Re: [Audacity-devel] How about trying DTMF Tones?

[Gale A. <ga...@au...>]

| From Leland <le...@au...> 
| Sun, 06 Jul 2008 05:25:06 -0500
| Subject: [Audacity-devel] How about trying DTMF Tones?
> Gale Andrews wrote:
> > | From Leland <le...@au...> 
> > | Sun, 06 Jul 2008 02:46:03 -0500
> > | Subject: [Audacity-devel] How about trying DTMF Tones?
> >> Try generating DTMF tones using 15 hours...you might get the impression 
> >> that Audacity is hung.  Progress dialog isn't taking all processing into 
> >> account.
> > 
> > We have a general note about all Tone Generators in "Release Checklist
> > not aiming for 1.4" :
> > 
> > "Long tone generations (e.g. 20 minutes) can crash Audacity from time to
> > time. i.e. the tone has not still not been generated after 30 minutes or more
> > and Audacity is no longer responding. Consider if the theoretical 99
> > hours, 59 minutes, 59.999 seconds generation allowed should be
> > restricted."
> > 
> > I observed the crashing on my older XP machine. On this more recent one 
> > I've been generating a 15 hour tone for five minutes in which time the 
> > "remaining time" has gone up from 20 to 45 minutes and CPU is 
> > about 15 -20 %. This isn't really useful. Should we escalate to main 
> > checklist with the intention of doing something about it e.g. generate 
> > a short tone and essentially repeat it (job done instantly)? 
> > 
> > Generate Tone works fine here. 
> > It's the DTMF fella that isn't including all of its processing
> > within the ProgressDialog updates.
> 
> I don't know about the other issue that you mention, but I'm starting to 
> suspect that we are still running into sample count overflows somewhere.

If I generate 1 hour DTMF and normal Sine tone, I can see that DTMF only
updates elapsed time every 10 seconds, Tone every second. I also found 
it very difficult to cancel DTMF, it took 30 seconds and several attempts
to do so, but Tone cancelled almost at once.  

Both generators suggest 1 minute ETA at start, but end up taking 3
minutes, and I don't like the amount of CPU this is taking either. 

So when you say "Generate Tone works fine", how long does it take to 
produce a 15 hour tone? Is anyone going to wait an hour to do that?
And not everyone has fast machines, I honestly believe long tone
generation could crash slower machines. 

And now I tried to generate (at 44100 Hz) a 14 hour silence, I *do* see
what you do, an empty track, but not generating a 12 hour silence, where
I get the zero samples. I get the same result generating at 96000 Hz.
So it seems we are up against the same type of limit you have been
seeing with Zac's issue, but it is not related to number of samples? 


Gale (Win XP)