NetNews Offline 2/NetNews Offline Volume 2.iso/news/comp/sys/amiga/programmer/2744

Internet Message Format | 1996-08-05 | 2.8 KB

Path: informatik.tu-muenchen.de!fischerj
From: [e-mail] (Juergen "Rally" Fischer)
Newsgroups: comp.sys.amiga.programmer
Subject: Re: playing 14 bit samples. how?
Date: 2 Feb 1996 16:30:36 GMT
Organization: Technische Universitaet Muenchen, Germany
Distribution: world
Message-ID: <4ete7c$[e-mail]>
References: <[e-mail]> <4enpm4$[e-mail]> <[e-mail]>
NNTP-Posting-Host: hphalle5.informatik.tu-muenchen.de
Originator: [e-mail]


In article <[e-mail]>, [e-mail] (Mathew Hendry) writes:
|> Juergen "Rally" Fischer ([e-mail]) wrote:
|> : John Hendrikx ([e-mail]) wrote:
|> :
|> : You calibrate the audio hardware, which does maybe not exactly play
|> : 1/64 of max volume if volume=1 selected.
|>
|> In fact, the hardware reference manual states that volume 1 plays at -36.1dB
|> (rather than -36.0dB) relative to volume 64, so even the nominal value
|> is slightly off. In practice the difference in volume varies with frequency as
|> well, and this too has to be compensated for.
|>
|> : : sample and using the volume to compensate for which top 8-bits your playing.
|> :
|> : : (The other method to use 2 channels playing at different volumes seems easier
|> : : to use though).
|> :
|> : Yes, because you don't need to sync them so exactly.
|>
|> If they aren't synced exactly, you're adding more noise than you would get by
|> playing a single 8 bit sample - as soon as the two "halves" of the sample
|> become desynced, you'll be playing the 8 most significant bits of one sample
|> along with the 6 least significant bits of another => nasty noise.
|>
|> The volume method is simple to use because it only requires you to split each
|> sample in two and right shift the less significant "half" by two bits. After
|> that it's all up to the DMA. The modulation method requires a bit more initial
|> processing than that.

Yeah, but "you don't need to sync them so exactly" :)

Well, I assume both dma channels running a wave.

I think the "initial processing" has to find data that results in 2 waves.
you can get a special value by different combinations of data and volume,
but only the ones building continous waves to both channels imho will
run without noise.

But, if both channels run a wave, you don't need to sync them so exactly,
because if they run a bit asyncron, there is not noise but kind of errorwave.
(you won't loose 14bit for just running 1 word async).

|>
|> -- Mat.
------------------------------------------------------------------------
   [e-mail] (Juergen "Rally" Fischer)   =:)