Page 1 of 1

Tips to recover and convert a difficult wav file.

Posted: Sat Feb 09, 2008 1:30 am
by waskol
Symoon, it's for you !!!

It is a method I use successfully to clean up and reshape most of wav files.

1) First I load this program on my real Oric : Private copy

2) You load each program of the wav file (PC audio output <--> Oric) or of the real tape on your real Oric

3) "Private copy" is ready to save it back...
... on the PC

and you obtain a nice wav file easier to convert into a tap file

:D

Posted: Sat Feb 09, 2008 8:06 pm
by Symoon
Quick reply from RTC connection (how slow ;))

Does this method allow to check parity errors during the reading ?
If not, I'll keep improving the PC tools :)

Posted: Sat Feb 09, 2008 10:56 pm
by waskol
Symoon wrote:Quick reply from RTC connection (how slow ;))

Does this method allow to check parity errors during the reading ?
If not, I'll keep improving the PC tools :)
Really, I don't know, but at least it permits to take benefit of the oric hardware in order to read the audio signal when it is "not nice". Sometimes, the hardware does far better than the software.

Otherelse, about the OricTools, so far I have tried, without success, to use some advanced signal processing techniques (FFT, Goertzel algorithm, Dynamic Time Warping...). I am still searching for something else than peak detection (Fabrice Frances' method) or zero crossing (the method I use in my wav2tape) in order to guess smartfully the good frequencies of the succeeding signal periods, because it's really too basic and not always efficient.

Posted: Tue Feb 12, 2008 8:31 pm
by Symoon
waskol wrote:
Symoon wrote:Quick reply from RTC connection (how slow ;))

Does this method allow to check parity errors during the reading ?
If not, I'll keep improving the PC tools :)
Really, I don't know, but at least it permits to take benefit of the oric hardware in order to read the audio signal when it is "not nice". Sometimes, the hardware does far better than the software.
Indeed, I've been using "real Orics" with Deblock or Multicopy for a long time now, when then PC transfer tools gave no good results. But the weak point of this method is that you can't be sure there were no errors. It takes then ages to do multiple transfers and try to see and correct the wrong bytes.
We'd need modified loading routines checking parity and warning in real-time which bytes might be faulty.
waskol wrote:Otherelse, about the OricTools, so far I have tried, without success, to use some advanced signal processing techniques (FFT, Goertzel algorithm, Dynamic Time Warping...). I am still searching for something else than peak detection (Fabrice Frances' method) or zero crossing (the method I use in my wav2tape) in order to guess smartfully the good frequencies of the succeeding signal periods, because it's really too basic and not always efficient.
I've added an option in wavclean to use what I call "low-level" detection, i.e. using the lowest value of the wave to detect a period. With this method and Fabrice's, I have much more good results than before.
I've also modified a few things in order to transfert tapes that were running too fast or too slow. This has been helpful a few times.

Sadly I'm still in lack of spare time and, IIRC, still waiting for Fabrice's answer about a few questions (well I'll have to double-check if I actually sent them ;)) before distributing this. :?