Yes, please do what you can, I'm hungry for information.

And sorry folks, post back to the list.  Whoops, pesky Thunderbird.

Paul Sanders.

On 04/04/2018 17:31, Gordon Rankin wrote:


You would have to bypass CoreAudio to get more than 24 bits. As I think someone else said CoreAudio is always going to present in Float32. That would require integer mode.

I basically design hardware and don't do much system programming anymore. I have built a couple 32bit ADC converters in the past and thinking of some ProAudio stuff in the future.

If you are interested in Integer Mode, I could probably hook you up with Damien @ Audirvana. Damien's product is one of the best sounding audio programs and he bypasses CoreAudio in Integer mode for his product.

If you want hooked up with Apple 3rd party just tell me and I can do that as well.


On 4/4/18 12:22 PM, Paul Sanders (nightowls) wrote:

Thanks, I'll do that when this thread appears in the archive so that I can refer to it.  And Microsoft's audio API is horrible, but never mind.

For what I am trying to do, 24 bits of precision is just fine.  If someone is trying to sneak some information through in those other 8 bits then obviously it wouldn't be.  And if anyone has any information on the APIs one might use to bypass Core Audio, I'd be interested.  A casual Google search didn't turn up anything and I might be prepared to go the extra mile here.

Paul Sanders.

On 04/04/2018 15:57, Gordon Rankin wrote:


Actually Apple is really good about this (compared to Microsoft which just lost all their audio device driver programmers---again). You are considered a 3rd party developer so they have a group of engineers that work with you. So do this and I will flag it to Michael with your contact info and they will arrange to put someone in charge of it.


FYI you don't want the data in floating point it will be 32bit floating point and therefore you have lost all the info you may need. 24 into Float and back again to integer works fine... but as you can see 32 would nowt. You could of course do integer mode <> which is really what you want to do for something like this.

I wrote an audio device driver like 4 years ago and everything in it supported 32bit samples so I would imagine your best bet is bypass CoreAudio and go integer mode and you can get all your info.


On 4/4/18 10:48 AM, Paul Sanders (nightowls) wrote:

I don't mind getting the data itself in floating point...  And yes, I think you're right, file a bug report and sit back and wait.  And wait.  And wait.

On 04/04/2018 15:21, Gordon Rankin wrote:

Paul Sanders,

I have been toying around with this as well. I probably know why your doing this and yes if the interface supports 32bits then CoreAudio should support that as well. But do note that you don't want CA to convert to floating point as that would not be bit true.

Probably best to file a bug report and see what comes of it.

Paul Davis,

Really? Look we all know no ADC can even approach 24 bits yet. But who cares, we can now buy 32 bit ADC and no matter how they work or even if they are not as good as other vendors 24 bit ADC it doesn't matter. Some data no matter what it is maybe there that develops better low level information that someone can hear. So let's move forward and not get in a pickle about bits.


J. Gordon Rankin
Owner and Chief Scientist
====== Wavelength Audio, ltd ======
High-End Audio since 1981
SET Tube Amplifiers, DACS & Preamps
Computer Audio, USB DACS
Tube Guitar Amplifiers
NAMM Member since 1998
6902 Murray Avenue
Cincinnati, Ohio 45227 USA
(513) 271-4100 phone/voicemail

Do not post admin requests to the list. They will be ignored.
Coreaudio-api mailing list      (
Help/Unsubscribe/Update your Subscription:

This email sent to

Reply via email to