At 11:24 PM -0800 1/9/05, Rudy Trubitt wrote:
>I've done a few things with two recorders and assembled them into a
>single mix in editing. As Dan says, there is drift. I had about a 45
>minute concert which drifted by an un-usuable amount from beginning to
>end. Here's how I fixed it:
>
>I began by measuring the difference in time from a sharp transient near
>the start of the recording to another transient another near the end. I
>then used this ratio and calculated a small sample rate conversion,
>followed by a header change. (for instance, convert one file from 44.1k
>to 44.087 kHz, then do a header change (not a sample rate conversion)
>back to 44.1 (this makes it play back at an ever-so-slightly higher
>pitch that before). At that point, the two 45 minute stereo files were
>really, really close in length, almost no drift. I tried time aligning
>the front and rear, but ended up liking the sound of the rear channel
>about 15 milliseconds behind the front, so as Dan points out, a bit of
>drift when you've got a big offset is not going to break anything.
>
>Rudy
>
>
>On Jan 5, 2005, at 4:09 AM, wrote:
>
>>>
>>> Multi-track using multiple recorders. Something I have considered
>>> before, but always dismissed. I may need to reconsider...
>>
>> Of course they drift, about one foot (sound travel time) per minute
>> is typical. But if the rear pair is fifteen to forty feet away from
>> the front pair, that won't matter. As Rich Peet said, effectively
>> decorrelated.
>>
> > -Dan Dugan
Bet I'm missing something here. As long as the signals are not mixed,
wouldn't a header change be enough? 44 100 / 44 087 =3D 1.00029487.=20
Think anyone could detect such a tiny pitch change?
With +15 feet mic spreads and two MT-90's last summer, I was
consistently getting ~1/30 second or ~33ms differences with 80
minute takes (digital transfers). Perhaps my recorders "speeds" were
closer matched, but I was able to get no more than ~17ms difference
by aligning the head slates, cutting the longer file in the middle,
dragging the 2nd half backwards until the tail slates matched and
cross faded the over-lap. MD drift did not surface as an issue at
all; there were only a few situations where some additional sliding
seemed to make a better image. Rob D.
--
Rob Danielson
Film Department
University of Wisconsin-Milwaukee
[Non-text portions of this message have been removed]
________________________________________________________________________
________________________________________________________________________
|