Topic: Curious about what everyone does with matrix

[atxwolfattack]

Do most people use 2 recorders, one for mics, and one for into the board? Or do you use a mixer or a recorder with multitracks, something like that?

I am curious about trying to get a soundboard recording soon, i have never done it before, but from what i have been told, if there is a choice between only sbd or only mics, then choose mics, since not every instrument does through the board.

[wbrisette]

I've done it both ways. I've done a few shows (long ago) where I plugged in one DAT recorder to the board and then another for my mics. The amount of work involved sucked for what ended up being a mediocre recording. I've also tried to use a mixer. The issue that you run into there is the delay that you have to put on the soundboard channels if you're doing a mixdown out of the mixer (as opposed to placing all channels on their own track/channel). Some of the newer ENG mixers and other types of mixers have some delay options, so those help.

Today I use a multi-channel recorder, so I just place everything on a separate track and mix everything later. But, it's an expensive option and obviously not one that is available to everybody.

Depending on what it is you are trying to do and how much work you want to do in post production, you should try the multiple recorder option first since that's the cheapest. Then try the mixer. If all else fails, then try a multi-channel recorder, but it might not be worth it to you to go that route.

Wayne

[bdasilva]

You can pick up a small mixer..  Mix your inputs and record in stereo. there are battery powered mixers. Trying to matrix two seprate recordings just never worked for me. The mixer option however  leads you down the path though because now you need to worry about the time alignment of the mics and board signal. So your mics need to be on the stagelip... in the area between the monitors and the house stacks. Once you figure all this out you want to include more channels in the mix... (See Team Multitrack)

[powermonkey]

I know very little about matrix recordings, but surely as long as the sample rates of the two recordings are the same, then once the two sources are lined up, then there shouldn't be any great difficulty in mixing them? I was under the impression - correct me if I'm wrong, obviously - that the problems only really arose when trying to synch up sources that were different sample rates (and hence one source would be frationally quicker than the other)?

As I say, I know very little about matrixes, so I could be hopelessly wrong... it wouldn't be the first time.

 

[Brian Skalinder]

I know very little about matrix recordings, but surely as long as the sample rates of the two recordings are the same, then once the two sources are lined up, then there shouldn't be any great difficulty in mixing them?

Not that simple.  In order to easily line up two recordings of the same sample rate, they both need to be synced to the same clock, e.g. using a single ADC for all tracks like an Edirol R-4, or syncing two different ADCs (two recent threads about this:  1, 2).  Without the same or synced clocks, even two 48 kHz recordings will differ due to clock drift.  No two ADCs are clocked exactly the same (unless synced), kinda like no two snowflakes are exactly identical.  So no matter what, even if using the same sample rate, using two un-synced ADCs will produce recordings that are not in sync.  Syncing the beginning of each source together is only part of the battle - due to drift, one must re-sync over and over throughout the course of the recording.  How often one re-syncs depends on how far off the clocks are from one another.

[Shawn]

I know very little about matrix recordings, but surely as long as the sample rates of the two recordings are the same, then once the two sources are lined up, then there shouldn't be any great difficulty in mixing them? I was under the impression - correct me if I'm wrong, obviously - that the problems only really arose when trying to synch up sources that were different sample rates (and hence one source would be frationally quicker than the other)?

As I say, I know very little about matrixes, so I could be hopelessly wrong... it wouldn't be the first time.

 

looks like brian beat me to it.

[bgalizio]

Drifting is one of the main reasons I moved to an R-4. I was having so much trouble lining up the JB3 and iRiver sources that I ended up not bothing with a matrix.

[powermonkey]

I know very little about matrix recordings, but surely as long as the sample rates of the two recordings are the same, then once the two sources are lined up, then there shouldn't be any great difficulty in mixing them?

Not that simple.  In order to easily line up two recordings of the same sample rate, they both need to be synced to the same clock, e.g. using a single ADC for all tracks like an Edirol R-4, or syncing two different ADCs (two recent threads about this:  1, 2).  Without the same or synced clocks, even two 48 kHz recordings will differ due to clock drift.  No two ADCs are clocked exactly the same (unless synced), kinda like no two snowflakes are exactly identical.  So no matter what, even if using the same sample rate, using two un-synced ADCs will produce recordings that are not in sync.  Syncing the beginning of each source together is only part of the battle - due to drift, one must re-sync over and over throughout the course of the recording.  How often one re-syncs depends on how far off the clocks are from one another.

Ahhh, I see! Cheers for that... there's clearly more to this taping business than I thought... I think I'll stick to either board or audience for a little while, then, before I try anything more complicated.

[easy jim]

Syncing the beginning of each source together is only part of the battle - due to drift, one must re-sync over and over throughout the course of the recording.  How often one re-syncs depends on how far off the clocks are from one another.

Or, much better, time-scale one source to the other by stretching or shrinking it appropriately so that you do not end up with drift at the end of each track and brief gaps in one source at every point you have to re-sync it.  Time-scaling is much easier w/ software that supports it, and much more accurate IMO, than cut and paste re-syncing...and it is only a couple steps (sync at the beginning, calculate drift, time-scale) instead of very many steps (re-sync and chop between every track).

[Brian Skalinder]

Or, much better, time-scale one source to the other by stretching or shrinking it appropriately so that you do not end up with drift at the end of each track and brief gaps in one source at every point you have to re-sync it.

I've tried a couple apps that support time-scaling and was never able to get it to work to my satisfaction.   Probably use error. 

[Kevin Straker]

Personally I like to do this on the fly. So far the results have been nice. I power my mics with the phantom power on my preamp. Then I send the analog out from the preamp in to a small mixer. I also send the sbd signal to this small mixer. That gives me four channels. Mix these down to two and send it out to your recorder.

[easy jim]

Or, much better, time-scale one source to the other by stretching or shrinking it appropriately so that you do not end up with drift at the end of each track and brief gaps in one source at every point you have to re-sync it.

I've tried a couple apps that support time-scaling and was never able to get it to work to my satisfaction.   Probably use error. 

Some app's are better than others for this, for sure.  AudioDesk/Digital Performer is really good and allows me to get ~ 1milisec accuracy as long as I have a good transient peak near the beginning and at the end to accurately calculate the drift.  I know a lot of video editing software is also very functional with time-scaling.  Other programs are not as accurate, however, and make it difficult/frustrating.

Out of curiosity, what apps did you try and give up on?

Personally I like to do this on the fly. So far the results have been nice. I power my mics with the phantom power on my preamp. Then I send the analog out from the preamp in to a small mixer. I also send the sbd signal to this small mixer. That gives me four channels. Mix these down to two and send it out to your recorder.

Mixing on-the-fly limits you, however, on mic placement for your room pair to within ~ 20-25' of the PA unless you also have a delay box in front of the mixer for the SBD feed.

[wbrisette]

Mixing on-the-fly limits you, however, on mic placement for your room pair to within ~ 20-25' of the PA unless you also have a delay box in front of the mixer for the SBD feed.

That's one of the nice things about the Deva, you have up to 40 ms of delay for each channel. However, even with this, I couldn't get rid of the echo during a recent SXSW show that I was recording and feeding to a camera.  See photos here:

http://homepage.mac.com/wayneb/mixer_1.jpg

http://homepage.mac.com/wayneb/mixer_2.jpg

I ended up feeding the mics only to the camera so they could use those as the guide track, then gave them the timecoded BWAV files so they could put those into their video editing software.

Wayne

[Kevin Straker]

Or, much better, time-scale one source to the other by stretching or shrinking it appropriately so that you do not end up with drift at the end of each track and brief gaps in one source at every point you have to re-sync it.

I've tried a couple apps that support time-scaling and was never able to get it to work to my satisfaction.   Probably use error. 

Some app's are better than others for this, for sure.  AudioDesk/Digital Performer is really good and allows me to get ~ 1milisec accuracy as long as I have a good transient peak near the beginning and at the end to accurately calculate the drift.  I know a lot of video editing software is also very functional with time-scaling.  Other programs are not as accurate, however, and make it difficult/frustrating.

Out of curiosity, what apps did you try and give up on?

Personally I like to do this on the fly. So far the results have been nice. I power my mics with the phantom power on my preamp. Then I send the analog out from the preamp in to a small mixer. I also send the sbd signal to this small mixer. That gives me four channels. Mix these down to two and send it out to your recorder.

Mixing on-the-fly limits you, however, on mic placement for your room pair to within ~ 20-25' of the PA unless you also have a delay box in front of the mixer for the SBD feed.

Yeah, I've only done it from a good celing mount front and center, or from the lip/on stage.

[Brian Skalinder]

Out of curiosity, what apps did you try and give up on?

CEP/Audition and...probably WaveLab (not positive, though, it's been quite a while).

[capnhook]

Personally I like to do this on the fly. So far the results have been nice. I power my mics with the phantom power on my preamp. Then I send the analog out from the preamp in to a small mixer. I also send the sbd signal to this small mixer. That gives me four channels. Mix these down to two and send it out to your recorder.

+T

I did it like this all weekend in the dancetent at Great Blue Heron last year, and had a nice result.  My mics were centerFOB, and the board was only about 24' from the stage.  I used about a 80:20 board:mics mix ---- just enough to get a little of rowdy ambience into the recording.  Take a listen to this one.....nice one-disc d/l........................


http://www.archive.org/details/Donna2006-07-09.buffalozydeco.flac


The little bit of delay helped "wetten up" the result.....sort of a free bit of reverb.   

[dgodwin]

Syncing the beginning of each source together is only part of the battle - due to drift, one must re-sync over and over throughout the course of the recording.  How often one re-syncs depends on how far off the clocks are from one another.

Or, much better, time-scale one source to the other by stretching or shrinking it appropriately so that you do not end up with drift at the end of each track and brief gaps in one source at every point you have to re-sync it.  Time-scaling is much easier w/ software that supports it, and much more accurate IMO, than cut and paste re-syncing...and it is only a couple steps (sync at the beginning, calculate drift, time-scale) instead of very many steps (re-sync and chop between every track).

I spent a couple hours doing the chopping/resyncing today in audition.  On some of the longer tracks, the 2 sources start drifting enough so that is sounds like theres some reverb.  I'll look more into the stretch/shrink method....  http://www.upstatetapers.org/tracker/index.php if you want to hear my attempt.  It's the top show (Ron Hawkins and Lawrence Nichol)

[easy jim]

Out of curiosity, what apps did you try and give up on?

CEP/Audition and...probably WaveLab (not positive, though, it's been quite a while).

Interesting...is anyone aware of a freeware program that offers detailed time-scaling?  I thought you could do it in CEP, but I don't know from experience. 

Once I learned on AudioDesk/Digital Performer since I have the MOTU gear, I've stuck with what works (for me, that is).  In these two programs, you click and 'grab' the end of the wav on your time ruler and drag it to where it 'should be' based on your calculated drift.  The software then re-calculates the samples/wav file and creates a new time-scaled version.

[SmokinJoe]

I realize if clocks aren't synced there will be difference, but with modern digital recorders it should be a consistant linear difference, right?  I.e. if one records a little faster than the other, chances are this will be consistant over the course of an evening?

Let's say I take two sources, then pick a particular drum beat for the beginning, and another one an hour later, and they are seperate by 2.13 seconds, then mathematically, I should be able to dither it by a factor of 2.13/3600 = 0.0005916666, and apply that to the whole file using accurate software, right?  Unless you have a supercomputer, this will probably run very slowly (think over night).  But when I get done, I should have a mixable result, no?  I haven't done it, but that's my plan for when I do.

Now, the real heros are the people to mix 1970's vault tapes with lineage like MSR > Reel(2) > DAT > WAV against a MAC > DAT > CDR > EAC > WAV, because there you have 2 fluctuating analog signals, plus several other clocks.  That would be no where near as consistant as modern digital recordings.  You read stories about dan@amdig doing that for many many hours.