MixPre-D — AES out converted to S/PDIF for input into Tascam DR-100 MKII?

[jimbee]

I'm wondering if anyone has tried using the MixPre-D's AES/EBU output to send a digital signal directly to a low-budget recorder, using the recorder as only a "bit bucket," and bypassing its preamps and A/D converter?

Let's say you make a custom cable that goes from female XLR to 3.5mm TRS (matching the TRS converter cable that comes with the Tascam DR-100 MKII; this digital input cable has the tip, plus three rings and three sleeves). And this custom-made cable also converts the signal from 110 ohms to 75 ohms, which I believe is what it needs to turn the AES signal into the S/PDIF signal that the Tascam is expecting to see.

So you're using the preamps and the A/D converter on the MixPre-D and sending a pure digital signal to be recorded by the Tascam. Theoretically, wouldn't that give you a great, low-noise recording, much better than you could get with the Tascam by itself?

Problem is, I have not found anybody who has A/B'd this combo in the real world with their own ears.

I did find a guy who has gone digital out with a cable like this from the MixPre-D to an Edirol R44, and much to my surprise, he said, "I bought the custom cable, so I had the ability to use the Edirol as a bit bucket. That is what I use it for, ocassionally. Mics go into the MixPre-D, which sends it digital to the Edirol. This way I only have to set the levels properly on the MixPre. The Edirol does not add anything to the sound."

"And to be honest, I have not really found any difference compared to recording straight into the Edirol. But it's just nice, I can record from the MixPre-D to a Sony M-10 (analog) and R44 (digital) at the same time (backup)."

Now maybe the Edirol R44 is a really good recorder (price, around $1,100), but I am a big fan of Sound Devices after recording with the 722 for a few years. It seems like if you could capture the pure digital out from the MixPre-D, and then compare it to a recording done with the same mic directly into the Tascam's preamp, using the Tascam's A/D converter, the MixPre-D recording would be much better. Anyway, that's my theory.

The question I'm asking myself is: Would it be $800 better, basically the cost of the MixPre-D plus the custom cable?

Recommendations and advice much appreciated,

jimbee

 

[davefisk]

$800 better than just going straight into the recorder? I'm not sure. I haven't done that. I do remember the last time I tried to convert AES to SPDIF using an adapter cable I had issues, so I don't do that, but that was some time ago. Maybe something has changed to make it more reliable. If you have the 722, do you have a need for the Tascam?

 

[jimbee]

Sadly, Dave, I no longer own the 722, but the memories remain. Now I'm thinking, the USBPre 2 would send S/PDIF to the Tascam, no AES-to-S/PDIF conversion needed. So, Sound Devices preamps and A/D conversion dumped into the DR-100 MKII … theoretically I think it would be great, but there's always the pesky real-world result. Maybe somebody has heard the difference between this setup and the Tascam by itself. If I was recording loud music, the Tascam would be fine, but for quiet film dialogue, when you can't get the mic in close enough — the DP's fault, not the K-Tek boompole's — a lot of gain with low noise is the holy grail. If only grail wasn't so expensive.

 

[TheDingo]

but for quiet film dialogue, when you can't get the mic in close enough — the DP's fault, not the K-Tek boompole's — a lot of gain with low noise is the holy grail

The problem is that the ambient noise at most locations is too loud when trying to boost low audio levels to fix the problem. From my experience, ambient room noise is rarely below -55 dB, unless you are filming on a dedicated sound stage or somewhere very isolated from the modern world, and so this puts a huge limitation on how much your can boost your audio recording to fix this type of problem.

The Tascam DR-100 MK2 in analog mode using the LOW or MED preamp setting has a noise floor between -84 dB to -79 dB depending on the adjustable GAIN setting. If your mics are low noise ( and your location has very low ambient noise ) then you can boost your overall audio levels by 15 to 20 dB when recording in analog mode and still obtain a very quiet sounding recording with a -60 dB noise floor. Digital recording is not necessary at all.

 

[jimbee]

Yes, an excellent point, Dingo, one that I forget, one that crushes everything else. The world is too noisy to make clean outdoor recordings, unless it's the S/N ratio at a Metallica concert … jet planes, you have been beat down.

As usual, Hollywood has warped my expectations and desires. Everytime I listen to a "real" movie, I'm thinking, How did they record the dialogue at that outdoor, streetside cafe in New York? Even a simple moving car dialogue scene seems to be a feat of magic, given the engine and road noise in reality (at least in my car). OK, the cafe dialogue is probably ADR'd and a quieter street tone is put underneath, and the car maybe is being towed on a trailer and there is no engine noise. Nevertheless, I get sucked into thinking, Why can't I do that?

Maybe someday I will get to hear some actual location sound recorded by a big-time mixer with the best gear and best mics, and then I'll think, How is this even usable? I would love to know how much bad sound is passed on to the sound editors, and then how do they do their magic? Can they make bad sound usable, or even good? Is their software a lot more powerful (and expensive) than, say, iZotope RX2?

And that raises the question, How can no-budget filmmakers get sound that is not embarrassingly bad? Hmm, don't film a scene next to a freeway, which, everytime I see a director in my town do that, I laugh (as I am crying quietly on the inside).

I agree, the DR-100 MKII on low or medium preamp setting is very quiet. Unfortunately, at medium, even at full gain of +10 on the wheel, that does not give you a loud enough signal for normal talking. Maybe the talking would peak around -35 with a shotgun mic about four feet away. And a wide shot, where you can't get the mic in close? No chance, unless you've got $5,000 for a couple of Lectros and COS-11s to go with them. And then you've got clothing rustle ruining the take. I've found that the Tascam, on high preamp gain, with the control wheel at +9, will get you a decent, usable dialogue level, but of course it's noisier. The noise floor might come in around -42 in a quiet room, not terrible, but the audible hiss is objectionable to my ear.

My experience has been that in a fairly quiet room, 55db or 60db of gain on a Sound Devices 722 with a Sennheiser 416 on a boom about four or five feet away will get you a good dialogue recording that might peak at -12, and if the actors yell, the limiter will save you. I can no longer afford to own that gear, but the memories of it drive my quest — and my delusion. Thus, I begin to think, how about the reasonably quiet Rode NTG-2 (on battery power, since it will supposedly run for 50 hours on the AA) into a USBPre 2 feeding S/PDIF into the Tascam. Could you get the same sound for $1200 that previously cost $3000? Someday, if I am able to get a USBPre 2, I will report back.

Then again, your original point may make the whole thing irrelevant. In a noisy exterior, I guess you would get crappy sound with a Shoeps CMIT 5U into a Nagra VI. And it would sound great in the theater.

 

[Steve House]

...
I agree, the DR-100 MKII on low or medium preamp setting is very quiet. Unfortunately, at medium, even at full gain of +10 on the wheel, that does not give you a loud enough signal for normal talking. Maybe the talking would peak around -35 with a shotgun mic about four feet away. And a wide shot, where you can't get the mic in close? ....

My experience has been that in a fairly quiet room, 55db or 60db of gain on a Sound Devices 722 with a Sennheiser 416 on a boom about four or five feet away will get you a good dialogue recording that might peak at -12, ....

I have to jump in to point out, 4 to 5 feet distance with a shotgun is much too far. 'Guns like the 416 work best for dialog when they're between 18 and 24 or so inches away from the subject.

 

[jimbee]

Of course. In a perfect sound world, every shot would be a closeup or medium closeup. Those are easy and the sound is great. Four or five feet is, hopefully, worst case, even on a master shot. It never worked for me to ask the director or the DP if he could tighten the framing so I could get the mic closer.

 

[TheDingo]

Is their software a lot more powerful (and expensive) than, say, iZotope RX2?

I am curious to know if the $1,200 Advanced version of RX2 ( with adaptive noise reduction ) is worth the extra $850 over the standard version.

How can no-budget filmmakers get sound that is not embarrassingly bad?

Know your audio gear and it's limitations, which will likely mean budget for ADR when there are no other options to getting good sound.

I agree, the DR-100 MKII on low or medium preamp setting is very quiet. Unfortunately, at medium, even at full gain of +10 on the wheel, that does not give you a loud enough signal for normal talking.

I think it really depends on the mic you are using. I've had good luck with the Rode NTG-3 and Sennheiser MKH 8060, as long as the mic is 3 feet or closer to the person speaking. Otherwise, I use Sennheiser UHF packs with Tram TR50 lav mics. I own one Sanken Cos-11 lav, but I rarely use it because it picks up so much ambient sound. ( I was going to switch to all Sanken lavs until I bought one to try out )

Maybe the talking would peak around -35 with a shotgun mic about four feet away.

We both know that this is too far away from your subject, and lowers the sound quality significantly compared to having the mic closer.

55db or 60db of gain on a Sound Devices 722 with a Sennheiser 416 on a boom about four or five feet away will get you a good dialogue recording that might peak at -12

Gah! ...I've never recorded sound that I have been happy with when the shotgun mic is so far away. If I can't get the shotgun within three feet then it's lav mics all the way.

In case you are curious, I am putting together a portable recorder test for my first blog post ( I still have to do a live music recording test for the internal mics ), where I'm testing eight different recorders between $100 - $400. Here's the noise floor results for the Tascam DR-100 MK2:

Tascam DR-100 MK2

MIC GAIN SETTINGS ( LOW, MED, HIGH settings )
LOW
MED = LOW + 27 DB
HIGH = MED + 28 DB

MAX MIC GAIN RANGE : 78 dB


RMS MIC NOISE FLOOR : MIC GAIN = LOW

MIC LEVEL 2 : -83 dB
MIC LEVEL 4 : -84 dB
MIC LEVEL 6 : -84 dB
MIC LEVEL 8 : -83 dB
MIC LEVEL 10 : -84 dB

RMS MIC NOISE FLOOR : MIC GAIN = MED

MIC LEVEL 2 : -83 dB
MIC LEVEL 4 : -83 dB
MIC LEVEL 6 : -83 dB
MIC LEVEL 8 : -80 dB
MIC LEVEL 10 : -79 dB

RMS MIC NOISE FLOOR : MIC GAIN = HIGH

MIC LEVEL 2 : -79 dB
MIC LEVEL 4 : -72 dB
MIC LEVEL 6 : -63 dB
MIC LEVEL 8 : -55 dB
MIC LEVEL 10 : -53 dB


...The Tascam DR-100 MK2 starts to get quite noisy when the mic pre-amp is set to "High", so I would avoid it if you can.

 

[Steve House]

Of course. In a perfect sound world, every shot would be a closeup or medium closeup. Those are easy and the sound is great. Four or five feet is, hopefully, worst case, even on a master shot. It never worked for me to ask the director or the DP if he could tighten the framing so I could get the mic closer.

Of course you don't ask the director to tighten the framing. But even a two-shot doesn't always have a lot of daylight at the top of the frame, thus letting a boom op with a long boom get in close by just caressing the upper frame line - that's why they make 16 foot or 20 foot booms. What you do is plan your mics, going to lavs if you can't work your 'gun close enough for proper pickup.

 

[jimbee]

I will definitely check out your blog post on recorders, Dingo. That sounds really interesting and helpful. What's the URL going to be?

Your results for the DR-100 MKII seem pretty good, slightly better than what I've been able to achieve. I'm probably not measuring RMS, though. To compare various mics or, say, the difference between running an NTG-2 off the Tascam phantom or running it off AA, I'll record something in a quiet room, not anechoic, but a room where my ears hear almost nothing. Then I'll put that file into RX2 and analyze it. What I call the "noise floor" is the recording when I'm not talking or whispering, and my results seem to be about -45 to -48 on full gain with the NTG-2 on AA battery power. To my surprise, I heard a difference between running the NTG-2 off AA compared to running it off Tascam phantom — it was slightly quieter and the noise that I did hear sounded less objectionable somehow, not to mention that when you run phantom on the Tascam, it drains the battery shockingly fast.

Yeah, I would love to know whether RX2 Advanced is worth the extra $850. That sounds like a lot, but if it performed magic, that would be cheap. Of course, that's nothing for a Hollywood post house. Sometimes, depending on the noise, I'll find that the RX2 (cheap version) denoiser module comes as close to performing magic as I can imagine. The Spectral Repair module, too. It's case by case, and of course there is no substitute for getting good sound to start with, but I do love RX2.

I find it is easy to reduce my Tascam "quiet room noise-floor" results from -48 to -66 with RX2, and at that point, any objection I have to the noise goes away — I can't really hear anything at -66. It's not scientific, and I am monitoring through headphones, so that's no good, but that's the gear I have. Surprisingly, if I have a good section of "noise" to profile, and in my home testing, I do, the reduction does not hurt the voice at all. It's like I was talking and underneath I would hear a little hiss, and after the RX2 pass, I'm talking exactly the same, but the hiss has disappeared. And there's no weirdness happening to the voice like you would get in the so-called denoiser in, say, Soundtrack Pro.

If you haven't used the envelope filter in the denoiser module, it is worth checking out, and that is a feature that I have never found any documentation on. But iZotope support answered by questions, and now I will often include it in my algorithm to try to target the noise better.

Steve, that was a joke when I mentioned "asking the director to reframe." That's not to say I've never done it, I have, but only with my friends. True story, I was doing the sound for a friend's short, and the DP was a photographer from New York (shooting on the 5D MKII) whom I hadn't met before. It was a constant battle to get him to tell me or show me where the top of the frame was so I could get a few inches outside it. After a lot of this, he finally, very annoyed, said, "Why should I help you? I'm only concerned about the image." Maybe that's how it goes in New York. He was, seemingly, a very important, big-time guy. I had the chance to work with him again and chose not to.

Also, I probably should have spelled out that when I mention "no-budget" filmmaking, in my little Portland, Oregon, circle, that really does mean no money. We are lucky if we can afford to buy food. One of my short films did have a budget of five dollars for a crucial prop, a can of saddle soap. We do not have a choice of microphones, we have a microphone. So, no lavs for us, no wireless, no 20-foot boom. I do have an 11-foot carbon fiber VDB, which, when I got it, my head was spinning around as I suddenly realized how much a good piece of gear can help you. Same with the Rycote InVision shockmount, which, at $70, seems like a bargain. I do aspire to have better gear someday. Until then, I love manufacturers who make good stuff for reasonable prices.

 

[Steve House]

...
Steve, that was a joke when I mentioned "asking the director to reframe." That's not to say I've never done it, I have, but only with my friends. True story, I was doing the sound for a friend's short, and the DP was a photographer from New York (shooting on the 5D MKII) whom I hadn't met before. It was a constant battle to get him to tell me or show me where the top of the frame was so I could get a few inches outside it. After a lot of this, he finally, very annoyed, said, "Why should I help you? I'm only concerned about the image." Maybe that's how it goes in New York. He was, seemingly, a very important, big-time guy. I had the chance to work with him again and chose not to.

... Same with the Rycote InVision shockmount, which, at $70, seems like a bargain....

The Invision is truly a bargain, one of the best, if not the best, shock mount available at an incredibly low cost. To get better, and I'm not sure if it's really better, you'd need to go to a Cinela OSIX at MUCH higher price.

To position the mic you don't need to elicit the director's help, work with the camera operator instead. When things are setup and close to ready to roll, with a boom long enough for the boom op to be out of frame to the side, position the mic above the talent over the lens-to-subject line and drop it down until it enters the top of frame, then back up until the camera operator tells you it just clears the frame line.

 

[TheDingo]

I will definitely check out your blog post on recorders, Dingo. That sounds really interesting and helpful. What's the URL going to be?

I'm not sure just yet. I have a few addresses I use. I had originally hoped to have it up by July, but my real work keeps getting in the way, so the project keeps getting pushed back. I will definitely post a link here once it's done.

 

[Noiz2]

Thought I would jump in with a few things.

As you guessed, no film with a budget is going to try and shoot dialog in a car that is moving on it's own power. No the scene was probably not ADRed. It was recorded in a quiet environment where a lot of effort was put in to keep any extra noise from happening.

The exterior was also a LOT quieter than you think it was. It also probably was not ADRed but maybe. Look at the extra content scene on the PIPE DREAMS DVD for a very interesting behind the scenes of a low budget film.

And then it was worked on by dialog editors.

Look at the 'death scene" in Mosquito Coast for an example. The shot on location in a canoe with lapping water. The water had to go and it was stepping on lines. It had to go because there are cuts so the lapping popped in and out in a very unnatural pattern.

ADR didn't sound right so the dialog editor went through all the other takes finding replacement words and syllables to patch over the laps. What you end up with is a smooth seamless performance that is actually pieced together from a stack of different performances. That is pretty amazing on it's own but the really amazing part is that it was all done on Mag (magnetic 35mm film) with razor blades and splicing tape.

Time and skill, that is how they do it.

Low budget films CAN get similar tracks. It takes thinking about it BEFORE you roll the camera. Thinking about it while you are rolling the camera and also thinking about it after. The after part most mini-indi filmmakers have down it's the other two that are often lacking.

The second part is you have to build your sound track like Hollywood does if you want to sound like a Hollywood film. That means that you allow ENOUGH time to do a proper sound post and you get someone who knows how to do a proper sound post.

There is a sticky on sound post that goes into detail but in general it means a lot of tracks. Only using dialog from production and then really covering the FX.

You talked about layering a quiet ambience. No you will be layering a stack of ambiences. You will be putting in seat sound when people move. You will change the road texture every once in a while. You will put in suspension sounds when the car bumps, the engine will be changed to match the driving that is supposedly going on. You will add car bys for cars that go by, going over the bridge WILL sound different than driving on pavement. You will get an ever changing touch of the world outside the car. Keys will make sounds, blinkers will make sounds, turning in your seat will make a sound.

Once you fill all of that out your clean dialog edit will seem totally "in the world" of your movie.

It will take time and if people are getting paid it will take money. Also it's near impossible to do this kind of post in sound Track Pro. OK maybe just very difficult. It's not the the tools are not there (most are) but STP is very hard to work with when you have over 12 or so tracks. You will spend endless time scrolling around trying to find stuff and wishing you could get rid of some of those panes (or pains as I used to call them) and still maintain functionality.

In fairness I know someone who did a decently complicated show in STP and I did one with around 30 tracks. It was not "impossible" but it wasn't fun. On a short a "real" post is probably going to run 30 - 60 tracks, a feature closer to 90. "Big" films these days are running closer to 200 tracks. In reality they always were but systems couldn't handle that many tracks so a lot of pre mixing went on to keep the working tracks at any one time to around 16. If you counted everyones full track count it would be closer to 500. But a lot of things still get mixed down before they hit the stage and alt tracks often are not sent to the mix unless needed.

 

[TheDingo]

Look at the 'death scene" in Mosquito Coast for an example. The shot on location in a canoe with lapping water. The water had to go and it was stepping on lines. It had to go because there are cuts so the lapping popped in and out in a very unnatural pattern.

ADR didn't sound right so the dialog editor went through all the other takes finding replacement words and syllables to patch over the laps. What you end up with is a smooth seamless performance that is actually pieced together from a stack of different performances. That is pretty amazing on it's own but the really amazing part is that it was all done on Mag (magnetic 35mm film) with razor blades and splicing tape.

Thanks for posting this. I will definitely have to take a look at Mosquito Coast again. ( I saw it once when it first opened )

Also it's near impossible to do this kind of post in sound Track Pro

Out of curiousity, what's an affordable audio tool that would be up to this task?

 

[Steve House]

... Out of curiousity, what's an affordable audio tool that would be up to this task?

"Affordable" is a matter of perspective. The big contenders are ProTools and Nuendo

 

[Noiz2]

ProTools used to (may still) have an "affordable" academic version. There are a few other contenders (in the very affordable range) but they don't have solid video support yet so...

Nuendo is the only Mac alternative that is used much for film. There are a few Windows programs that are have some market share.
But ProTools still is 99% of the post market in the US and a bit less in Europe, partly because it used to be Mac only and Macs cost a LOT in Europe.