Mic solutions - on/off camera - TC synching

[Alex Filacchione]

OK I am looking at several solutions and need a couple of opinions...

My camera is a PD-150.

I am looking to get a decent, cheap wireless lavalier mic. Right now I have narrowed it down to the Audio Technica Pro-88W and the Azden WMS pro.

Both have pretty much the same specs. Any opinions on reception quality? Audio quality? This would be for a single person in front of the camera speaking. THere will be times when that person is talking or interviewing someone else, so I would need the capability to easily & cheaply replace the lavalier MIC with a standard hand held mic. So, even though I am not buying a handheld unit just yet, that IS in the future, and the quality of the handheld units that work with the recievers WILL come into play in my final descision. I am leaning towards the Audio Technicas for that reason, but not set on it.

For recording indoors in a stationary environment (recording a concert, or several people in a room) I am probabky going to get a pair of Oktava MC-012s and get the Hypercariiod capsules for them.

So my question her relates to recording on or off camera.

My options are running the mics directly into the PD-150, or into a mixer and then into the PD-150.

However, if possible, I would rather record to a PC instead just because it gives me the flexibility of moving the camera around and not worrying about mic cabling. I can set up the mics, send them to a mixer and into a PC, and just leave that alone as I reposition the camera as needed, esp if doing anything hand-held in a room, which I see myself doing a lot of.

I would record into Steinberg NuEndo on a PC probably using a USB recording interface, running at 16/44.1

My understanding is that the PD-150, when recording in either DVCam mode, or recording directly to disk, gives me true SMTPE timecode, whereas recording in MiniDV mode gives me timecode, but not true SMPTE.

So, I would record the picture w/ the PD-150 (DVCam) and the audio to a PC. Both would be SMPTE timecoded, and then I can merge the audio & video together on post.

Am I asking for trouble doing this?

Opinions? Tips? Comments?

Thanks!

Alex F

[Matt Stahley]

I would record directly into the 150 esp. if its mainly dialog.No need for a mixer unless you are using more than 2 mics at a time.Just to simplify for post.If using 3 or more mics I would want to multitrack all the mics so i have total control in post so the PC may be a good idea. Record each mic to its own track.I dont think USB could handle more than 2 inputs at a time so firewire would be a better option. Ive never had a problem syncing audio to video by recording a hand clap at the start of each shot with the on cam mic and a boom.Also you would want to record on the PC at 16/48k since the dv audio is at that sample rate to begin with and will save some time when you dump everything into your NLE.

[Douglas Spotted Eagle]

I'd go with the AT for the following reasons:

1. No compander
2. Awesome sound for low cost, but keep distance short
3. Great customer support from AT
4. I've never had good luck with anything Azden, and never known any professional happy with anything Azden.
5. You see the cheap AT's on quite a few film sets, so even Hollywood likes em'.

[Alex Filacchione]

I have never used any Azden stuff, but I have seen their catalogs and heard others' opinions. They have always seemed very overpriced for the quality, so I figured that between the Azden & AT at around the same price the AT would be much better, but you never know. Behringer has a couple of good products amongst a sea of really bad stuff, so the same might have been true for Azden. Thanks for the opinion, I'm going to get the AT.

As far as Hollywood using AT, well I have never seen any live musicians actually using Azden for wireless either. :-) In the wireless world *usually* you get what you pay for, and when for the same price you can get a better Samson or Sony or AT.

Too bad X-Wire units are all but impossible to find. Those were digital wireless units that were simply amazing, esp for instruments. THey could simulate cable length and sounded better than anything else. THey were around $900 for the unit. The company was supposed to come out with a half priced model, but went under before they could.

As far as the rest of it goes, I will be recording (when setting up the stationary mics) sometimes talking (interview type stuff) and sometimes music (live, in a recording studio, people playing solo, etc.) I want to get all of that in stereo. Using a mixer or even recording via a USB soundcard and micing in software will give me more flexibility and control, and probably better sound since I can control the quality of the pre-amps, and the audio engine that is doing the recording and converting. That and it will be easier in terms of if I have to move the camera as mentioned before.

As far as synching up with a clap, my concern is that of drift, esp. if it is a music video or a concert film. I am concerned that if I do a lot of post processing (effects or otherwise), that the video will start to drift from the sound. I do have tools that can fix the audio to match the video and you won't notice, but it is a tedious and long process!

Are these drift concerns founded, or have you guys in general not really noticed this, or what?

BTW, thanks for the 16/48 tip. I never realized that minDV, etc. recorded in that format.

Thanks,

Alex F

[Phil Lemoine]

did you HAVE TO HAVE an xlr adapter for you azden wms pro? or is it possible to just plug it directly into the microphone input... i have a gl2...

[Steve House]

Originally Posted by Alex Filacchione

...

I would record into Steinberg NuEndo on a PC probably using a USB recording interface, running at 16/44.1

My understanding is that the PD-150, when recording in either DVCam mode, or recording directly to disk, gives me true SMTPE timecode, whereas recording in MiniDV mode gives me timecode, but not true SMPTE.

So, I would record the picture w/ the PD-150 (DVCam) and the audio to a PC. Both would be SMPTE timecoded, and then I can merge the audio & video together on post.

Am I asking for trouble doing this?

Opinions? Tips? Comments?

Thanks!

Alex F

Why record using 16/44.1? The standard sample rate for video is 48kHz and that's what you should use if you're going double system to a an external recorder or a laptop.

The problem you're going to run into with using timecode for sync is there's no convenient way to get 'code from the PD150 to the recorder or computer or vice versa, AFAIK the PD150 doesn't have timecode in or out. That means that while your audio will have a timestamp and the video will have timecode, they won't be slaved to each other so as to make them identical, thus destroying their usefulness in establishing frame accurate sync. You'll still need a clapper slate on each scene in order to align the audio and video. Whether the PD150's code is SMPTE or not is irrelevant.

[Peter Moretti]

Originally Posted by Steve House

...
The problem you're going to run into with using timecode for sync is there's no convenient way to get 'code from the PD150 to the recorder or computer or vice versa, AFAIK the PD150 doesn't have timecode in or out. That means that while your audio will have a timestamp and the video will have timecode, they won't be slaved to each other so as to make them identical, thus destroying their usefulness in establishing frame accurate sync. You'll still need a clapper slate on each scene in order to align the audio and video. Whether the PD150's code is SMPTE or not is irrelevant.

Steve, some cameras have time code in/out, e.g. the XH-G1. And some recorders have time code, like the Sound Devices 702T. Is it necessary to have BOTH camera and recorder with time code, or is necessary to only have one with time code, since I believe you slave one off of the other?

And if only camera OR recorder time code is necessary, is it better for audio purposes to have the camera generating the time code or the recorder generating the time code, or does it not matter?

Thanks a lot (and I hope this ? made sense, LOL)!

[Steve House]

Originally Posted by Peter Moretti

Steve, some cameras have time code in/out, e.g. the XH-G1. And some recorders have time code, like the Sound Devices 702T. Is it necessary to have BOTH camera and recorder with time code, or is necessary to only have one with time code, since I believe you slave one off of the other?

And if only camera OR recorder time code is necessary, is it better for audio purposes to have the camera generating the time code or the recorder generating the time code, or does it not matter?

Thanks a lot (and I hope this ? made sense, LOL)!

If you're recording double system, i.e. a recording device separate from the camera, and you're using timecode to establish sync you need code on both devices and they must be set identically to less than a maximum of 33 milliseconds deviation (the length of one frame). If someone in the scene claps their hands one time and the clap occurs at 00:14:37;18 (zero hours, 14 minutes, 37 seconds, and 18th frame) on the timecode recorded with picture, the sound of the "pop" in the sound recording must also occur at 00:14:37;18 in the timecode recorded recorded with the sound. This means the camera and the sound recording device must talk to each other and either the camera sends code to the recorder or the other way around (the norm in video production is sound slaves to video, for film it's the other way around but via a smart slate - most film cameras don't do code).

Slaving on off the other means they have to talk to each other. If the camera is the master it has to send timecode to the recorder so it must have a timecode out connector. Using the audio recorder as the master means the camera must receive code from the recorder, hence it requires a camera with a timecode input connector. The PD150 has neither. Ambient has annouced a reader that can retrieve camera timecode from its LANC terminal but I'm not certain they're shipping yet and it won't be cheap when it does.

You also have to complication not only of establishing sync but the problem of maintaining it over long'ish shots. Timecode in itself only serves to establish sync but if the sample clocks (different completely from the timecode clock) aren't running at absolutely identical rates the sync will drift over time, the length of time it takes to become noticable depending on how close or far away the two clock rates are from each other.

[Steve House]

Originally Posted by Alex Filacchione

...
As far as synching up with a clap, my concern is that of drift, esp. if it is a music video or a concert film. I am concerned that if I do a lot of post processing (effects or otherwise), that the video will start to drift from the sound. I do have tools that can fix the audio to match the video and you won't notice, but it is a tedious and long process!

Are these drift concerns founded, or have you guys in general not really noticed this, or what?

BTW, thanks for the 16/48 tip. I never realized that minDV, etc. recorded in that format.

Thanks,

Alex F

FWIW, I'd record 24bit/48 kHz or 24/96.

Drift isn't an issue with most scenes but then most shots are fairly short. But you're talking about music concert shoots it's going to be a different ballgame. "Austin City Limits" is on the tube in the background as I type this and if that's the sort of thing you're looking for as concert footage be prepared for some sticker shock on the hardware they use to get everything coordinated and synced up.

[Peter Moretti]

Originally Posted by Steve House

...
You also have to complication not only of establishing sync but the problem of maintaining it over long'ish shots. Timecode in itself only serves to establish sync but if the sample clocks (different completely from the timecode clock) aren't running at absolutely identical rates the sync will drift over time, the length of time it takes to become noticable depending on how close or far away the two clock rates are from each other.

Steve, could you possilbly explain the difference between timecode clocks and sample clocks? And is synching sound and video for long scenes only an issue when you are not slaving?

BTW, I just ordered a two books on the subject of sound and will be ordering an third, so hopefully these basic ?'s will end soon. THANKS for all your help.

[Steve House]

Originally Posted by Peter Moretti

Steve, could you possilbly explain the difference between timecode clocks and sample clocks? And is synching sound and video for long scenes only an issue when you are not slaving?

...

As I'm sure you know, all digital recorders convert the analog signal to digital by sampling it X times a second and recording the resulting string of numbers. 44.1kHz means that 1 second of audio contains 44,100 samples, the video standard of 48kHz means 1 second contains 48,000 samples, etc. The device's sample rate clock is the 'timepiece' that controls this process and all digital recording and playback devices have 'em. So imagine you're recording some sound destined to be matched up later to some video you're shooting at the same time but your audio recorder's clock is a terribly fast, running at, say, 50kHz instead of 48kHz like it should. If we record a 10 second scene, the 10 seconds of sound will be represented by 500,000 samples. Now we load it into our editor to sync it to picture. The editor enforces the standard and says each 48,000 samples we've given it represents 1 second of sound so as far as its concerned, that file we gave it containing 500,000 samples is 10.42 seconds long and that's how it plays it back. Meanwhile, the picture file we've given it, recorded in a camera that DOES have an accurate clock, plays back in exactly 10 seconds. So if we have lined them up so they start together, after 10 seconds the picture is done but the sound still has almost half a second left to go - they've started in synch but drifted apart. Now real world clocks are pretty accurate these days so the situation isn't as severe as my example but the principle still holds true and sound will inevitably drift out of synch with picture over time unless measures are taken to make sure the sample clocks in the camera and the audio recorder are running at exactly the same rate. Cheap recorders like iRivers and consumer mini-discs will drift out relatively quickly; professional recorders like the Sound Devices series (just as an example) will be accurate enough to hold to within a frame for perhaps 15 to 20 minutes or so but they'll all eventually drift.

A single master clock called "house synch" is commonly used in broadcasting to drive all the devices in a facility through a process called "genlock" to insure multiple cameras and audio recorders are running together. For field production a set of clock devices such as Ambient's Lockit boxes that can be tuned to each other can allow you to keep cameras with genlock inputs and audio recorders with wordclock inputs locked together to within 1 frame every 24 hours accuracy without a physical connection between them but be prepared for a bit of sticker shock - they run about US$1000 each and you need one for each camera and recording device.

Timecode is based on what is essentially a time of day clock and there are a lot of variations on how it is recorded and used, with differences between various film and video workflows and differences with the various methods of recording audio (analog, DAT, file-based) as well. While in the device itself, it may ultimately be derived from the same timebase oscillator as the sample clock, it is a separate counter altogether, meaning that it is essentially a second clock. It timestamps the video file with the exact time each specific frame of video was recorded and in the audio file the time a specifc instant of sound was recorded. But for those recorded timestamps to be useful for aligning the two files, obviously the clock in the camera and the clock in the recorder must be set to read exactly the same time.

Note that receiving external timecode only sets the slave device's timecode clock counter to read the same as that of the master, it DOES NOT necessarily synchronize the two devices' sample clock rates. The SD 7xx series of recorders, for example, do NOT derive sample clock from external timecode sent to them - instead you have to send them wordclock to sync the sample rate to an external source. Think of two clocks on the wall - you can set them togther to read the same any time you choose to do so, but if one of them runs faster or slower than the other they wont stay together once you've set them. As the ancient proverb says, a man with one watch always knows the exact time but a man with two watches is never really sure <g>.

So the upshot of it is, slaving the sample clocks in the two devices makes them run at the same rate while jamming the timecode makes them read the same time at that specific instant. Matching timecode is one way to align them to each other at a single point in time. Slaving the rates makes sure they don't drift out of alignment as you move away from the single line-up point.

[Ty Ford]

Steve,

There's a place for you in Heaven.

Other thoughts.

Double recording to a 702T would give you better quality audio (if that's a factor), but you'd STILL need eyes and fingers on to capture the sound. Audio recording is NOT set and forget.

Does the PD150 have a LANC port? There's a box out there that strips SMPTE from the LANC signal.

"For recording indoors in a stationary environment (recording a concert, or several people in a room) I am probably going to get a pair of Oktava MC-012s and get the Hypercariiod capsules for them."

For reasons so great in number that only someone like Steve House has the patience to answer (especially on a national holiday), this micing approach will not yield the best results for either situation.

Regards,

TyFord

[Seth Bloombaum]

Thanks Steve, for that comprehensive summary!

Couple comments:
Sample clocks in Prosumer and Professional gear really have gotten so much better than they used to be. Timecode started way back, and analog video and audio magnetic tape players had many difficulties in synchronizing. So, when we spoke of "timecode lock", we were talking about servo-locking the motors of physical players for recording or playback/recording (editing). These were some of the first uses of digital technology in video - the actual clocks & synchronizers.

The world has changed. Nonlinear editors enable all kinds of sync tricks on the timeline that used to be impossible. It has become much more possible to fix sync, because: cameras can record reference audio tracks; 2nd system sound can be stretched or shrunk to match; and those sample clocks really have gotten much better, meaning stretch & shrink is rarely neeeded.

Having said all that, it's one thing to sync up a several hour concert recording with multiple cameras and separate audio recording, and quite another to sync several hundred documentary clips, no matter whether they were sample-rate-locked or not!

I've relatively easily synced many recordings, some with wild (unlocked) timecode, some with no timecode at all. YMMV, but, you should try this out yourself with your style of shooting and editing before deciding about double-system sound. (note: Vegas is a particularly good NLE for sync work)

Which means that there is no one right answer for all shoots. Double-system sound is not universally better than sound on the camcorder because time and money are always considerations. So are crew size & capability, so is the experience and speed of the editor.

Originally Posted by Steve House

Timecode is based on what is essentially a time of day clock and there are a lot of variations on how it is recorded and used, with differences between various film and video workflows and differences with the various methods of recording audio (analog, DAT, file-based) as well. While in the device itself, it may ultimately be derived from the same timebase oscillator as the sample clock, it is a separate counter altogether, meaning that it is essentially a second clock. It timestamps the video file with the exact time each specific frame of video was recorded and in the audio file the time a specifc instant of sound was recorded. But for those recorded timestamps to be useful for aligning the two files, obviously the clock in the camera and the clock in the recorder must be set to read exactly the same time.

Note also that for audio recording, time code stripes that referenced every "frame" of audio went out with timecode dat acquisition. Now, with very digital flash or HD recorders using formats like broadcast wave, all we get is an initial timestamp when recording starts. Which works, because of those better sample clocks.

[Carlos E. Martinez]

Originally Posted by Ty Ford

Does the PD150 have a LANC port? There's a box out there that strips SMPTE from the LANC signal.

As long as you can find it...

This is a box I might be interested in too, but it's a mystery where to buy it. Does anyone know of an actual place to get one?

[Steve House]

Originally Posted by Seth Bloombaum

...

Note also that for audio recording, time code stripes that referenced every "frame" of audio went out with timecode dat acquisition. Now, with very digital flash or HD recorders using formats like broadcast wave, all we get is an initial timestamp when recording starts. Which works, because of those better sample clocks.

Very true, but when you playback a recording on devices like the SD 7xx series they take the timestamp in the file header and regenerate the timecode for output just as if it had been recorded in a parallel track to begin with. The important point is that in DV workflows, timecode in and of itself doesn't prevent drift but merely provides an identiable common point of reference in the audio and video streams to allow them to be aligned to each other in post.