Subject: Re: [ecasound] need some buffering advice
From: Julian Dobson (juliand_AT_braverock.com)
Date: Thu Sep 25 2003 - 15:45:12 EEST
Is there any reason for the int->float->int conversion? I could see that
you might have situations where you would need to do this, but there
might as in this case, be situations where you don't want to convert, or
you want absolute bit for bit output.
I guess what I'm trying to say is, would it be worth/feasible having an
option that turns off the conversion to floats, and does everything as
Jamie Lutch wrote:
> yes, i see this, when compared with a source file, there was a bit of
> rounding, if you invert the source and recorded wav, it cancels down
> to -141db, so it looks like the least significant bit is being rounded
> fortunately, thats well below the resolution of any practical device..
>>[mailto:ecasound-list-owner_AT_wakkanet.fi]On Behalf Of Kai Vehmanen
>>Sent: Wednesday, September 24, 2003 6:07 PM
>>Subject: RE: [ecasound] need some buffering advice
>>On Wed, 17 Sep 2003, Jamie Lutch wrote:
>>>>No dithering, but the data is converted from 32bit integers to
>>>>32bit float and then back to 24bit integers.
>>>the question being, is the 24bit file recorded bii-for-bit
>>identical to the
>>>24-bit input source? seems so, but I figure I'd ask.
>>Yes, it should be. I say 'should' as the all the audio data does go
>>through two conversions (int->float, float->int) and bugs and/or rounding
>>errors are always possible. Ecasound does have a small test suite for the
>>conversion routines, so they should be ok, but I don't want to give any
>>guarantees. If bit-level accuracy is important, you should verify the
>> Audio software for Linux!
>>To unsubscribe send message 'unsubscribe' in the body of the
>>message to <ecasound-list-request_AT_wakkanet.fi>.
> To unsubscribe send message 'unsubscribe' in the body of the
> message to <ecasound-list-request_AT_wakkanet.fi>.
This archive was generated by hypermail 2b28 : Thu Sep 25 2003 - 15:44:33 EEST