Re: lmc bugginess

Kristian G. Kvilekval wrote:
>>I just wondered if we needed to distinguish (both internally and in the
>>ui) between an estimated length, and an actual known length from a file
>>that has been played before?
>>
>>Rob
> 
> 
> I'm not usually that bothered by the time being a little off, but
> I guess it's better to be right.

actually i thought i was pretty clear that it was better to not be 
right.  are you implying that we should be scanning (reading) all the 
files we add to the playlist?
What we do now does not give us the "right" times, it makes a guess 
given the data available to it. With mp3s it uses the first data header 
or xing tag if it has it compared to filesize and with ogg, well it 
should use bitrate_nominal but instead it does read the entire file (not 
for long though).  The way we determine time length is the best way for 
zinf, the more cpu intensive but accurate methods should be left to 
other programs or be non-standard and hence not-priority right now in zinf.

> 
> I like this idea of storing the time after playing in some metadata
> field and use that when available instead expensive calculations.

even if you had an option to determine exact time lengths or did it 
automatically after playing a file, you still wouldn't be able to 
accurately seek in a file. If there is a metadata tag for duration, it 
should be written as an option, automatic writing to files is not a good 
thing, we should make sure the user has to explicitly give zinf 
permission to write duration data to files being played or perhaps to 
files being added ... etc.






-------------------------------------------------------
This SF.net email is sponsored by:
The Definitive IT and Networking Event. Be There!
NetWorld+Interop Las Vegas 2003 -- Register today!
http://ads.sourceforge.net/cgi-bin/redirect.pl?keyn0001en