. 24 bit= 16,777,216 steps. midi is only a 7 bit number which will give 0-127. theoretically you could 65,536 signals together and they would never clip a 32 bit number limit. but who needs to do that.
the alesis adat records at 24 bit but in the manual they say to record at -15 to -12 which will use 21-22 of the bits. for every -6db you don't use 1 bit. with a 16 bit signal losing 1-2 bits is a higher percentage of information lost than with a 32 bit system.
for mixing even 24 bit is overkill but eq's, filters and effects(especially reverb) can use all of the bit depth and sample rate they can get. a 10 killohertz sine wave on my ocillioscope looks like a triangle. a 20khz wave looks almost like a square wave.
dithering adds fake noise to the signal to replace the information that's lost in the conversion.
does any one know when the new dvd audio format will be out.
