milosz
500+ Head-Fier
- Joined
- Aug 17, 2009
- Posts
- 971
- Likes
- 123
Yeah, when you RECORD a file, it's a good practice to use 24 bit, or even 32 I guess, because then when you edit / mix / filter / sound-design etc you have greater flexibility to combine, filter, add, subtract etc - and, when DSP is working with 24 bit data (or 32) and then your final product is 16 bit Red book, you'll have greater precision expressed within those 16 bits than if you done all your mixing and other DSP work in 16 bits all along. The rounding errors, etc, tend to add up and if those LSB's are bits 23 and 24 - which more or less get truncated when going from your 24 bit working file to 16 bits final output render- that little bit of grunge from rounding etc down in the least significant bits will just be discarded.
I think this is one of the reasons that some modern digital recordings can sound so good, people are paying attention to these things.
So if you are doing some DSP (like EQ, etc) on playback, it's useful to do that DSP at 24 bits rather than at 16, even though your original data is just 16 bits. The rounding errors and so on in DSP can build up from various iterative things going on inside the DSP algorithm, and those rounding errors are harder to hear (impossible to hear?) when they are in the 24th bit rather than in the 16th bit.
Your music- the 16 bit data- won't magically grow into 24 bit data, but the operations done to that 16 bit data will be more precise and likely better sounding if those operations are done at 24 bits.
I think this is one of the reasons that some modern digital recordings can sound so good, people are paying attention to these things.
So if you are doing some DSP (like EQ, etc) on playback, it's useful to do that DSP at 24 bits rather than at 16, even though your original data is just 16 bits. The rounding errors and so on in DSP can build up from various iterative things going on inside the DSP algorithm, and those rounding errors are harder to hear (impossible to hear?) when they are in the 24th bit rather than in the 16th bit.
Your music- the 16 bit data- won't magically grow into 24 bit data, but the operations done to that 16 bit data will be more precise and likely better sounding if those operations are done at 24 bits.