synth specific settings. was: delaytime of litetalk driver

Chris Nestrud ccn at uark.edu
Tue Jan 23 15:03:43 EST 2001


Thanks for the information. I had decided not to attempt much more in the
way of experimentation without getting some clarification beforehand.

Chris


On 23 Jan 2001, Kirk Reiser wrote:

> It has become quite popular to play with these settings since Frank
> discovered them a few days ago.  You all may have noticed that you
> need to be root to make changes to these settings.  The reason for
> that, and there is a reason, is that they can fuck up your system big
> time if you're not careful.  The common consensus is well okay
> sometimes they may hang my system and I'll have to reboot; well they
> can have much more violent reactions than that.  If they are set to
> far a field they can infact currupt your file system.  Okay you've
> been warned.
> 
> Let's just get the functions of these variables straight.  Delay_time
> is the amount of time that speakup goes to sleep after sending
> characters to the synth.  Speakup sends out jiffy_delta worth of
> characters before going to sleep for delay_time.  Speakup does not
> start talking to the synth until trigger_time worth of characters have
> been sent to the buffer.  Trigger_time is a one-shot in that it
> doesn't get set again until the buffer has been cleared.  Jiffy_delta
> is the amount of time speakup holds the kernel while sending
> characters to the synth.  A word about these values, delay_time,
> trigger_time and full_time are in miliseconds.  Jiffy_delta is in ten
> milisecond increments.  So a trigger_time of 50 is the same as five
> jiffies.  Jiffies are the basic scheduling increment of Linux on
> 32-bit processors they are ten miliseconds.  On systems such as alphas
> and sparcs they are one milisecond.  The higher jiffy_delta is the
> longer you hold the processor.  This wouldn't be bad except that we're
> talking multitasking here, so what you take someone else doesn't get
> in any given time period.  We can go into a rather indepth discussion
> about interrupts and all that but I'd rather save that for the
> reflector if anyone is interested.  The bottom line is you grab to
> much time and you'll have a totally fucked file system, network layer
> and who knows what else.  So be careful.
> 
> The default values are not optimal because of the amount of time I
> have to spend with each synth.  Some drivers I never even have the
> synth to test with.  We have already incorporated some of the new
> values that people like Frank and Bill have come up with.  Over time
> we will hopefully be able to tune all the different synths.  I just
> want you to  know what you are dealing with if you are going to muck
> about.  I was worried because it looks to me as if you didn't know
> what these values are for Chris.
> 
> So you've all been warned.  If you fuck your systems over don't even
> bother mentioning it to me.  meanwhile happy tuning.  If you get better
> values for your synths please post them to the list and on the
> reflector.
> 
>   Kirk
> 
> -- 
> 
> Kirk Reiser				The Computer Braille Facility
> e-mail: kirk at braille.uwo.ca		University of Western Ontario
> phone: (519) 661-3061
> 
> _______________________________________________
> Speakup mailing list
> Speakup at braille.uwo.ca
> http://speech.braille.uwo.ca/mailman/listinfo/speakup
> 





More information about the Speakup mailing list