I will take another look at the SDIF spec. If the format is still fluid,
than this may be an excuse to get it solidified. I remember feeling that
too many essential items of information were delegated to optional
chunks.

A pvoc analysis file can be megabytes in size - easily larger than the
soundfile it represents. Can anyone really want to deal with it in text
form?

Richard Dobson


Michael Gogins wrote:
> 
> I suggest you use SDIF as this is more in keeping with the public-domain
> character of our work. I also believe it would be technically adequate.
> 
> As an alternative, I suggest using Extensible Markup Language (tagged text
> files) as it is easier to code readers and writers for this type of format,
> and to use it from and for other applications.
> 

-- 
Test your DAW with my Soundcard Attrition Page!
http://wkweb5.cableinet.co.uk/rwd

Richard Dobson wrote:

> A pvoc analysis file can be megabytes in size - easily larger than the
> soundfile it represents. Can anyone really want to deal with it in text
> form?

Plenty of people would want to, I would say, but there is probably more to  
be gained by keeping it binary. To keep both camps happy, concurrently  
distribute some C routines to convert to and from the a text format. These  
could exist as stand-alone utilities, or a static lib.

Whoever designs the data structure is going to have to write routines to  
access the data anyway, so it shouldn't be a big job.