Never put mic on camera?

[jcs]

I meant particle physics at the atomic level- air molecules are particles, vs. quantum physics (sub-atomic: fermions, quarks, etc.). We can model waves on the computer as vibrating particles to get a good basic understanding of wave dynamics. Surprisingly couldn't quickly find an example on YouTube. This should provide the basic idea:

I was thinking of 3D particle simulation, where particles can also move in any direction, so spherical wave propagation as with sound in the real world vs. what is shown in the above video (where it's really 2D simulation with particles constrained to 1 degree of freedom: up and down).

Many years ago I needed a jet engine sound and didn't have the means to record a real one. So I wrote software to simulate the turbine (multiple sine waves) and a simple "particle" simulator to model low-frequency turbulence. It sounded pretty good! (was for an early VR videogame)

Agree if it's possible to simply record the real thing, that's way easier and faster! At the same time, if it's possible to model a sound, we can use it for interactive software applications, which can sound better than doing cross-fades, which can still work pretty good (actual Z06 Corvette engine cross-faded in real-time (recorded with the ancient Sony MZ-1. The hardest part of creating that simulator was the tire simulation- there were no good books/resources at the time and still probably aren't any: it's a surprisingly hard problem)):

Since it's possible to get excellent results simulating the Navier-Stokes fluid dynamics equations, in real-time no less(!), it's possible to model sound as well:

This is closer (2D):

Surprised to see not that many good sound simulation visualizations on YouTube!

Back to the mic-on-camera question about how sound changes with distance. Since sound travels around 761MPH at sea level, there's not much time for turbulence to really affect the sound as it travels from the source to the mic. Thus I'd expect the sound not to change very much with distance in an anechoic room, and in a quiet outdoor environment with only the ground as a reflector (perhaps at a certain distance sound reflections will start having an effect as they'll be less off axis (directional mic)). Any real-world examples (or simulations!) showing significant changes to sound with distance (and no reflectors (anechoic room) and/or just the ground)?

 

[Noiz2]

Actually molecules are not particles, they are made up of particles. Electrons, protons, neutrons etc. are particles. Sound though is a compression wave and diminishes by the square of the distance so distance matters. And everything being "perfect" frequencies have different energy levels so as the propagation wave looses energy as that energy is transformed into heat the energy loss is not linear so thee will be frequency shifts depending on exactly what the media is that it is propagating through.

RE: turbulence, ever see the heat shimmer on a hot day? That is turbulence affecting light, which travels a tad faster than sound.

The basic error of thinking about sound in terms of particles is that particles travel in a direction and sound propagates in all directions. On the receiving end you would be capturing the particle that was "sent" so it is the original, on the sound side you are getting pressure waves of your local "air" no "thing" has traveled just a pressure wave. The "original" never left the source. In my "shimmer" example the light is being bent by changing density but the photon you recieve is the one that started the trip. That is an over simplification but the essence is correct. A sound wave on the other hand is very interactive and is physically changed by everything it interacts with.

If you want to go down this particular rabbit hole though I suggest you start a new thread because as interesting as this is it has little to do with camera mounted mics.

 

[Noiz2]

Many years ago I needed a jet engine sound and didn't have the means to record a real one. So I wrote software to simulate the turbine (multiple sine waves) and a simple "particle" simulator to model low-frequency turbulence. It sounded pretty good! (was for an early VR videogame)

It's not that it can't be done, some things can be simulated quite well. For many things a recording will be faster easier and sound much better. Especially for any sound that lasts any length of time. Some things are very hard to record, or just really hard to get a clean recording of. Years ago I was on a film that needed a rural school bus honk. It was no problem to go down to the bus yard and record, and we did. But the bus yard was close to the airport and in the end we never got a clean enough honk to work for one on a rural road. But american horns are square waves a fifth apart so I pulled out a DX7 and made a bunch and the editor was pleased as punch.

In most cases the things that can't practically be recorded though are built up from recordings of other things. It's a fundamental skill in editing/ design, to be able to break down the sound you want into sounds you can get and then build the sound you want. Which is essentially what you did with synthesis. The general problem with synthesized sounds is that they are often too "sterile", like a lot of CG. CG has come a long way in adding real world "dirt" so I suspect sound will get there also. Right now mechanical sounds can work pretty well but many "natural" sounds lack "life". Also size matters, games can get away with more than a feature film on the big screen can.

 

[jcs]

It's not that it can't be done, some things can be simulated quite well. For many things a recording will be faster easier and sound much better. Especially for any sound that lasts any length of time. Some things are very hard to record, or just really hard to get a clean recording of. Years ago I was on a film that needed a rural school bus honk. It was no problem to go down to the bus yard and record, and we did. But the bus yard was close to the airport and in the end we never got a clean enough honk to work for one on a rural road. But american horns are square waves a fifth apart so I pulled out a DX7 and made a bunch and the editor was pleased as punch.

In most cases the things that can't practically be recorded though are built up from recordings of other things. It's a fundamental skill in editing/ design, to be able to break down the sound you want into sounds you can get and then build the sound you want. Which is essentially what you did with synthesis. The general problem with synthesized sounds is that they are often too "sterile", like a lot of CG. CG has come a long way in adding real world "dirt" so I suspect sound will get there also. Right now mechanical sounds can work pretty well but many "natural" sounds lack "life". Also size matters, games can get away with more than a feature film on the big screen can.

Yeah, it's cool to be able to make things from scratch, by whatever means, when "real" isn't possible.

The trick to making the jet engine sound realistic, was randomness: noise. By combining various sine wave frequencies and noise, I was doing Fourier Synthesis (didn't know it at the time). Years later I learned to do spectral shaping: FFT + filtering + iFFT. Similar things can be done with wavelets in the time domain (doesn't need windowing).

Regarding "life" for sound synthesis, we'll be seeing (hearing!) some very cool things by way of machine learning in the near future!

 

[jcs]

Actually molecules are not particles, they are made up of particles. Electrons, protons, neutrons etc. are particles. Sound though is a compression wave and diminishes by the square of the distance so distance matters. And everything being "perfect" frequencies have different energy levels so as the propagation wave looses energy as that energy is transformed into heat the energy loss is not linear so thee will be frequency shifts depending on exactly what the media is that it is propagating through.

RE: turbulence, ever see the heat shimmer on a hot day? That is turbulence affecting light, which travels a tad faster than sound.

The basic error of thinking about sound in terms of particles is that particles travel in a direction and sound propagates in all directions. On the receiving end you would be capturing the particle that was "sent" so it is the original, on the sound side you are getting pressure waves of your local "air" no "thing" has traveled just a pressure wave. The "original" never left the source. In my "shimmer" example the light is being bent by changing density but the photon you recieve is the one that started the trip. That is an over simplification but the essence is correct. A sound wave on the other hand is very interactive and is physically changed by everything it interacts with.

If you want to go down this particular rabbit hole though I suggest you start a new thread because as interesting as this is it has little to do with camera mounted mics.

Particles are short-hand for things we study as single objects, even though they may be made up of smaller things (particles again!): https://en.wikipedia.org/wiki/Particle

In the physical sciences, a particle (or corpuscule in older texts) is a small localized object to which can be ascribed several physical or chemical properties such as volume or mass.[1][2] They vary greatly in size or quantity, from subatomic particles like the electron, to microscopic particles like atoms and molecules, to macroscopic particles like powders and other granular materials. Particles can also be used to create scientific models of even larger objects depending on their density, such as humans moving in a crowd or celestial bodies in motion.

That's the same way I describe them, as when modeling physics on the computer we need a finite object to reference elements which interact with each other: particles. A particle system is a common term for visual effects used for sparks, smoke, explosions, etc. Ultimately, at the fundamental level of reality (as far as we now know- this may change!), everything is a wave + probability function: no particles at all!

To get a very accurate simulation of motion in fluids (and air), we can use the Navier-Stokes equations. They are accurate enough to model turbulence to design wings, aircraft, turbochargers, boats, propellers, etc. However, for sound recording, the relative speed of air turbulence is far lower than the speed of sound, and thus it doesn't seem likely that turbulence will have a (statistically) significant affect on recordings as perceived by human beings. Recent advances in algorithm design and GPUs allow the Navier-Stokes equations to run in real time for water and smoke simulation, which is pretty impressive. It would also work for air/sound-wave simulation.

When we model sound with particles, the particles represent air molecules, and they won't be whizzing about like in a particle accelerator. They'll be vibrating and not really moving much, just like real air molecules (acting like particles). Sure, there will be air movement and turbulence, and again the particle motion of turbulence is very slow compared to the speed of sound, and thus won't effect the sound in a significant way. This is in open space, with little or no reflectors, such as an anechoic room or outside as described by combatentropy. Obviously turbulence plays a major role in sound coming from a wind instrument, organ, etc.

As high frequency sound carries less energy, sure, the higher frequencies will attenuate faster than the lower frequencies via air friction. For sound recording, especially mic-on-camera, when will this become significant so that we can hear it? Just from real world experience, I don't recall a symphony having significant frequency changes sitting close or far from the orchestra. Thus for sound recording, mic-on-camera, it doesn't seem like this will be significant. The major factor is the inverse square law, and when reverb isn't causing an issue, if we have a low-noise mic and AD converter, distance won't change the sound significantly (beyond the bass boost of the proximity effect when very close). If this is not true, where is the real-world evidence?

How does this sound?

1. We can use a mic mounted camera when it fits the scene/shot, when there aren't any other options, or when 'good enough' is good enough (haha). It's important to understand the benefits of booming and lavs, when possible/available.
2. Sound quality doesn't change significantly as the mic moves away from the source (beyond the proximity effect). Instead, the signal to noise ratio drops, as reverb level rises. In an anechoic room or quiet outdoor, non-reverberant space, distance primarily increases system noise relative to the signal.

 

[Noiz2]

2. Sound quality doesn't change significantly as the mic moves away from the source (beyond the proximity effect). Instead, the signal to noise ratio drops, as reverb level rises. In an anechoic room or quiet outdoor, non-reverberant space, distance primarily increases system noise relative to the signal.

Won't argue the science but practical experience is that the quality of sound does indeed change with distance. The whole "anechoic room" is kind of irrelevant from a practical stand point since that is not where films are shot.

 

[indiawilds]

Hopefully some moderator can cull some of the posts from this thread to a new thread

 

[jcs]

Won't argue the science but practical experience is that the quality of sound does indeed change with distance. The whole "anechoic room" is kind of irrelevant from a practical stand point since that is not where films are shot.

The anechoic example was a thought experiment; combatentropy mentioned a real-world example outside in a quiet space (no reflectors). We could generalize the statement saying that as reflectors (reverb) increases, sound quality also decreases with distance. For example, a single sound reflector (the ground) didn't significantly hurt sound quality when the mic was placed relatively far from the source.

Here's an example using a shotgun vs. parabolic mics. The Pro Mono parabolic at 55 feet sounds pretty impressive. The wooded/tree space with no wind has no apparent reverb and sounds amazingly clean and clear + birds are high frequency
As per the thought experiment, the shotgun was still pretty good, however the system noise became apparent, which reduced quality (as opposed to reverb and/or frequency changes with distance):

Birds and animals from up to 350' away:

Pretty amazing dialog from a long distance away (in a noisy environment no less)!

Thus sound quality doesn't degrade much with distance.

Parabolic > long shotgun > short shotgun > super/hyper/cardioid > omni with distance.

 

[IronFilm]

None of us ever record in a perfectly quiet environment (often it is very far from that!).

A camera mic is usually always further away than the ideal location for a mic.

R^2 rule therefore means this a bad idea. As simple as that.

Pretty amazing dialog from a long distance away (in a noisy environment no less)!

Thus sound quality doesn't degrade much with distance.

Odds are a lot of that was done with a wireless lav.

 

[jcs]

None of us ever record in a perfectly quiet environment (often it is very far from that!).

A camera mic is usually always further away than the ideal location for a mic.

R^2 rule therefore means this a bad idea. As simple as that.

Odds are a lot of that was done with a wireless lav.

1. There are no rules
2. Mic on camera is exactly the right choice for some shots: cinema verité / first person point of view (include 3D audio with VR)
3. Sound quality does not fundamentally decrease with distance. Signal to noise ratio decreases with distance, and with the right tools, excellent quality can be captured from large distances
4. Understanding principles and physics is useful vs. rote memorization without understanding
5. While it's possible some NFL recordings were lavs, it's likely the birds and animals were not
6. The fundamental point of this thread is to challenge limited thinking: question everything!
7. There are no rules

 

[David Jimerson]

1. There are no rules
2. Mic on camera is exactly the right choice for some shots: cinema verité / first person point of view (include 3D audio with VR)
3. Sound quality does not fundamentally decrease with distance. Signal to noise ratio decreases with distance, and with the right tools, excellent quality can be captured from large distances
4. Understanding principles and physics is useful vs. rote memorization without understanding
5. While it's possible some NFL recordings were lavs, it's likely the birds and animals were not
6. The fundamental point of this thread is to challenge limited thinking: question everything!
7. There are no rules

There are no rules, but there are most definitely practices which work better than others, and purposes for which different practices may be appropriate.

You need to understand why those practices work before you can understand when it's appropriate to switch something up. Simply saying "there are no rules" doesn't really help inform on that.

 

[Noiz2]

None of us ever record in a perfectly quiet environment (often it is very far from that!).

A camera mic is usually always further away than the ideal location for a mic.

R^2 rule therefore means this a bad idea. As simple as that.





Odds are a lot of that was done with a wireless lav.

Actually it is almost definitely a very directional mic. Even a good shotgun picks up some amazingly "close" sounds from a distance. We were recording a NFL practice game and I could easily hear the huddle from the sideline with a long shotgun. There is an extremely directional mic that was developed with this use in mind, think Sennheiser but?, that uses a bunch of capsules and a special processor to get super tight patterns.

You can't use a mic like that camera mounted because the aim will be off.

 

[Noiz2]

1. There are no rules
2. Mic on camera is exactly the right choice for some shots: cinema verité / first person point of view (include 3D audio with VR)
3. Sound quality does not fundamentally decrease with distance. Signal to noise ratio decreases with distance, and with the right tools, excellent quality can be captured from large distances
4. Understanding principles and physics is useful vs. rote memorization without understanding
5. While it's possible some NFL recordings were lavs, it's likely the birds and animals were not
6. The fundamental point of this thread is to challenge limited thinking: question everything!
7. There are no rules

Great summation of why so many new filmmakers films sound so bad.
Success in challenging rules requires understanding those rules and why they are rules.
But of course there never were any "rules", just best practices. If you swap those two you see how ridiculous #1 and #7 are.
#2 ALL audio is 3D
#3 when someone actually accomplished this let us know.
#4 practical experience trumps both
#5???
#6 I question the point of the thread.

For me this thread has gone off the rails into silliness so I shall bow out and leave.

 

[GaryNattrass]

But don't rode make one of those mics that you can zoom in on the sound ? rogar-Happy(DBG):

 

[Paul F]

I think everyone is getting a little stressed on this thread. It's time for one of JCS's videos to relax everyone. From his Cosmic Flow series: https://www.youtube.com/watch?v=M5J9N1xLHRg. I wonder how it was mic'ed. Was that a parabolic or a mic on the camera?

 

[combatentropy]

I suspect the NFL video is mainly lavs, because:

- the title of the video and others in the series is: "Mic'd up"
- miking up players is a thing, Reddit discussion
- by contrast, a parabolic mic of a person talking at a distance is quiet and thin, Youtube demo. I've never used parabolic mics. This is just what I could scrounge up in a minute of searching.

I myself am happy when a thread turns from the normal specification speculation, though I skipped over some of the longer paragraphs

 

[Publimix]

Never put microphone on camera?

I am a simple person, my camera (FS7) has a microphone holder so I put a microphone on it. Doesn't mean I use the recorded sound. And doesn't mean I use other microphones as well and doesn't mean I use other sound-sources. The camera has an intern microphone too but that sucks.

I think a microphone on camera can always be handy, and why not, audio tracks enough. For syncing with other audio or just making audio-notes while shooting or during rough cut editing and dozens of other reasons.

 

[Run&Gun]

I suspect the NFL video is mainly lavs, because:

- the title of the video and others in the series is: "Mic'd up"
- miking up players is a thing, Reddit discussion
- by contrast, a parabolic mic of a person talking at a distance is quiet and thin, Youtube demo. I've never used parabolic mics. This is just what I could scrounge up in a minute of searching.

I myself am happy when a thread turns from the normal specification speculation, though I skipped over some of the longer paragraphs

Yeah, we're straying from the OT, but... One of my friends/audio guys works for NFL Films for home games here doing the "wired" sound. They have a system down and can wire the pads up very quickly. The players know they are mic'd, but the way they are wired, it doesn't affect them. It's not in the way. And believe it or not, they almost never lose equipment to damage. Another friend that shoots for the team in-house uses the Q5X mentioned in the reddit article. I needed to wire up a college football player last year for a shoot, My guy that does NFL Films wasn't available that day, so he walked one of my other guys through how to do it before hand. I first had the producer sign off that they were 100% responsible for replacement of any equipment damaged. We used one of my Lectro belt packs(a UM400a) and a Sanken COS-11. Sounded great(better than either of us thought it would, for sure) and nothing got damaged. And the battery lasted the entire time(pre-game and entire game). The Zaxcoms that the NFL uses can be put into standby remotely. UM400a's can't.

 

[IronFilm]

There are no rules, but there are most definitely practices which work better than others, and purposes for which different practices may be appropriate.

You need to understand why those practices work before you can understand when it's appropriate to switch something up. Simply saying "there are no rules" doesn't really help inform on that.

Agreed, if there is "no rules" you could just record total white noise and say it doesn't matter because "no rules"!

Welllll... NO! Life just isn't like that in reality.

And the better and deeper you know what "the rules" are, then you'll better be able to to know when it is appropriate to "break" the rules.

I suspect the NFL video is mainly lavs, because:

- the title of the video and others in the series is: "Mic'd up"
- miking up players is a thing, Reddit discussion
- by contrast, a parabolic mic of a person talking at a distance is quiet and thin, Youtube demo. I've never used parabolic mics. This is just what I could scrounge up in a minute of searching.

I myself am happy when a thread turns from the normal specification speculation, though I skipped over some of the longer paragraphs

Just from watching the video, and using your eyes and ears, you can tell there are clips there that were not done with a long shotgun (or a parabolic mic) because what you hear simply doesn't match up with what you see in the scene that you'd get with a long shotgun.

 

[jcs]

There are no rules, but there are most definitely practices which work better than others, and purposes for which different practices may be appropriate.

You need to understand why those practices work before you can understand when it's appropriate to switch something up. Simply saying "there are no rules" doesn't really help inform on that.

Lol did you read items (2) & (4) in the above list or the 5 pages before this post? It was an epic battle of egos covering just about every angle!

A simpler, more helpful comment for beginners would be: as distance from microphone source increases the signal to noise ratio decreases. In other words, as we move the mic closer to the source, the speaker's voice will sound increasingly louder compared to all other sounds. This isn't "better" or "worse", any more than changing treble or bass is good or bad, it's what's the purpose of the recording? For some shots we want background noises to be louder when the speaker's voice is supposed to be farther way.

Sound quality doesn't magically decrease with distance: only signal to noise ratio changes. Many focused on the NFL examples possibly using lavs, but didn't pay attention to the animal recordings with parabolic mics: confirmation bias holding on to "mic on camera is 'bad'".

What's the point of this thread? Don't listen to so-called experts or pros. Think for yourself, and if 'experts' don't make sense, do what's right for you. Don't blindly listen to me either (except think for yourself)! I don't mind being challenged, and there were many challenges in this thread. I learned new things. Hopefully others did too and enjoyed the process.