Re: Core Audio integer mode
Re: Core Audio integer mode
- Subject: Re: Core Audio integer mode
- From: Brian Willoughby <email@hidden>
- Date: Wed, 31 May 2017 12:40:42 -0700
Gordon,
I was restricting my comments to those situations where you can afford to not care whether you're dealing with integer or float. In those cases, 24-bit is the maximum you can support losslessly. CoreAudio is proven to be bit perfect from 24-bit integer input to 24-bit integer output, even with 32-bit floating point conversions in the middle.
It's often important to maintain more than 24 bits for intermediate DSP formats, but not for conversion.
As for hardware, I am highly dubious of 32-bit converters. As it stands, we can barely exceed 110 dB of dynamic range. 24-bit audio has enough dynamic range to extend from the quietest sound the human ear can discern, all the way to the sound pressure level near an atomic blast. Admittedly, signal to noise ratio is not the only metric, because human hearing can extract sound details from below the noise floor, but 144 dB conversion doesn't even make sense with 112 dB S/N.
Are you saying there are ADC chips which guarantee 192 dB signal to noise?
Brian Willoughby
On May 31, 2017, at 12:13 PM, Gordon Rankin <email@hidden> wrote:
> Why is integer only to 24 bits? I was actually thinking of making a 32 bit stereo ADC, but not if it's going to chop off the lowest 8 bits.
_______________________________________________
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