Cyber ears

[Jos]

Richard Brekne wrote:
> 
> Hi list
> 
> I am intrigued with these computer tuners ie. cyber-ear and tunelab 97. But as I
> dont have access to a portable puter I havent been able to fool around with the
> latter yet.
> 
> What I would like, if anyone out there uses them and would bother to take the
> time to gather the data, is a read out of a fully sampled piano after tunning,
> with as many partials as the program can pick up for all 88 notes.
> 
> I suppose I am asking a lot <grin> and will have to wait till I can justify
> buying my own portable. But I thought I might give it a shot anyways.
> 
> If anyone can be of help you can mail me at
> 
> Richardb@c2i.net
> 
> thanx

Hi!

I tried tunelab

http://www.wwnet.net/~rscott/

but for me, it was only useful to set the temperament  in 
the right direction. I had to make ajustments to the temperament 
afterwards, to be really satisfied,  to remove some of the annoying
beats. But then, maybe I did something wrong...

I think the problem is, that programs base their tuning on the absolute
frequencies measurements of partials. This becomes a major problem,
especially in the higher octaves; it's just difficult to gather accurate
high-frequency data with a pc soundcard and a laptop microphone. Then,
 the absolute frequencies of 
the tone to be tuned has to be moved according to those measurements. I 
think this process is error-prone: imagine you'd have to use absolute
hearing 
to perform a tuning, while your ear-canals would be stuffed with cotton
waddings,
this would be very difficult indeed.

I think, tuning programs should follow the same strategy as aural tuners
do. An aural tuner doesn't tune using absolute hearing, he listens to 
beats! So maybe, a tuning program should try and  listen to beats to:
record
the pitch and modulating frequency of a beat, and use that to tune. This 
would be an advantage in removing annoying beats, which makes an aural
tuners
tuning sound so well. Ofcourse, a program cannot judge, if at a ceartain 
stage the tuning sounds 'well', but I think one could make a program
that
detects all the beats, and then tries to optimize the tuning so as to 
reduce the beats as much as possible. Any programmers on this list?

(Btw, richard, I'll  look if I can find those measurements).

regards,

Jos.

-- 
___________________________________________________________
J.G.A. van Riswick, Eindhoven University of Technology,
Eindhoven, The Netherlands. mailto://j.g.a.v.riswick@tue.nl
mailto://josvanr@xs4all.nl http://www.dse.nl/~josvanr