Re: AudioConverter gives non normalized samples?! (Doug Wyatt)
Re: AudioConverter gives non normalized samples?! (Doug Wyatt)
- Subject: Re: AudioConverter gives non normalized samples?! (Doug Wyatt)
- From: Brian Willoughby <email@hidden>
- Date: Wed, 28 Jan 2009 12:43:39 -0800
Good question, Ernest. In CoreAudio code, the data is (almost)
always 32-bit float. Nobody is confused about that, hopefully. The
question is what kind of hardware is being used for the DAC, or what
depth is being used for the file format if saving to disk. CoreAudio
allows code to be written which is not tied to a specific output
device or format, and AudioUnits and AUGraphs allow chaining of multi-
purpose processing units. With such flexibility, it's quite easy for
one piece of code to have no information about the details of the end
of the processing chain.
For the most parts, many algorithms can ignore the range and just
implement the math. But some algorithms explode with unbounded input
ranges. Ultimately, it's up to the user to manage the gain staging,
but the gain units may need to make limited assumptions about the
final bit depth to control clipping.
Brian Willoughby
Sound Consulting
On Jan 28, 2009, at 12:35, Ernest Burghardt wrote:
and I was curious why one would not know the bit depth of the data
being being discussed?
_______________________________________________
Do not post admin requests to the list. They will be ignored.
Coreaudio-api mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
This email sent to email@hidden