[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [fluid-dev] Clicking, continued

From: Josh Green
Subject: Re: [fluid-dev] Clicking, continued
Date: Fri, 28 Jan 2005 11:20:31 -0800

On Fri, 2005-01-28 at 00:52 -0500, address@hidden wrote:
> Hi again,
> I tried my hand at changing the code in 1.0.3 to disable the filter off 
> optimization. It compiled and ran without errors. But the clicks persist, 
> just as before! So I'm thinking that these clicks have nothing to do with the 
> filter off optimization.
> Now, I've made a rather interesting observation. I have two sound cards: my 
> laptop's built-in ESS Maestro and an external USB Griffin iMic. I'd been 
> using solely the iMic until now, so I decided to give my built-in card a try. 
> And guess what, it's click-free. The problem only happens with the iMic.
> So there's something strange about this particular card. Something about it 
> is 
> causing this clicking *only* while FluidSynth is producing the audio in real 
> time. If I record FluidSynth's signal sent to JACK using jackrec, and play 
> back the wave file - through the iMic, of course - there's no clicking at 
> all. But there was plenty of clicking coming out of the iMic while I recorded 
> it.
> I wonder if there's a relationship between my computer's CPU load and the 
> amount of clicking... I also wonder if this is a problem with USB sound cards 
> in general...
> Anyway, I'll keep working at this for now.
> Paul

You may want to try different settings for the audio buffer size and
count, as well as perhaps the sampling rate.  Also make sure you are
running FluidSynth with root privileges (I imagine you are already,
since you mentioned you get click free output with another sound card).
The switches for the audio buffer are:
-z for audio buffer size (try numbers like 128, 256, 512, etc), -c for
audio buffer count (1, 2, 3, 4, etc).  The higher the numbers the higher
the latency of course.  Some cards like certain combinations better it
seems.  Its unlikely that its the sample rate, but you could always set
this as well with -r (48000 might be worth trying).

        Josh Green

Attachment: signature.asc
Description: This is a digitally signed message part

reply via email to

[Prev in Thread] Current Thread [Next in Thread]