[Ardour-Dev] SuperRapidScreenUpdates are neither rapid, nor super.

John Emmas johne53 at tiscali.co.uk
Tue Mar 10 05:09:52 PDT 2009

----- Original Message -----
From: "Kjetil S. Matheussen"
Subject: Re: [Ardour-Dev] SuperRapidScreenUpdates are neither rapid, nor
>>> Are you using the closed-source binary driver from nvidia on the
>>> laptop?
>> I don't know, to be honest.  It's running 64studio 2.2 which was
>> installed about a year ago from a CD image.  It'll be using whatever
>> driver gets supplied by 64studio
> Probably not. Installing the binary driver from
> http://www.nvidia.com/Download/index.aspx?lang=en-us
> will probably solve your problem.
I might have to leave that for a week or two because I can't afford any
unexpected driver problems just at the moment,  It does puzzle me though
that I'm getting identical refresh timings on different machines with
different hardware which leads me to believe that Paul is right about this
problem being caused by changes to Ardour's code.  The only bit I don't
understand is this:-

----- Original Message -----
From: "Paul Davis"
> that said, nick mainsbridge did come up a way (a rather obvious way, in
> retrospect) for us to use the canvas and make this work. but it involved
> ceasing all use of the canvas's own scrolling/zoom mechanism, and setting
> up redraws in our own code.
It's entirely possible that I've misunderstood this (or maybe I'm looking in
the wrong place) but that doesn't tie in with what I can see (at least, not
in Ardour 2).  Scrolling the canvas in 'moving timeline' mode appears to be
accomplished by programmatically 'repositioning' its horizontal scrollbar
and letting gtk do the work.  Unless Nick's changes were done a long time
ago (or after version 2.7.1) that seems to have been the mechanism for a
very long time (at least dating back to Ardour 2.4)..

I'd love to be proved wrong but from what I've observed, around 90 percent
of Ardour's work is being done in a single thread - and the more work that
gets added to that thread, the slower it gets.  Therefore signals that were
originally intended to fire every 1/40th of a second (or whatever) simply
cannot be achieved (at least, not on my hardware) because the thread is
bogged down doing too much work.  Secretly, I'm hoping that I'm just totally
wrong about this.  I'd love to think that updating my graphics driver will
kick in some currently unutilised hardware acceleration or something like
that.  But at this point in time - and expecially, given the evidence that
the problem has crept up slowly and progressively - I'd be surprised if it's
anything more complicated than an overworked thread.

It has to be said though - I've been surprised many times before....  :-)

Best regards,


More information about the Ardour-Dev mailing list