When kAudioFormatFlagIsPacked is set, it means that all the bits in the sample are significant. When it is unset, it means that some bits in the sample aren't used. Which bits those are is determined by whether kAudioFormatFlagIsAlignedHigh is set or not.
In your case, you should set kAudioFormatFlagIsPacked. In fact, you might just want to go ahead and use the pre-assembled flag constant, kAudioFormatFlagsNativeFloatPacked.
On Jan 18, 2008, at 12:17 PM, Roland Silver wrote: I want to output (using Audio Queue) a time-series of algorithmically generated sound signals in linear PCM format: left-channel-signal #1, right-channel-signal #1, left-channel-signal #2, right-channel-signal #2, etc. Each signal is represented as a 32-bit float. Question: Should the format flag kLinearPCMFormatFlagIsPacked be set? I've read Core Audio: Audio Stream Basic Description, but I'm still not clear on the answer to my question, because I don't understand what "packed" means, exactly. I think my data is packed, but I'd like to make sure.
--
Jeff Moore
Core Audio
Apple
|