Re: long double data type
Re: long double data type
- Subject: Re: long double data type
- From: tahome izwah <email@hidden>
- Date: Sat, 10 Jul 2010 09:32:20 +0200
In most (not all) of these cases however the algorithms can be
reformulated to work with a lower floating point dynamic range as
well. This affects their efficiency, but it is possible.
--th
2010/7/9 Christopher Penrose <email@hidden>:
>
> Sorry to respond to this belatedly, but I had to weigh in.
> On Jun 15, 2010, at 5:38 PM, Dominik Peklo wrote:
>
> Personally, I would take fixed point over floating point any day. In audio,
> there's no need for extended dynamic range afforded by exponent. The range
> of interest is only 0dBFS and down. 32b float wastes 8b on exponent,
> extending dynamic range by 256dB over the 144dB afforded by mantissa.
>
> There are many, arguably common, DSP applications, usually employing
> transform mathematics, where extended floating point dynamic range is
> beneficial for maintenance of precision.
> Christopher
>
_______________________________________________
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