Simplest Lip-sync equipment?

[Greg Miller]

I'm posed with a situation that is pretty new to me. Hopefully someone here has already pondered this and has found a good answer.

The situation involves a pianist playing a 4-hand duet with herself. We need to record the first piano part in stereo, which is simple. Then we need to playback that recording (stereo would be nice, mono would be acceptable) for use as a sync track, and simultaneously record her playing the second piano part in stereo.

Although it will complicate things, I think I will be forced to do the playback with a speaker, rather than headphones. Thus, I will probably try to subtract out some of the bleed from the playback, from the recording of the second part. (I will do my best to minimize this problem with judicious mic placement, but it won't be perfect.) Because of this probable need to null out, I would like to have the first two tracks *perfectly* in sync with the second two tracks (i.e. zero-sample drift).

I realize there will be some issues with nulling: the time delay from speakers back to the mics, and also the fact that the speaker response is not at all flat. Be that as it may, I would like to have zero-sample drift between the two pairs of tracks.

I would like to find one piece of hardware that can play two tracks, and simultaneously record two tracks, using the same clock for both functions. Does anyone know of such a device? Suggestions?

Other approaches? (I could use a separate playback machine, and then re-record that audio on two tracks of a four-track recording, with the second-part mic on the other two tracks. That might be easier and/or cheaper, might be harder and/or more expensive.)

Since the forum has been rather sparse for the last few days, I thought I'd liven things up with this off-the-wall scenario. Thanks in advance for any suggestions!

[Jay Massengill]

Why will you have to use a loudspeaker for playback? What about using a single very high quality earbud for the playback of the first part? That could be hidden pretty easily and still leave her one ear open to listen to her own second part.

Given the volume of playback necessary for her to hear the first part over her playing the second part, I doubt you will be able to absolutely cleanly record the second part if you use a loudspeaker.

There are plenty of 4-track recorders that will playback 2 and record the second 2 tracks using the same internal clock. I think the human element (player), and the inevitable bleed from loudspeaker to mics will be the more difficult to solve problems.

What if you record both the first part AND the second part as audio-only first? Then record the video (are you recording video?) using the combined parts as both playback sync for the video performance and as the actual audio in the final product. The scratch audio recorded during the video will be discarded, like in a normal music video shoot.

[Chris Medico]

One of these - M-AUDIO - ProFire 610 - High-Definition 6-in/10-out FireWire Audio Interface with Octane Preamp Technology

and a copy of ProTools will do it for sure.

+1 on using a very hide-able ear bud for monitor.

[Greg Miller]

Jay:

I think I will have to use a loudspeaker for playback because the pianist isn't willing to use an earbud. Her job is to play the piano. My job is to solve the audio puzzle.

If this project is going to work at all, the technician will need to accomodate the musician. I can't change that. Nor do I think it would be right to try to: if a musician is forced into an uncomfortable situation, the performance likely will suffer, and this musician is used to hearing another piano (or another pianist sharing the same piano), with both ears, in order to adjust dynamics and subtle nuances of the performance.)

Yes, I'm sure I won't be able to get a 100% clean recording of the second part, given some inevitable bleed from the playback speaker. That's what prompts my question, and that's stated in the question.

Mind you, the bleed won't be terrible. I will mic the piano as closely as reasonable, without getting an "up-close rock piano" sound. And I'll keep the playback speaker reasonably close to the pianist, so the level [at the mics] will be fairly low. Glenn Gould's humming isn't overpowering, it's just sometimes audible in the background. I am hoping I can achieve a similarly low level of bleed. In fact, if I'm lucky, I won't need to tinker to remove it at all. (I think most of the pieces are "one piano four hands" rather than "two pianos four hands"; so ultimately the two pairs of tracks will be mixed together anyway... I just want to be prepared to avoid any obvious coloration from the playback loudspeaker.)

Be that as it may, I would still like to use a single machine for simultaneous playback and recording. So, as I originally asked: Can you recommend any specific recorders that would be appropriate for this usage? I guess I could read a lot of catalogs and spec sheets, but I thought it would be more efficient to get some recommendations from someone on this forum with relevant experience.

Unfortunately, your last question is moot, because in reality we will have to record the second piano part using loudspeaker playback of the first piano part. So the question of when we shoot video doesn't help with the audio question I originally posed.

[Greg Miller]

Chris:

Thanks for the M-Audio lead. I will read through the literature and see what it's all about. I certainly can't complain about the price.

OK, an earbud would be nice, but in this case it really won't fly. And I have to sympathize with the performer. Back in my college days, I did some glee club singing. It was a matter of hearing the other people around me, and adjusting my voice to blend in correctly. I don't think I could have done that with any kind of headphone or earbud that prevented me from hearing the other voices and my own voice all mixed together acoustically.

I suppose if I had been singing a solo track, against a pre-recorded chorus, I could have just done my best, and let the correct mix evolve in post. But it would have required a new learning curve on my part, before I could have done that comfortably.

With the current recording project, such a learning curve just isn't an option. I fully understand that classical music is a special animal. It's going to be enough of a learning curve for this pianist (who's used to sharing the bench with another pianist, or facing another pianist at a second piano) just to listen to loudspeaker playback. Blocking even one ear with an earbud would be too drastic a change. The pianist doesn't want to do it, and IMHO that's where it ends. I'm sure there is another [technical] option, and I feel I'd be way out of line to tell this pianist how to play her music on her instrument.

[Bill Davis]

There are a bunch of amateur ways - and one professional way.

Which you use depends on your budget.

All the amateur ways require workarounds to balance a recording using live mics in a sound field where another recording at nearly the level of the original piano is taking place. As you note, the issue is that you can't isolate the results, everything gets mixed up.

Which is precisely why the pros seldom do live on stage recordings this way. The secret is that they virtually NEVER use actual acoustic pianos. It's really common for touring bands to roll a baby grand on stage - but that's just a shell into which they've installed a top quality sample driven electronic keyboard.

The digital signal from the performance can therefore be split out electonrically with no danger of feedback or other issues. You can send it to the PA at whatever level is needed, and a secondary feed to the stage monitors, and a third, perfectly clean version can be split off for a recording - and any or all of these can be as loud as you like since there are no microphones picking up the sound to worry about.

That's the professional problem solver here. But it requires enough budget for a quality keyboard and a stage monitor mixer who can feed the correct levels to the performer so that they can play to the prior recording as IF it's live, knowing all the time that their new performance is being recorded cleanly without any bleed at all for proper mixing in post.

Source: A good friend of mine is a top level piano tech that tours with giant stage shows of the big artists everyone on the planet has heard about. His job is to service and maintain these keyboards. He travels with at least 3 fully functioning keyboards and can totally swap one out of a "dummy" grand piano in about two minutes flat if something goes bad. Show business!

FWIW.

[Paul R Johnson]

I've just started the latest project for a concert pianist, and completely agree with the way their needs (not wants) have been presented here.

I'm recording in the pianists home because he will not play an alternative piano - the music was composed on this instrument, with two ears and we've tried these options before. One eared monitoring doesn't work. he simply cannot play accurately with one ear blocked. In the music room is also a high quality stage piano he uses for live work. We were experimenting with recording one piece that I need to add some sound effects to - and we spent a long time just getting him to the end on the plastic piano.

Speakers will be a major problem. A grand piano (which I assume is what we're talking about - although the problem is similar for an upright) is not a point source, so attempts to 'phase-out' the monitor sound don't work that well - however, one old stage trick that does help let you null out the monitors is to use two - but wire one with the polarity reversed. if these are equidistant from the mic (singular) the nulling out works better. If you are recording in stereo - less effective of course. Small compact speakers can be stuck on microphone stands and placed facing towards the ears, but with their backs to the mics - that also helps. The addition of video to a multi-track recording really complicates things - most of the solutions look visually naff! Getting the recording right, if you can do it, then trying very hard to convince the pianist to just play along for the second part, without the speakers visible may be best for you - but they'll probably hate it. At the end of the day, it may be best to do a rough recording and let them decide on the success of the recordings - with a reshoot to solve the problem once they've understood what went wrong. My technique now is very different from the original method, because he composed pieces so hard he could only just play what he'd composed! What we now do is record the 2-3 minute compositions in short blocks - based on ending at the end of a musical phrase, then recording the next section and then editing them together. A recording session can generate hundreds of takes. He can detect minute tenp changes between them, but I compile the edits from the best ones - then he sits with me and we make minute changes, and perform crossfades to match them up. On the video versions - these edits are ideal places to cut anyway. The biggest hassle and potential cockup is writing accurate notes. I get the pianist to announce each attempt. "Allegro Vitore in Gm - second section take 6, last one ok but I can do better". This really helps.

Based on doing this a lot, with a simply excellent, but demanding pianist on a superb instrument - record the audio and video for the first part - then record the second part with very close to the ear, and quieter small monitors wired with one reverse polarity. Lastly record the video for part 2 as a play-along.

Expect resistance to the play-along, but visually the isolation technique looks dreadful - BUT - to get the separation you, it's essential. Spill will ruin any attempt to use more distant monitors, and nulling it out will really spoil the sound

[Greg Miller]

Bill:

I appreciate your reply but based on the actual situation, which I have tried to explain in detail, a lot of your comments do not apply to my scenario. The key to the disconnect is in your saying that "... the pros seldom do live on stage recordings this way. The secret is that they virtually NEVER use actual acoustic pianos. It's really common for touring bands..."

If you think that your description is the only description that applies to "pros" then I think you have a rather uninformed view of classical music. Do you mean that Van Cliburn wasn't a pro? That Leonard Bernstein wasn't a pro? That Leon Fleischer and Martha Argerich aren't pros?

If you read my posts, I am talking about a classically-trained pianist, playing classical music for two pianists (4-hands) playing one or two pianos. Period. There are no touring bands. There are no other instruments at all. It's not a live performance with audience; it's a recording session. There is nobody mixing for stage monitors, because there are no stage monitors at all, and no reinforcement, either. These are classical piano duets, and I can't imagine classical music ever being performed on anything other than real pianos. That's what I'm asking about here... not amplified synthetic sound and "touring bands." It's not "show business," it's classical music.

Thanks for your insight into an entirely different kind of performance. It may well be useful for somebody, some day, and it makes for interesting reading. But, as I said, it doesn't apply to my question.

--

Paul,

Thank you for chiming in! I knew eventually I'd hear from someone who would understand and sympathize with my dilemma.

Your point about bleed is well taken. I will be recording in a real-world acoustical environment, not an anechoic chamber. So there will be multiple reflections from the speaker(s) into two mics. It won't be possible to completely null it out, nor will I try to do so.

I was hoping to get away with one playback speaker, and start the playback track with some digitally-generated clicks. Then I could at least time-align the recording of those clicks on the two mics, and partially null out the recording of the direct sound... and let the reflections fall where they may. (As I said, in a typical performance, both piano parts will be mixed anyway, so a small amount of bleed won't be a game killer.)

I definitely like your idea of a small playback speaker on a stand, near the pianist. That way, I can have a playback at a realistic level at the pianist's ears yet have a much lower soundfield at the mic location.

I'll have to ponder the reverse-phase playback trick. It might work if I were recording in mono, but I anticipate recording this with a pair of mics. Perhaps if I record X-Y with the mics essentially in the same location... hmmm. This might be interesting to try, although it wouldn't be my first choice for micing a piano.

Your project certainly sounds interesting, and a lot more demanding than what I anticipate doing. I pray that I won't have anywhere near that amount of editing to do. And since we'll have the playback track for reference, matching tempo should be less problematic. That only leaves matching level and matching expression in general. Sigh...

I wish you luck with your project, and thanks again for your suggestions!

[Steve House]

I don't understand why the options are limited to a mono earbud or speakers for the playback. What you're describing is standard practice for tracking and the way it's usually done is for the artist to listen to the playback of the previously recorded tracks through headphones while recording the new ones. For recording all you need is your DAW such as Adobe Audition and an audio interface that is full-duplex capable (most decent interfaces are). Mic the piano as for a normal stereo recording and record your first stereo track. Play the track back through headphones for the artist to record the second part to. Mix enough of the live signal into the playback track for her to hear herself properly. The result will be two stereo tracks completely in sync with each other on the workstation's timeline. You can even record an entire band this way ... first laying down a click track, then the drum tracks, then the bass line, then the various instruments and vocals. Multitracking like this is a very common technique in the pop recording world. Five or six years ago I happened upon an instructional video of Roger McQuinn (The Byrds) demonstrating the technique as he recorded what became a Grammy winning album in his living room with nothing more than decent mics and headphones and a Dell laptop running Audition. You might Google it and see if it's still available.

[Garrett Low]

Gregg, are you trying to record audio only or audio and video? If it's audio only, then Steve's description of how to do it is how the "pros" do it. A simple M-Audio Fastrack or similar device will do it. You could even go really cheap and get an Alesis Multimix USB for about $150.

If you are want to also capture it on video so you don't want to see the headphones, you're a lot more limited. If the artist won't live with headphones or an ear bud, they your workflow is the same except you send the first track to some monitors instead of the headphone. The problem as you've noted is isolating the second piano performance. It can be done so that it is acceptable but it won' be as clean.

[Bill Davis]

Well, then we disagree. The parts of my response that I think DO apply have to do with the nature of sound recoding and reproduction.

BTW, I brought up the "touring band" stuff to highlight the process of ISOLATION of instruments to achieve quality - not to imply that a touring rock band and a classical performance are similar. The point is that the techniques that work on stage for the former - may be the key to solving your frustrations with the latter - using the kind of modern tools that Brenstein and Cliburn simply did not have access to in their performance eras. As to Mr. Fleisher (now 84) and Ms.Argerich now what? 74? I doubt they every considered recording other than the way they've been doing it for their entire careers - which may or may not reflect the current state of the recording arts.Tradition is a powerful thing. But not always the best solution when new techniques are at hand.

Anyway...

The task you've set for yourself is demanding because, at least in my view, of the unyeilding (and eternal!) physics of sound propagation.

A live two piano duet creates incredibly complex harmonic richness. It's the interplay of vibration with the acoustic space.

Now suddenly, you want to record one of those pianos - and then record a second piano and try to re-create the merging harmonic profile of the second with the first. But if you think about it, every technique (especially those that rely on phase cancellation playback mixed with the second performance) simply can not re-create anything close to the complex interplay of harmonics that a two piano simultaneous recording will create.

So I would tend to argue that they are flawed approaches from square one.

The only way I can think of that you could accomplish anything better would be to install a theoretical "perfect" playback chain that actually reproduced the recording of the original piano with such pristine fidelity that the reproduction would blend with the new performance as if the two instruments were actually interacting in the space - and that's kinda impossible with modern playback gear, as far as I understand current transducer limitations.

Now, I don't do this type of classical recording for a living. So if you guys do and tell me that the system you're using is what gets the best resuts, I'll bow to your superior experience.

But I've been in and around sound and recording for a lot of years, and I simply can't imagine that any system that mixes a purposely phase cancelled playback of Piano 1 - with an "open mic" pristine recording of piano performance 2 is going to be nearer to the actual sound of duet than mixing two "isolated" recordings of the two performances in post.

If the pianist is unwilling to isolate performance 2 from performance 1 using the proven techniques of headphone isolated multi-tracking - then the next best thing I believe would be to isolate each recording and you simply can't do that if there's an open mic present while doing first performance playback.

That's how I see it. If you can articulate why I'm wrong, I'd be interested - and would be most willing to change my thinking if the explanation is at all compelling.

Sound is like light. You turn on a red bulb, even in a room with generally white light, then the red will be noticably mixed into the scene.

To me, any playback of the initial recording - audibly - to allow the pianist to synchronize their playing of the second part, inexorably damages the recording of the second part.

That's how I see it, anyway. Educate me if I'm wrong. I'm very willing to listen.

(BTW, if you have two open mics, I'm not sure I'd count on using phase cancellation as a way to suppress the first recording. Remember the temporal delay between two open mics means that any "mixed" recording will be hopelessly mis-aligned as to phase since room reflecting signals will be hitting the two diaphragms at different times and likely causing a good bit of unavoidable comb filtering. FWIW.)

Originally Posted by Greg Miller

Bill:

I appreciate your reply but based on the actual situation, which I have tried to explain in detail, a lot of your comments do not apply to my scenario. The key to the disconnect is in your saying that "... the pros seldom do live on stage recordings this way. The secret is that they virtually NEVER use actual acoustic pianos. It's really common for touring bands..."

If you think that your description is the only description that applies to "pros" then I think you have a rather uninformed view of classical music. Do you mean that Van Cliburn wasn't a pro? That Leonard Bernstein wasn't a pro? That Leon Fleischer and Martha Argerich aren't pros?

If you read my posts, I am talking about a classically-trained pianist, playing classical music for two pianists (4-hands) playing one or two pianos. Period. There are no touring bands. There are no other instruments at all. It's not a live performance with audience; it's a recording session. There is nobody mixing for stage monitors, because there are no stage monitors at all, and no reinforcement, either. These are classical piano duets, and I can't imagine classical music ever being performed on anything other than real pianos. That's what I'm asking about here... not amplified synthetic sound and "touring bands." It's not "show business," it's classical music.

Thanks for your insight into an entirely different kind of performance. It may well be useful for somebody, some day, and it makes for interesting reading. But, as I said, it doesn't apply to my question.

--

Paul,

Thank you for chiming in! I knew eventually I'd hear from someone who would understand and sympathize with my dilemma.

Your point about bleed is well taken. I will be recording in a real-world acoustical environment, not an anechoic chamber. So there will be multiple reflections from the speaker(s) into two mics. It won't be possible to completely null it out, nor will I try to do so.

I was hoping to get away with one playback speaker, and start the playback track with some digitally-generated clicks. Then I could at least time-align the recording of those clicks on the two mics, and partially null out the recording of the direct sound... and let the reflections fall where they may. (As I said, in a typical performance, both piano parts will be mixed anyway, so a small amount of bleed won't be a game killer.)

I definitely like your idea of a small playback speaker on a stand, near the pianist. That way, I can have a playback at a realistic level at the pianist's ears yet have a much lower soundfield at the mic location.

I'll have to ponder the reverse-phase playback trick. It might work if I were recording in mono, but I anticipate recording this with a pair of mics. Perhaps if I record X-Y with the mics essentially in the same location... hmmm. This might be interesting to try, although it wouldn't be my first choice for micing a piano.

Your project certainly sounds interesting, and a lot more demanding than what I anticipate doing. I pray that I won't have anywhere near that amount of editing to do. And since we'll have the playback track for reference, matching tempo should be less problematic. That only leaves matching level and matching expression in general. Sigh...

I wish you luck with your project, and thanks again for your suggestions!

[Richard Crowley]

In a quick review of this thread, I didn't see any explicit statement that this was a live performance with an audience vs. a recording session?

I would be strongly tempted to consider using some small monitor speakers in perhaps unconventional positions. One possibility would be to place a pair of small (5-inch) speakers on either side of the music stand. With the speakers relatively close to the performer's ears and aimed AWAY from the business-end of the instrument, you can reduce the playback bleed into the recording microphones.

Or, alternately, a pair of small monitors on mic stands just behind the performer's head. Again to get the speakers as close to the ears as possible to reduce the amount of acoustic bleed into the recording microphones.

I would also remind the producer/client that limitations imposed on the recording situation will result in unavoidable compromises in the quality of the recording. In the immortal words of Scotty, engineer of the star-ship Enterprise: "Ye cannae change the laws of physics!"

[Greg Miller]

It's not so much a matter of whether we disagree. It's just that we're talking about two different things, or talking about the same thing from two different perspectives.

You're talking about

Originally Posted by Bill Davis

the current state of the recording arts.

and what is technically possible.

I'm talking about what this particular performer is willing to do. The only playback option this performer is willing to consider is loudspeaker playback. Therefore, some of the techniques you suggest, while they are valid as techniques, are not applicable in this case. They're current, they're interesting, and I don't question that they work technically when they're used, but they don't answer my question: my question relates to what options I have with this situation which is defined by what this performer is willing to do.

I don't disagree with everyone's comments about how difficult it will be to achieve any useful degree of cancellation. That is a big bugaboo. I understand why the problem exists, and I understand that it will be very imperfect.

Originally Posted by Bill Davis

if you have two open mics, I'm not sure I'd count on using phase cancellation as a way to suppress the first recording. Remember the temporal delay between two open mics means that any "mixed" recording will be hopelessly mis-aligned as to phase since room reflecting signals will be hitting the two diaphragms at different times and likely causing a good bit of unavoidable comb filtering.

Yes, I understand that and I agree completely. The more I think about this, I'm concluding that maybe I should be thinking not about cancellation, but rather about addition. Perhaps I should strive for the best possible playback (within reason), keep the level as low as possible (within reason), and then consider the bleed from that to be a small part of the sound from the "first piano." Then I can add in the direct track from the first piano, to come up with the final mix. It won't be perfectly accurate reproduction, but it might just sound like a slightly different piano, and be entirely believable and acceptable.

I still need to think long and hard about time relationships between the playback speaker and the "second piano" mics, but perhaps there's a solution lurking in here somewhere.

Given all that, I'm going to take a long look at the M-Audio gear that Chris Medico recommended, it might make this all possible. And I'll clearly need to perform a few tests to see if the above scenario produces acceptable results.

Thanks again for all the suggestions.

[Greg Miller]

Originally Posted by Richard Crowley

In a quick review of this thread, I didn't see any explicit statement that this was a live performance with an audience vs. a recording session?

Richard:

It's buried in the third paragraph of post #8. It's a recording session with no audience. That gives me some extra leeway.

I'll need to try some tests with different speaker locations, before the actual session. Your suggestions sound pretty valid to me. Luckily I can play a few notes on the piano, so the test will have some validity as far as checking levels and bleed, etc.

Originally Posted by Richard Crowley

I would also remind the producer/client that limitations imposed on the recording situation will result in unavoidable compromises in the quality of the recording.

Yes, they're willing to trust me to "do my best," whatever that may be... I just don't want to use this particular scenario as an excuse to do less than the best I can possibly do.

[Garrett Low]

Gregg, what is the reason the artist does not want to use some form of isolated listening? If it is that they won't be able to hear themselves playing, I think that is has already been addressed that they will be able to hear themselves play along with the first recorded track.

From working with actors and musicians I know that everyone of them has their own particular desires and way of performing. Just curious about this one.