Topic: matrixing analog sources. How to synch them?

[nicegrin]

Hi!

I´d like to matrix 2 recordings (both from analog sources). Any program or suggestions how to do this?
I have made matrices from digital sources before which is alot easier due to the speed diferences of the recordings being linear,
but how to do when you want to synch 2 analog sources (with unlinear speed)?

Thanks
N

 

[mattmiller]

I've never had to do it, but I think the general consensus is that there's no way to do it other than chopping the sources up into as small of pieces as you can tolerate and sync them individually.

[nicegrin]

yeah, that was my original approach aswell and I managed to synch 4 min of recording pretty good after chopping it up in peices of 30 seconds and then I kind of got tierd of it! So I was curious to learn if there is any tool that can manage this in a smarter way!

[admkrk]

did you try stretching? i think that would be more appropriate here then matching a delay(as in sbd/aud sources).   

[su6oxone]

I'm not sure what you mean by linear/non-linear speed, but if you're just trying to sync up two audio sources one of the easiest ways is to use Sony Vegas.  You just open both tracks, line them up at the beginning so that they sync up, and then go to the end of one of the tracks and 'stretch' it so that it matches up with the other track.  That way, both the beginning and end (and presumably the middle parts) will be synced up easily (press '~' while clicking the end of one of the tracks and then move it left or right to sync with the other track).  There is a bit of a learning curve, but once you get a hang of it it's actually quite quick and easy.  I recommend using discrete sounds (like cymbal or drum beats) to help sync them up. 

[nicegrin]

I'm not sure what you mean by linear/non-linear speed, but if you're just trying to sync up two audio sources one of the easiest ways is to use Sony Vegas.  You just open both tracks, line them up at the beginning so that they sync up, and then go to the end of one of the tracks and 'stretch' it so that it matches up with the other track.  That way, both the beginning and end (and presumably the middle parts) will be synced up easily (press '~' while clicking the end of one of the tracks and then move it left or right to sync with the other track).  There is a bit of a learning curve, but once you get a hang of it it's actually quite quick and easy.  I recommend using discrete sounds (like cymbal or drum beats) to help sync them up.

By linear I mean that the speed is constant so when stretching source 1 to match source 2 at the end they will match throughout from the beginning to the end. This is the case when mixing digital sourced stuff, but for analog tapes the speed varies so much that you will have missmatches in the middle of your wave-form when you match the beginning and the end.
I´d really like to see a function where you define two points on one of your your waveforms and then fix point one and "graphically" stretch/compress the other to match the corresponding point on the other waveform as you need to match points like every 10th second or so not to get them too much out of synch.
Hope that made sense! 

N
 

[Charlie Miller]

You need to make sure that they are both on pitch before you do anything else. That will make things much easier.

[nicegrin]

You need to make sure that they are both on pitch before you do anything else. That will make things much easier.

In what way would that be easier?

[su6oxone]

By linear I mean that the speed is constant so when stretching source 1 to match source 2 at the end they will match throughout from the beginning to the end. This is the case when mixing digital sourced stuff, but for analog tapes the speed varies so much that you will have missmatches in the middle of your wave-form when you match the beginning and the end.

Okay, I see what you're saying now... that does sound like a challenging situation. 

[mattmiller]

By linear I mean that the speed is constant so when stretching source 1 to match source 2 at the end they will match throughout from the beginning to the end. This is the case when mixing digital sourced stuff, but for analog tapes the speed varies so much that you will have missmatches in the middle of your wave-form when you match the beginning and the end.

Are digital recordings even necessarily constant speed from beginning to end?  I use two digital recorders to pull MIC and SBD sources, and the clocks on them vary from night to night and, most troubling, from SET TO SET on a given night.  For every set I have to carefully measure the length from point A to point B on both sources, and the correction amount is always different.  My M10 (SBD) source is always shorter than my HD-P2 (MIC) source, but sometimes it's shorter by only 2,000 samples over a 90 minute set and sometimes it's shorter by 6,000 samples.  If the clocks vary from night to night, and from set to set, is there anything inherent to the way the clocks work that keeps them from varying from sample to sample?

[pistolpete71]

 is there anything inherent to the way the clocks work that keeps them from varying from sample to sample?

Not sure if this will help or hurt, but it's a good read on the pitfalls of digital recording,  the section "Evil #1 - Jitter" may be of particular interest:

http://www.benchmarkmedia.com/discuss/feedback/newsletter/2010/08/1/unique-evils-digital-audio-and-how-defeat-them


 Good luck with the clocks, that would make me loose my mind.

[faninor]

I second the suggestion of Sony Vegas to fix non-linear drift. It is super easy to split one track at different places and stretch the chunks to fit the other recording. Won't be fun, but it's the only application I've tried that I would use to attempt this. Just make sure you have the setting that auto-fades splits turned off before you start.

[ghellquist]

Are digital recordings even necessarily constant speed from beginning to end?

No. The clock frequency varies all the time. Changing temperature of the clock crystal and surrounding capacitors  is probably the largest contributor, but other things do influence such as changing supply voltage around the crystal drive circuits.  This is why many high-spec crystal oscillators as one counter measure has the clock in a small owen that is kept at a constant higher temperature. In addition these clocks are generally calibrated to be quite close to the exactly correct frequency. Cheaper circuits will generally be further off from the exact frequency.

If the clocks vary from night to night, and from set to set, is there anything inherent to the way the clocks work that keeps them from varying from sample to sample?

Nop. They vary all the time. Typical result inside a box though is that the frequency slowly changes. But there can also be larger or smaller changes going on at a faster pace, even down to a scale of a few samples. When you connect two boxes together, this "sample-to-sample" variation generally gets much larger.

// Gunnar

[page]

Are digital recordings even necessarily constant speed from beginning to end?

No. The clock frequency varies all the time. Changing temperature of the clock crystal and surrounding capacitors  is probably the largest contributor, but other things do influence such as changing supply voltage around the crystal drive circuits.  This is why many high-spec crystal oscillators as one counter measure has the clock in a small owen that is kept at a constant higher temperature. In addition these clocks are generally calibrated to be quite close to the exactly correct frequency. Cheaper circuits will generally be further off from the exact frequency.

I guess that's a real case for wordclock connections.

[stevetoney]

Personally, I think I'd just suck it up and go a minute at a time using software that gives you the most control to stretch and shrink conveniently and quickly.  Some people have said Vegas...I can't disagree.  Decide which of the two tapes has the more pitch accurate sound and then reference the other source to that one for stretching or shrinking. 

If you try to take bigger chunks you'll just gonna end up breaking it down into smaller pieces anyway. 

Also, big chunks are great, but what you usually will do is match it end to end.  That works fine until you discover something in the middle that isn't matched correctly because of tape stretch (isn't that called wow or flutter).  So the middle section needs to be adjusted, but as soon as you move something in the middle, you've just messed up everything at the end. 

So that's why I just say to build the master piece by piece from the start.  I don't know of any software that does this.