[nas] NAS with XVision

J. Thompson jt at xsound.demon.co.uk
Thu Nov 16 13:43:18 MST 2000


I have just checked the source and Jon is correct on this. The AU format reader
only does a fseek if its not reading from stdin. I few years back I did write my
own AU I/O code because there were limitations with the implementation in NAS,
but thinking back it was because I wanted the output piped to stdout, which NAS
can't handle as there definitely is an fseek.

So to correct my original post, it should be possible to pipe the output of sox
into auplay, provided the output from sox is in the AU format and get good
audio quality.

Jonathan Thompson


On Thu, 16 Nov 2000, Jon Trulson wrote:
>On 15 Nov 2000, Raymond Toy wrote:
>
>> Date: 15 Nov 2000 18:00:26 -0500
>> From: Raymond Toy <toy at rtp.ericsson.se>
>> To: J. Thompson <jt at xsound.demon.co.uk>
>> Cc: jon at radscan.com, Ola Andersson <rand at ling.umu.se>, nas at radscan.com
>> Subject: Re: [nas] NAS with XVision
>> 
>> >>>>> "J" == J Thompson <jt at xsound.demon.co.uk> writes:
>> 
>>     J> real-time under Linux and is optimised for audio quality. The difficulty then
>>     J> becomes how do you pipe the output of SOX into NAS. I'm afraid you can't as the
>>     J> audio I/O bits in NAS don't work with stdin, there is an fseek in the code.  So
>>     J> you have to reasmple your audio files and store the resampled files before
>>     J> attempting to play them, which is probably not what you want. 
>> 
>> Are you sure?  It's been a long time since I looked at the guts of
>> NAS, but I'm pretty sure that Sun's au format doesn't need a seek at
>> all.  So it shouldn't be too hard to modify auplay to read from stdin
>> for au files.
>> 
>> Maybe I'm totally wrong,
>> 
>
>
>	Close... There was a previous thread on this that culminated in a
>post by Charles Levert: http://radscan.com/nas/nas-ml/msg00348.html.  
>
>	Basically, it's the fact that au format is the first in the list
>of openFile hooks that are tried in order to determine the right format of
>the data. The first one (au) wins if it matches, so no fseeks are done in
>the following hooks.
>
>	This would be a good thing to properly handle in the next
>release...  Anybody want to take a stab at it? ;-)
>
>-- 
>Jon Trulson    mailto:jon at radscan.com
>ID: 1A9A2B09, FP: C23F328A721264E7 B6188192EC733962
>PGP keys at http://radscan.com/~jon/PGPKeys.txt
>#include <std/disclaimer.h>
>Free Mars!




More information about the Nas mailing list