25 Sep
2012
25 Sep
'12
12:05 p.m.
On Tue, Sep 25, 2012 at 11:28:42AM +0200, Atte André Jensen wrote:
On 2012-09-25 08:21, Nils Gey wrote:
Zita-at1 will do the job, but in a rather invasive way. It will
not only resample by some near-unity ratio, but also repeat or
skip cycles in order to preserve to original lenght (it has to,
since it's meant for real time work rather than batch processing).
And while the resampling is of decent quality, it's nowhere near
what the secret rabbit or zita-resampler can provide.
You could still use it to measure the error. An alternative is JAAA
- freeze the display, put a marker on the fundamental frequency and
it will tell you the nearest musical note (for 440 Hz tuning) and
the error in cents. From this you can compute a resampling ratio.
Resampling to the nearest note pitch will change the lenght of a
sample by at most 3 percent, so that's probably acceptable.
The second problem is that you need a resampling app that does
the computations but does not modify the sampling frequency in
the output file header. Never seen such a thing, but I could
write one in a very short time.
Ciao,
--
FA
A world of exhaustive, reliable metadata would be an utopia.
It's also a pipe-dream, founded on self-delusion, nerd hubris
and hysterically inflated market opportunities. (Cory Doctorow)
Quality is the top priority. Since the tuning is
very fine I don't
care if the speed of the recording changes or not. I take what sounds
best.
Although zita-at1 sound really good, I'm pretty sure you'll get the
best result by speeding up/slowing down the recordings, provided
that each file needs the same retuning throughout. I have no idea
how to automate that, though :-(