New Release of GuitarTune v0.20 (based on Alsa)

Ed Kapitein ed at
Fri Feb 6 10:02:03 CET 2009


Thanks for the reply.
I put my comments inline, so see below.

On Thu, 2009-02-05 at 05:11 -0800, c_c wrote:
> Hi,
> Ed Kapitein wrote:
> > 
> > ... so when playing a G the tuner hears a A.
> > 
>  Well, I only own two accoustic guitars - and it seems to be quite accurate
> for me. But then fundamental freq detection is not a guarantee with the
> signal processing I'm doing. There are many things that can throw the tuner
> off - like background noise for eg or the particular characteristics of the
> guitar.
> So, here are some questions that would help me understand the approach to
> take to make the program more accurate.
> Are you using an accoustic guitar? 
Yes, vincente sanchiz 38B, beautiful instument btw.
I do have another one, a bit older and of less quallity, i will try that
one later tonight.

> Is the tuner consistently hearing a higher note?
It switches every now and then. the G is almost always regonized as an
(out-of-tune) low A, but sometimes it switches to te correct G.
> Could you verify by using some other tuner what the error is?
the other tuner i have has no frequency scale, just a led bar indicating
way to low, to low, right, to high, way to high.

> Could you try after keeping the freerunner really close to the guitar?
I almost inserted it! without any difference. i also adjusted the volume
to no avail.

I could record the sound of my strings and send them to you, if that
would help debugging the problem.

Please let me know if you need more info, i will make some pure sinus
waves at the right frequency, to see if that does work.

Kind regards,

> Ed Kapitein wrote:
> > 
> >  is it possible to select manually which string you try to tune? it could
> > filter the input signal at that frequency and perhaps make recognizing the
> > correct tone easier.
> > 
> Well, the current method I'm using (FFT followed by a harmonic spectrum
> product) wont be affected by knowing which string you're tuning. 
> Then again, I could use another method of determining the fundamental freq
> (goertzel algorithm) perhaps, which would need to know the target freq.
> Though that will be totally different program.
> So, right now, I would prefer trying to get some feedback and improving the
> code to see how it can be made more accurate. And your feedback would help
> the most since you're the first one to tell me its off :-)

More information about the community mailing list