Questions about sound in Red One

[craesmeyer]

As a sound guy, i would ask some questions to the Red people.


The 4 mic inputs are mic/level switchable?
There would be 48V Phatom?
The camera will have some kind of cooler or another component tha can produce an louder than aceptable sound (higher than an modern - not effect -35mm camera)?
The Red Hard Disk recording solution will be fan cooled?
And the Red Flash?


Just a sugestion... There will be an mic company who will be endorsed by Red? Suggested microphones?

You know about Core Sound's Tetra Mic? It seens to be an intersting companion to an doc/run & go type of production... I think it only needed to be a little shorter, or have an angled output to fit Red perfectly...

http://www.core-sound.com/TetraMic/2.php
http://www.core-sound.com/TetraMic/1.php

I didn't test it (nor see an review about it), but sit hould be an nice mic.

Thanks for watching

PS: Theres no slight chance that Red will get any digital audio input? AES/EBU yould be great... SPDIF not so, but... Red will be an evoution in picture, why not be an revolution in sound too? :cool:

[jbeale]

All good questions... (surely the Red team is not still adding new features, new circuits and new connectors at this late date?)

Just one note, it is slightly non-trival to add external digital audio input to a digital video camera, because the input has to be re-clocked. In general the external audio clock will drift relative to the camera master clock so you cannot simply transfer the bits directly. I'm sure there are chips that handle this, but if that chip is not already on the board, I assume this would be something for Red 2, not Red 1.

[Graeme Nattress]

Interesting mic. Michael Gerzon was a sound genius, and this, if based on his work like they say, should be superb. The 4 tetrahedrally placed mics would allow you to record the full 3dimensional soundfield on 4 channels, that you could decode into practically an surround sound format you want, as you have full XYZ information, just waiting to be pulled out of the data you recorded.

Graeme

[Stuart English]

All four audio inputs are mic/level switchable

All inputs support 48V phantom power for mics

The camera can suppress any active cooling (and hence sound) during a take

RED-DRIVE and RED-RAM and RED-Flash are all fan-less designs

Each audio input is sampled at 48KHz at 16 or 24 bit depth

We have considered methods for pre-digitized audio capture, but we have no news to share at this moment.

[Emmanuel Cambier]

Would 96khz be a huge problem?
Because it does sound better, the sound is noticably clearer and more define and would make the camera much more future proof.
24 bit is great allready, but 96khz is a very nice iceing on the cake, and would fit better in a demanding audio environement.

Yours "at any freq" Emmanuel

[jbeale]

I've taken good care of my ears, but I'm over 40 now and I have no hearing at all above 15 kHz. I envy you if you can detect 96 vs 48 kHz sample rates, all other things being equal! Of course they usually aren't... when you go to 96 kHz sampling there are normally other differences at various places in the audio chain- at the very least the details in the reconstruction filter- so unless it's a very controlled test, there is more than just a sample rate effect.

[GlennChan]

- How many channels of audio monitoring will there be?

- Is it safe to presume that you can control how all the channels get mixed to the monitoring outputs?
i.e. you can hear all 4 channels at once, or just input 1 in both ears, etc.

[Emmanuel Cambier]

I've taken good care of my ears, but I'm over 40 now and I have no hearing at all above 15 kHz. I envy you if you can detect 96 vs 48 kHz sample rates, all other things being equal! Of course they usually aren't... when you go to 96 kHz sampling there are normally other differences at various places in the audio chain- at the very least the details in the reconstruction filter- so unless it's a very controlled test, there is more than just a sample rate effect.

The difference is there to be heard on good monitors.
But I agree that it only get obvious when you are going to mix several tracks together, you really get a lot of clarity and definition, and things are being mixed a lot easier.

Yours Emmanuel

[Stuart English]

Because of the approach we have taken we have a fair bit of flexibility regarding what channel(s) can be monitored on the L and R outputs of the headphone and XLR outputs.

It would be interesting to collect inputs to what you guys would like, so we can compare it to what we've come up with as the first priority for product implimentation.

[Jochen Schmidt-Hambrock]

Hi Stuart,
flexibility about what channels to monitor implies that a 4 -> 2 channel monitor downmix is not possible?

[GlennChan]

1- Throwing out an idea here: It might be better to have 2 independent headphone outputs, as opposed to XLR outputs. Some mixers like the Sound Devices stuff take headphone/mini-jack as the return feed, not XLR.

If you want to record sync sound, a reasonable setup would be:
Two audio persons. Each sound operator holds a boom and a single wireless lav (or handle two wireless lavs). They also handle their own levels and mix their feeds. They can monitor right off the camera (via the return on their mixer).
This works well with the Sound Devices mixers, since they can't output more than 2 channels. They can have 2+ inputs, so you can have a boom and at least one lav. As well, if you make a custom snake/cable for a Sound Devices mixer (Trew Audio / Remote Audio does this), it will have a headphone jack (not XLR).

1b- If you are going wireless in and out of the camera (from your mixer[s]), 2 X mini-plug will still be ok?

2- Another situation would be using double system, where you might just record 2 tracks of guide audio, plus timecode. In this case 2 independent headphones feeds should still be fine.

3- Menu-wise, you'd need controls for:
*Monitoring/mixing. 4 channels of inputs, and 4 channels of monitoring (L+R on headphone out 1, L+R on headphone out 2) would mean a 4x4 matrix. A joystick-like control like on the DVX100 would be a sensible way of setting this if using a menu system.
*Line/mic on each input.
*Limiters on inputs 1 & 2. (I don't forsee needing limiters on inputs 3&4.)
*Phantom power (Probably ok to omit T power.)

I think with this many settings (24 or more), you'd need a menu system. Ideally, the menu system would be so intuitive that you can look at the camera and figure out how to setup the audio.

4- Stereo linking of the limiters can likely be omitted.


Flexibility about what channels to monitor implies that a 4 -> 2 channel monitor downmix is not possible?
My interpretation is that it implies that it IS possible.

[Graeme Nattress]

http://www.paia.com/ProdArticles/msmicwrk.htm

http://forums.dvdoctor.net/showthread.php?t=39513

Describes how the Mid - Side system works, which to me would be a great alternative to the tetrahedral technique linked to above.

Graeme

[Bruce Allen]

Graeme, I totally agreee about MS.

To be honest, for useful production sound, the Red should either have:
- really great preamps + DAC plus MS matrix monitoring for headphones
OR
- digital input (so that you can use an external mixer / preamp / DAC chain).
OR
- take $500 off the price of the camera, have 2 channels with cheap-ass preamps which we just record a reference track to and then we put that $500 towards a Sound Devices compact-flash timecode recorder with great pre-amps, MS monitoring and 192khz / 24bit recording (costs $2000-$4000 depending on how many channels you want)

...actually why don't you just cut a deal with the Sound Devices people to include their audio system on your camera?

The current system looks suspiciously like it has just low enough quality to not be useful.

All of this stuff about holographic 3D sound recording is likewise not that useful for real jobs. If you want to record stereo, MS is the way to go (especially with youtube, phones, video ipods, etc you always want a good mono-compatible signal).

[Emmanuel Cambier]

- digital input (so that you can use an external mixer / preamp / DAC chain).


This sure seems like a possibly good solution.

Yours Emmanuel

[Chris Kenny]

Because of the approach we have taken we have a fair bit of flexibility regarding what channel(s) can be monitored on the L and R outputs of the headphone and XLR outputs.

It would be interesting to collect inputs to what you guys would like, so we can compare it to what we've come up with as the first priority for product implimentation.

This is admittedly a fairly minor quibble, but... does the camera always record four channels, even when there's no input? It would be nice if channels could be disabled individually, to save a bit of storage space and make things just a little neater in post.

[Martin Drew]

I think well engineered 24bit 48khz is fine, 96khz would be nice but 48khz is good enough given that onboard sound will always be a bit of a compromise. If you are in pursuit of ultimate quality you will probably be using separate audio anyway. I think it is more important that the solution is implemented well and the AD stage is well engineered than pushing for maximum spec, that may lead to compromised performance.

There is an interesting paper here about 192kHz and why it maybe isn't such a great idea.
http://www.lavryengineering.com/documents/Sampling_Theory.pdf
(http://www.lavryengineering.com/documents/Sampling_Theory.pdf)

Martin

[REDHKSC]

Stuart,
may I say that RED one has the options to go for Surround Sound live recording from all the inputs ?

No Camera on earth can record surround sound from the Camera but THE RED ONE ?

Stewart

[Stuart English]

The RED_ONE camera has four discreet channels of audio at 48KHz 16 or 24 bits. What you do with that - 2 channels spilt high / low gain or surround sound matrixed recording is for the user to descide.

These user choices do have some impact on monitoring to headphones etc, which is why we'd like to gather likely use data for our consideration.

[Stuart English]

We can turn off channels 3 / 4 ( i.e not record those) if its required.

To answer the follow up question - does that mean we can use that "unused bandwidth" to record Channel 1 & 2 at 96Khz ?

That issue is not related to the above. The audio spec is 48KHz at 16 or 24 bits per channel.

(Usual rider - that all specifications subject to change)

[Jochen Schmidt-Hambrock]

Ok, if you want to monitor the surround mix somehow, the camera would have to do a downmix from 4 channels into the 2 channels of the headphones.

So a simple matrix (which input of the 4 provided do I want to hear on my left output channel, which one on the right) would not enough. You´d have to be able to adjust volume and pan for the 4 inputs.

That would indeed be nice also in a non-surround setup with maybe 2 Lav´s plus a stereo ambience mike

Anything more for the surround field would require the ability to encode the 4 inputs into some sort of Dolby-ish matrix for monitoring - bit of an overkill IMHO.

I´d be perfectly happy with the simple I/O matrix for monitoring.
Onscreen metering for all 4 channels though is necessary (to check for dead mikes, cables, etc.)

Jochen

[GlennChan]

If you need to record surround ambiances, I think a better idea would be to record them separately. A laptop with a decent audio interface is very cheap nowadays. In addition, you can take a mobile recorder to many different locations and record different ambiances.

2- For the few people who need to record a stereo mic (i.e. perhaps in documentaries), I think a reasonable workaround would be to use a field mixer than supports it (i.e. M/S headphone matrix, linked gain, linked limiters). You'll likely have a sound guy with a field mixer anyways.

[Trevor Meier]

Things I'd like to see on the input side:
- analog input limiters
- flexible options for digital input (USB, AES, ??)
- remote control of input level & remote level meters *

Options for headphone/output monitoring:
- flexible 4>2 channel downmix (all channels to stereo, all channels to mono, or separate channels/channel pairs to mono left, mono right)
- channel monitor: selectable stereo ch 1&2, 3&4 stereo mix; or mono ch. 1-4
- left ear 4>1 channel downmix; right ear, selectable channel or pair of channels (e.g. for using a channel as audio input for set radios)


* Scenario: In keeping with the indie nature of the camera, allowing a boom op to have remote control of audio levels on the camera reduces the need for a recordist or a field mixer.

[Peter Richardson]

I'm curious about a comment Stewart made:

The camera can suppress any active cooling (and hence sound) during a take.

How would the supression of active cooling work on a documentary, where the camera is almost always constantly rolling? Will there be overheating issues if the active cooling is left off? If you could elaborate on this that would be great as I principally plan to use my Red for doc work (made a reservation on Monday). Thanks,

Peter

[Steve Freebairn]

Can you record audio while you are over and undercranking? That is what I'd really like. (As long as it stays sync)

[David Fairbanks]

To keep it in sync you would have to sample at greater rates or it would have to interpolated to the greater rates. If you're shooting at 48fps that would mean recording the audio at 96KHz and so forth.

[GlennChan]

Another possibility:
What about automated syncing with .BWF files? This would enable 4+ channels of *high quality* audio recording, and a reasonable workflow. A Sound Devices recorder is only $4k, which is a reasonable price. Automated syncing with .BWF files would also improve the workflow where >4 tracks are desired. I think it's a better approach than recording better-than-normal audio onto the camera. In my opinion, 2 channels of really good analog inputs (with good soft limiters) is likely the best approach. Anything more than 2 channels can go with double system sound, with automated syncing of .BWF files. Just make sure that the timecode clocks don't drift and that the process is mostly foolproof (which I believe are challenges with Keykode, since TC does end up drifting and sync is off).


In keeping with the indie nature of the camera, allowing a boom op to have remote control of audio levels on the camera reduces the need for a recordist or a field mixer.
Trevor, I don't believe that feature will really solve anything. If you have the controls remote, you'll still need someone to operate those controls!

2- Also, you really only need to set the levels on the camera ONCE. If you're using a field mixer, you mix in the field mixer and calibrate the camera to tone.

3- Also check out
http://www.sonosax.com/BD1/SX-BD_InfosheetE.html
A mixer that fits onto a boom, so the boom op can hold the boom and mix at the same time. *Haven't used it myself, but seems like a good idea.

[Stokestack]

In my opinion, 2 channels of really good analog inputs (with good soft limiters) is likely the best approach.

That would be nice. GOOD preamps or a digital input is pretty important.

The nice thing about a digital input is that it would allow us to overcome any on-board shortcomings with an external A/D converter. I'd think that'd make it easier for the Red team to future-proof the camera; they could solve the problem of getting audio data in, and then worry about tweaking an analog section.

[Hrvoje Simic]

It would be really nice if the 4 peak levels were visually placed around the screen in the scheme they record the sound. Simply changing the color of the monitored channel's peak level would in my opinion be the most logical option for multy sound monitoring.

[Steve Freebairn]

I understand that there are issues recording audio while over and undercranking, I just want to know if the option is available?

[Michael Mann]

GOOD preamps or a digital input is pretty important.
Reference quality should be Sound Devices preamps.

[GlennChan]

IMO, digital inputs aren't necessary. If you need that quality (which is most real situations you won't), you're better off recording double system and hopefully Redcine implements an automated syncing solution *that works*. If you shoot film and transfer it, the available solutions don't entirely work perfectly and the telecine operator (+assistant) always have to manually check sync and fix problems.

But if Red could get an automated solution working well, then it allows very high quality audio without much workflow hassles. Perhaps matching timecodes between Red and .bwfs would allow this. Recording audio separately is ultimately the highest quality approach.

[Hrvoje Simic]

Err....has anyone seen a good quality preamps which could fit into RED, especially for 4 channels ?

I wonder how much space for audio have the guys left in there.

[Nick Shaw]

I understand that there are issues recording audio while over and undercranking, I just want to know if the option is available?

As I understand it, the 'issues' are simply to do with how you sync the sound in post. Graeme said previously that sound will be recorded while overcranking. You then just need to decide what to do with sound recorded while shooting at 120fps when the picture is played back at 24fps.

Nick

[Trevor Meier]

Since the RED's built around timecode, Redcine syncing to BWF in double-system should be relatively easy and workable... would save a hell of a lot of time for the online with RPL.

RE: remote control options, just trying to think outside the box... making RED the 'whole widget' isn't necessarily the right idea, but playing with different low-person-count workflows is where I live all the time.

[Nick Shaw]

Since the RED's built around timecode, Redcine syncing to BWF in double-system should be relatively easy and workable... would save a hell of a lot of time for the online with RPL.

It probably still involves a little bit of manual input to get the sync right at non standard frame rates though. As far as I know, timecode can be one of 24, 25, 30, 50 & 60 fps. RED can shoot at fully variable frame rates, but other external sound recorders are unlikely to allow non-standard time-code rates.

RED will allow shooting at eg 120fps, with a 'project setting' of 24 fps, so I assume 24fps time-code is what will be recorded in the picture meta-data. But what about sound? And what if that sound is recorded on an external device running 24fps time-code. The time-codes will not match.

Of course these issues have been around for years with film, and overcranked shots are usually done MOS. But digital overcranking presents the opportunity to improve the situation.

Nick

[Stokestack]

IMO, digital inputs aren't necessary. If you need that quality (which is most real situations you won't), you're better off recording double system and hopefully Redcine implements an automated syncing solution *that works*.

I can see the reasoning there. But given that the Red is going to ingest audio at all, it's going to involve recording a digital bitstream. What's the harm in allowing the user to pipe that bitstream in directly, bypassing the analog section? I don't know the engineering involved, but it doesn't seem like the additional work would be that great, especially if the bitstream is restricted to one of the sampling rates and bit depths that the Red records natively. And if you can eliminate (or render redundant) an outboard solution that has to be synced later, why not do it?

Plus, good stand-alone field recorders are costly (in terms of some of our budgets), whereas an outboard A/D converter and phantom-power pack are less so. I have a little belt-clip A/D converter and matching mic-power unit; I'd like to be able run the A/D's S/PDIF output right into the camera.

You could argue that this will be of minimal extra value once you convert your footage into a QuickTime or AVI file for editing, since you'll almost certainly want to break the audio out into separate files for sweetening in your audio app anyway. But at least they'll start out synced.

[David Fairbanks]

Why would you use an outboard A/D when the camera has one in it? It just seams like extras stuff to carry and get tangled.

I can definitely understand digital in, but that's only if it coming from a sound guys high quality digital mixer that has better pre-amps and A/Ds.

[Brian Broz]

FWIW I believe Sound Devices uses Lundahl pre-amps. Regarded to be clean, warm and Swedish :-)

Brian Broz

[jbeale]

I can see the reasoning there. But given that the Red is going to ingest audio at all, it's going to involve recording a digital bitstream.

Depending on their design, adding external digital input could be just a bit of code in a FPGA, or it could require resources that aren't on the circuit board.

1) Reclocking data: an external audio bitstream will come in based on a different clock than the camera, hence you must have a re-clocking circuit. It might be 48 kHz but it will never be *exactly* the same 48 kHz rate as derived from the camera master clock. Reclocking is not rocket science, but it does take some digital hardware, which may or may not be currently available on the board.

2) Analog feedback: if you want to monitor that bitsteam via the on-camera headphone outputs, you would have to route the incoming data through a D/A to regenerate it as analog. Not sure if Red was doing that already, or was just taking the audio inputs as pure analog signals (pre-A/D), which has the advantage of reducing some audio delay and maybe circuit complexity.

[ericyoung]

As I understand it, the 'issues' are simply to do with how you sync the sound in post. Graeme said previously that sound will be recorded while overcranking. You then just need to decide what to do with sound recorded while shooting at 120fps when the picture is played back at 24fps.

Nick

Surely if you're shooting 120fps, you aren't going to be using sound recorded at 120fps and played back at 24fps (if that's your projection rate)? Probably anytime you shoot "off-speed" you are going to be post-processing sound in some way, rather than using "sync" sound. "Fast" or "Slo-mo" is a visual effect, and rarely if ever will the sound be recorded and used at the same frame rate as the picture. At least I can't think of any examples! :)

Or maybe I'm being dense - it's late!

[Nick Shaw]

Surely if you're shooting 120fps, you aren't going to be using sound recorded at 120fps and played back at 24fps (if that's your projection rate)? Probably anytime you shoot "off-speed" you are going to be post-processing sound in some way, rather than using "sync" sound. "Fast" or "Slo-mo" is a visual effect, and rarely if ever will the sound be recorded and used at the same frame rate as the picture. At least I can't think of any examples! :)

I agree in the main. I did say in a later post in the thread that most overcranked shots are done MOS. But sometimes it can work as an effect. I cut a piece recently with some 100fps boxing shots. The sound recordist had been running the DAT continuously during the takes, with the intention of having a wild-track. I synced up the audio manually, and slowed it to match, and the result was very effective. The director was surprised how effective, and said he would record sync sound on all overcranked shots in future.

Nick

[ericyoung]

I agree in the main. I did say in a later post in the thread that most overcranked shots are done MOS. But sometimes it can work as an effect. I cut a piece recently with some 100fps boxing shots. The sound recordist had been running the DAT continuously during the takes, with the intention of having a wild-track. I synced up the audio manually, and slowed it to match, and the result was very effective. The director was surprised how effective, and said he would record sync sound on all overcranked shots in future.

Nick

Yes, is that the same as recording sound at "100fps". The DAT was just running at normal DAT speed. When you slowed the sound to match, were you really slowing it down an exact 24/100th of full speed, or just until it "felt" right. There was no need to have matching timecode on sound and picture for this purpose.

That's not to say that, when shooting offspeed, recording sound at the same time isn't useful. I would always record sound if available for exactly the reason your director found. But I wouldn't call that "sync" sound, just a wild track recorded at the same time. I'm probably just being overly pedantic. :)

I think the original query was about having recorded sound somehow be recorded at the same speed as picture frame rate so that it would be simple to sync up as they would have the same timecode? (I'm too lazy to go back and check!! :o )

[Nick Shaw]

Continuing the pedantry, I WOULD call what I had 'sync' sound, as I did indeed slow it to 25% speed and sync it with the images. I had no time-code though, and had to do the syncing manually.

If you are interested, you can see the film HERE (http://www.forward-thinking.tv/cnbc/boxing/on_air_film). Actually I just watched it again, and the off-speed sound (only used on a few shots anyway) is pretty inaudible in the final mix, particularly on the web version.

Anyway I'll stop the pedantry now!

Nick

[Stokestack]

Why would you use an outboard A/D when the camera has one in it? It just seams like extras stuff to carry and get tangled.

I can definitely understand digital in, but that's only if it coming from a sound guys high quality digital mixer that has better pre-amps and A/Ds.

Hey David,
doesn't your second statement answer the question in the first?

If you shop around for small flash-based field recorders, you find that many of them are condemned in reviews for having noisy preamps. And these are devices that are devoted entirely to audio. Do we consider the Red camera to be primarily an image-acquisition device, with limited space and resources to devote to audio? I would think so. If so many vendors are having trouble getting decent preamps into dedicated audio recorders, this gives us reason to worry that the ones built into a camera might suffer from similar limitations.

Thus the desire for a digital audio input, which frees us from such limitations while letting us sync audio at the point of acquisition. And a digital audio input requires an outboard A/D converter. So I'm perplexed by your question.

[craesmeyer]

After read, and read all the opinions i would vote to Red had an 2 channel analog input and another 2 channel digital input... It seems more interesting and future-proof than the actual solution EXCEPT if the actual preamps and A/D's in camera are really good...

[Len M.]

All four audio inputs are mic/level switchable

Are they well matched in gain? How close are they matched?



All inputs support 48V phantom power for mics

How much current can they supply per channel?


Each audio input is sampled at 48KHz at 16 or 24 bit depth

What are the noise and dynamic range specs on the mic inputs?

Are all four inputs truly sync'ed at the sample level?

Thanks!


Len Moskowitz
Core Sound LLC
Home of TetraMic