There was an article in a recent Electronic Musician (if I'm not
mistaken?) about 32 bit floating point (which is a 24 bit mantissa
and 8 bit exponent) and 24 bit integer formats.
ProTools, fwiw, stores sound as 24 bit integer (or 16 bit, if you use
that mode) and mixes using a 56 bit accumulator. There's an option
of a "dithered mixer" which uses a little extra DSP to dither the mix
before it's bumped back to 24 bit at the output.
Not a DSP whiz, so I leave someone else to interpret.
> > Dan Dugan, in a reply to Klas Strandberg (about digital manipulation, =
and
>> especially w.r.t normalization), wrote
>>
>> > In intermediate steps, you need headroom
>> > for processing.
>>
>> My question: I've always assumed that most digital manipluations were d=
one
>> as floating point (thus giving
>> more headroom than even long integers could require.) Is this true? If =
not,
>> what are the problems with using
>> floats?
>
>FWIW my own PC-based editor, Samplitude, by default performs all
>calculations in a 32-bit space, I don't remember if it's FP or integer.
>But the upshot is you never run out of headroom, signals can get much
>hotter than 0db and never clip. It's only when the mix is 'bounced' down
>to 44/16 for CD delivery (for example) that you need to normalize so that
>no signal is hotter than 0 db...
>
>best,
> aaron
>
>
> http://www.quietamerican.org
>
>
>
>To unsubscribe from this group, send an email to:
>
>
>
>
>Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/
--
-------------------------------------------------------------------
jeremiah lyman moore | san francisco | sound+media |
http://babyjane.com/timeweb/
http://northstation.net/ organic, mechanized, organized sound
________________________________________________________________________
________________________________________________________________________
|