Board:   [DoveNet] Pro-Audio Discussion
Subject: Re: Tape deck
From:    Nightfox
Date:    Tue Apr 28, 2020 01:56

@VIA: DIGDIST
  Re: Re: Tape deck
  By: Vk3jed to calcmandan on Tue Apr 28 2020 02:47 pm

 Vk> Yeah, can't do that with casettes. :D But I have done edits as precise as
 Vk> a single sample (1/44100 th of a second). In one case, I had 2 copies of
 Vk> an old song off YouTube. One was recorded on an old mono VCR linear track
 Vk> (think crap sound, high noise and inaccurate speed),and the other was
 Vk> recorded using a hifi VCR, which meant the audio quality was excellent.
 Vk> However, an announcer talked over part of this recording.

 Vk> Enter Audacity. First, it was noticed that the two recordings were of
 Vk> significantly different lengths. So I trimmed them to the same start and
 Vk> finish points, worked out the difference and assumed that was due to speed
 Vk> errors on the track recorded on the mono VCR. A couple of iterations and I
 Vk> had the secion I wanted accurate to sub mS precision

 Vk> Next was to identify the edit points where the lower quality mono copy
 Vk> woyld replace the higher quality one. Muting was the best way to handle
 Vk> this, because that preserved the timing. The lower quality file was
 Vk> trimmed after the last edit point and muted before the first edit point.
 Vk> Minor adjustments to timing were made, so the desited section of audio
 Vk> fitted the timing precisely. Test plays confirmed that there were no
 Vk> audible artifacts. Doing the edits at zero crossings helps too.

 Vk> Third stage was to set the relative levels of the two files. This had to
 Vk> be done by ear, because the recordings probably had different audio
 Vk> processing, especially AGC. I also may have done some light noise
 Vk> reduction. Not 100% sure on that.

 Vk> When I was happy, I merged the two tracks and the finished result sounded
 Vk> great.You can pick the edited part, if you listen carefully, because of
 Vk> the different audio quality and higher noise floor, but timing and levels
 Vk> are seamless to the ear. :)

That's some serious editing.  :)  I've done some digital audio splicing, butnot to that level.  Though, I have done some programming related to audio recording & mixing multiple audio files together.  For my senior project in college, I made a simple multi-track audio recorder for musicians.  I wrote it in C++ for Windows.  It would let you record invidiual tracks as separate WAV files, and then mix the audio together to a single WAV file when you're done.  And since then, I've made newer itrations of the WAV mixing code, once in C#, and then in C++ again.  The newer C++ mixing code, I actually wrote not too long ago, around November/December 2018.  One thing I wanted to accomplish by writing that again was to write better code, and also to use modern C++ features.  Also, I thought of a way to mix the audio files a bit faster than I was doing it previously.

Nightfox

---
 ■ Synchronet ■ Digital Distortion: digitaldistortionbbs.com
.