Post by George W.I stand corrected by Paul, who knows a LOT more about this than me.
I've been using the 24 bit setting myself because it's there and I
have the drive space. I admit I don't really hear a difference but
with givin my inexperience and cheapo equipments that's not
surprising. One distinction here, and correct me if I'm wrong: The use
of 24 bit is mostly for recording, where it probably makes sense, but
in the end most things are mixed to 16 bit. As far as I know that's
the standard for CD's. So.....does the improvement of 24 bit remain
after mixdown to 16 bit?
This gets complicated. :) And I may not understand this all that well, so
anyone out there, please feel free to jump in with corrections.
24 bits gives you more headroom, and less need to slam the A/D converters
to "use all the bits" and get a low noise floor.
But 16 bits can work fine... *if* you're not messing with the signal very
much after you record it.
The more important issue is what format you store your audio files in (16,
24, or 32 bit/float) and how much destructive audio processing you do after
recording.
If you record a multitrack song in 16 bits, hot enough for a low noise
floor, and then don't mess with it... you don't add a lot of EQ, or volume
changes, or compression... then bounce a final stereo 16 bit .wav file out
of your editing software... you may not lose any significant audio quality.
In fact it may even sound better than recording in 24 bits and then doing a
dither and bit reduction down to 16 bits (assuming your A/D converters are
very good).
The problem with 16 bits arises during processing like EQ, volume change,
and effects like compression during mixdown. These are all math operations
which result in something called digital word length expansion. You may get
rounding errors (digital noise) when the audio is converted back down to 16
bits for burning to a CD, or an MP3 conversion.
The major DAW software packages (Cubase/Nuendo, Logic, Sonar, ProTools,
Samplitude, etc.) operate at a high internal bit resolution when you're
mixing, usually something like 48 or 56 bits fixed, or 32 point floating.
This allows for rounding errors after word length expansion, and that's why
you want to write a 24 or 32 bit file after doing any destructive wave
editing. It's not so much of a problem with non-destructive editors. But
it's still a good idea, if possible, to store your intermediate mix files
in 24 bit or 32 bit format. Then convert to 16 bits at the very last stage
of the process, just before converting to an MP3 or burning an audio disk.
Make sure you use a dither process when moving from 32 or 24 bits to 16
bits.
In case that wasn't clear (and this stuff really makes my head hurt
sometimes), here's the short version:
If your recording is fairly basic... like a live 2-track recording and you
don't add EQ or volume changes, or compression, or reverb... then it can
sound very good going straight from a 16 bit recording to a 16 bit MP3 or
audio disk. Otherwise, use 24 bits. Especially with hard drive storage
being so cheap now.
P.S. And also stick to 44.1 kHz sampling rate, unless you have a very good
reason to use something higher.
--
Mike Barrs