Hi Karl. I have lots of experience with it - I've probably processed over
1,000 cues in the last couple years. In fact, I did a presentation at "The
Reel Thing" about various de-wow methods and specifically about Capstan.
It really is very simple to use. I haven't seen the demo video, but
basically there is a speed curve, which is generated from the audio file,
and smoothing and intensity curves which you can manipulate. It can work
miracles, but does have some limitations. It does not work at all with
clangorous sounds - bells, gongs, chimes, percussion, etc. Featured
instruments and vocals can also throw it off. And if a featured instrument
or voice is portamentoing around in pitch (my spell checker didn't flag
"portamentoing" so maybe it really is a word!), Capstan will often try to
"quantize" it to an equal tempered grid (that's basically how it works -
similar to Melodyne). Same goes for dialog (I'm in the film post
business). It also doesn't know wow from vibrato, so you have to be
careful to bring up the smoothing curve high enough so that it doesn't
reduce the vibrato. It also doesn't always know bad intonation from wow,
and sometimes it will correct a note that is a little out of tune.
It does rely on a good ear, so you can manually adjust it when it does
mis-track. If you're a musician, especially one who plays non-fretted
instruments, it is a big help, so you can hear subtle differences in
intonation more easily.
As far as less expensive alternatives, there really aren't any. Before
Capstan, I used to use Melodyne (and also the pitch bend wheel on my
sampling keyboard). Melodyne worked great for wow in a piano track. But
there is nothing out there like Capstan, besides Cedar re-speed. I tried
re-speed though and it couldn't fix wow in cues that Capstan could fix
effortlessly. I think re-speed was designed more for using a tone, such as
hum, in the recording to re-clock it, whereas Capstan uses the music itself.
It can handle slowly changing pitch. I used it to fix a recording from a
portable recorder with dying batteries. By the end of the recording, the
batteries were so low, it sounded like it was in fast forward, and the bias
oscillator was becoming audible! It took some hand tweaking, but I was
able to get intelligible playback from it.
It can handle long files, though it seems to work better with files under a
few minutes or so. I seems to have a "window" that is a fixed percentage
of the file length, so if a file is very long, then you can't "get in
there" as tightly and and adjust the curves with as much detail as with
If you don't need it that much, you can rent it for $200/day, so you could
save up some cues and process them all in a day and it might be more
economical. Or you could just send them to us ;-)
On Thu, Dec 26, 2013 at 12:00 PM, Karl Miller <[log in to unmask]>wrote:
> Wondered if any of you might have experience working with the Capstan
> Is the learning curve as simple as it is the demo video?
> Does it work well?
> Are there less expensive alternatives?
> Will it handle a slowly changing pitch…over the length of a side…I am
> wondering about side changes of recordings.
> How long can a file be?
> Any observations–thoughts would be most welcome.
[log in to unmask]