New Rough Guide to recording video location sound - comments please

[Ben Edwards]

Hi,

I have been writing a beginners guide to sound. As well as helping people get started it also aims to ensure people realise the importance of sound and having a good sound recordist. I am not a sound engineer but have written it with the help from some I know. Would be good if people could have a look and let me know what you think (I probably got at least one thing wrong). I am going to add more pictures and it is still in draft form but hopefully nearly there.

It is at Rough Guide to Location Sound Recording - iContact Video Network: Community and Citizens Video, Bristol, UK

Here is the first paragraph to give you an idea:

"It is easy to get carried away with the visual side of filmmaking and let sound take second place; this is a big mistake. Generally sound is as important as picture and sometimes even more so. While picture can grab us emotionally it remains detached and outside the body. Sound by its very nature can be subtle and subconscious or even felt physically. It gets inside you, vibrates you and grabs you emotionally from within. It is almost impossible to be totally immersed and feel something is truly real with images alone, but shut your eyes and listen to a high quality sound design and you will feel that you are actually there."

Regards,
Ben

[Rob Neidig]

Ben,

There's a bunch of good information in your audio article. There are also several half-truths or outright bits of misinformation. I started skimming after about the first half of it, but the one that got me the most was "A low level recording, amplified in the edit to ensure the contributor's voice is loud enough will have a lot less background noise than if the levels were higher when the recording was made and no subsequent amplification is needed." This is absolutely, and patently false. If the recording is done properly, just the opposite is true. Boosting the audio will also boost any self-noise of the recording device along with any background noise. Making a point that having slightly low levels and boosting is better than clipping would be valid. And with today's digital recording, boosting a slightly low level signal is not as bad as it was in the analog days, but your statement is just completely wrong. The important point should be getting the correct sound level to begin with and having the highest signal to noise ratio you can get.

Other wrong or misleading bits include "Lavelier mics are omni (non) directional." (they can be both directional and non-directional) "All the decent quality microphones require power." (There are tons of great-sounding dynamic mics out there that do not require any kind of power) "With boom poles the mic comes in from above but with a Pistol Grip it is positioned low pointing upwards towards the mouth." (a pistol grip is quite often also used from above, and in fact often mounted to the end of a boom pole for the task). And there may be others bits I missed once I started skimming.

I'd have a good audio tech go over your whole article before pointing people toward it as a guide.

Have fun!

Rob

[Ben Edwards]

Rob,

Thanks for that. The bit about level does need rewriting and the other info you provided is all good. The broadcast sound recordists I know is back this week so we will go over it in detail. Thanks for your help.

ben

[Ben Edwards]

Originally Posted by Rob Neidig

Ben,
.... If the recording is done properly, just the opposite is true. Boosting the audio will also boost any self-noise of the recording device along with any background noise. Making a point that having slightly low levels and boosting is better than clipping would be valid. And with today's digital recording, boosting a slightly low level signal is not as bad as it was in the analog days, but your statement is just completely wrong. The important point should be getting the correct sound level to begin with and having the highest signal to noise ratio you can get.

Rob

Have made a lot of changes that should make things better. Trying to get my head around the noise to signal ratio. Also someone mentioned that the signal will not have as much resolution if recorded low. Is this related to signal to noise ratio or is the resolution thing a red herring?

Basically I am trying to get a list of all the reasons that recording very low signals are bad and exacly what signal to noise ration means in this context (I understand the term generally). Iguess if you have a cheaper recorder the pre-amps will introduce more noise but I thouhht this would get worse as the levels increased. Maybe I dont understand what self-noise is, more insight into this would be good.

Regards,
Ben

[Dave Blackhurst]

Haven't read your article, but perhaps I can shed a bit of light on your question....

Pretty much any piece of electronics will add "color" and some background noise - IOW, as the device or combination of devices is used to record, it will impose a certain tonality that in theory is supposed to be "transparent" and exactly what was there, but in practicality the recording will reflect both the equipment used and the relative ambient conditions (such as distance of the source from the mic). So the resulting recording is a REPRESENTATION of what audio occurred, but may have varying degrees of accuracy and "pleasant" or "harsh" aspects (read some of the analysis of different mics for a perspective).

On top of this, electronics usually create a varying but hopefully relatively small amount of internal noise, usually a bit of "hiss", but there can be other frequencies - the hiss seems to be the tough one to eliminate. This is usually expressed with a spec (S/N or signal to noise ratio) to espress how much noise you can expect the device itself to generate against a given signal level. Think of it like this - there in theory should be NO noise added by the device, but in reality it's there, but at a low level so it's considered innocuous in MOST situations as long as a sufficiently strong signal is present to overwhelm the noise for the recorded signal or the listener.

The problem arises when the source "signal" is too week to rise above the noise floor effectively, meaning it is so close to the noise level that it mixes in with the noise. This is of course effectively the opposite of distortion where the signal is so "hot" that it hits the limitations of the electronic devices used and results in the waveform being clipped.

So when recording audio, the key is to get a signal sufficiently strong as to be clearly above the "background" noises, both internal to the devices used for recording, and ambient (i.e. using a "close mic" in a noisy environment), but not recording the signal so hot that it distorts - of course in music, distortion is a creative tool, artists having used "overdriving" EVERY device in the signal chain for artistic purposes! For "normal" purposes your goal should be to acquire as clean and crisp a recording as possible of the original source.


Let's take it out of the realm of electronics for a moment - you have a room full of people chatting amongst themselves vs. a stadium of crazed fans screaming - those would represent two different "noise" signatures, both challenging, but to varying degrees. For someone to be heard in the first instance, they might need only to speak loudly to be heard and understood - in the second, they probably would require a loudspeaker to augment their "signal". On top of this you run into all sorts of potential issues with the frequencies of the background noise vs. the "signal" frequencies, phase variances, cancellation, tonal coloration, etc. This is why one typically will find themselves "listening" to audio in a quiet and audio optimized theater designed to minimize both the color and background noise!

Hope that puts a bit of light on the subject... er... clarity since this is the audio forum, not the lighting one <wink>!

[Ben Edwards]

Dave, thanks for your reply.

"On top of this, electronics usually create a varying but hopefully relatively small amount of internal noise, usually a bit of "hiss", but there can be other frequencies - the hiss seems to be the tough one to eliminate. This is usually expressed with a spec (S/N or signal to noise ratio) to espress how much noise you can expect the device itself to generate against a given signal level. Think of it like this - there in theory should be NO noise added by the device, but in reality it's there, but at a low level so it's considered innocuous in MOST situations as long as a sufficiently strong signal is present to overwhelm the noise for the recorded signal or the listener. "

So basically as electronics are not perfect they introduces noise to the signal (machine noise?). So if the signal is very low and you have to amplify in post you will also amplify the 'machine noise'. Am I to take it that as you turn up the levels you do NOT also turn up the machine noise, therefor as you turn up the levels the S/N ratio increases?

[Paul R Johnson]

Having spent two days locked up in London conference room writing a new version of a popular UK school/college qualification, I've still got the right head on.

I picked this section

Halving the distance between mic and the sound source quadruples the quality of the sound

the square law and the inverse square law are of course valid, bit the language you are using in the body is aimed at people with a decent level of comprehension, so the oversimplification here could be confusing.

The halving of the distance increases microphone output on the wanted source by 6dB. It has no effect on 'quality' at all, this remains exactly the same. All that changes is the ratio between wanted an unwanted sound. This reduces the background noise and firms up intelligibility. With an omni mic, the tonal balance remains the same too, not cardioid bass emphasis as the mic gets close in. Less gain is needed, and noise goes down.

A singers vocal mic may well be omni if it is a boom mounted, over the ear, musical theatre typical mic, but the Britney, Madonna and of course hand held mics are invariably cardioid, or even hyper-cardioid. Mentioning singer could confuse?

I don't like the stereo section, X/Y is usually explained as being at 90 degrees to each other, not normally 45 degrees each. M/S hasn't been mentioned which is more common in video than in normal recording because of mono compatibility.

I'm afraid I disagree completely with your statement that minidisc is not good enough for music. The fact that they are on the way out as solid state is more reliable and convenient is very accurate, but having DAT and MD available on proper pro machines, I proved to myself a very long time ago that MD quality is perfectly sufficient for broadcast and theatrical playout - the early ATRAC compression was a little rough in the cheap machines, but the pro machines from Sony, Tascam, Denon and a few others were actually very transparent. Not as good as 48K DAT, but very, very close. In blind testing between DAT and the Tascam 801, although there was a difference detectable in the sound, our tests indicated that it was impossible to state that one was better without resorting to hi-fi technosnakeoil babble.

In fact, the MD machines produced in the past 5 years or so have amazing capability to handle over level signals, far better than Solid state devices recording uncompressed to cards.

I think the style and photos really help, but my comments above suggest you're wavering when it comes to explanation - some sections are quite high level, others over simplified so the meaning gets blurred. This is so easy to do, so you level wavers up and down - a good example is where you need to explain a complex technical area - like phantom power. I'd suggest that perhaps explaining it as the power superimposed on top of the audio might help understanding - because you use the analogy of the audio being the difference between the two cores, it is difficult to find quite the words to explain it to beginners. Perhaps a simple diagram, showing the circuit could work better - one 'sharing the screen', the other with the separate screen, and then maybe a dotted line with arrows going back to the mic for power. It's a good job you've done, but really green people could misunderstand. It's common when trying to write this kind of material to envisage the 'average reader' - for us, this is why we moan so much about manuals - they have to cope with people who find fitting the battery difficult, so take ten pages to get to the state when a product is turned on!

You are talking about recordists and 'proper' video stuff, so maybe trying to dumb down too much wrecks the professional end at the expense of real unskilled, non-technical beginners?

I spent an hour today, trying to decide if I could use certain words. If you specify "understand" as a requirement, it is fine - it can be tested. "Know" however can't! So you can say that somebody understands how to select an audio recorder after reading the article, but you will never be able to guarantee they know about something. Language is damn difficult. There's a notion nowadays that our tendency to oversimplify to guarantee understanding, is actually doing them a disservice, as they have learned it 'wrong'..

In lighting, for years we've always told students Amps x Volts = Watts, but strictly speaking this has never been quite correct in this context because we're dealing with AC and sometimes reactive loads. Let's not go there!

[Rick Reineke]

Well said Paul!
My sentiments exactly.

[Ben Edwards]

Paul, Thanks for taking the time to reply.

Originally Posted by Paul R Johnson

the square law and the inverse square law are of course valid, bit the language you are using in the body is aimed at people with a decent level of comprehension, so the oversimplification here could be confusing.

The halving of the distance increases microphone output on the wanted source by 6dB. It has no effect on 'quality' at all, this remains exactly the same. All that changes is the ratio between wanted an unwanted sound. This reduces the background noise and firms up intelligibility. With an omni mic, the tonal balance remains the same too, not cardioid bass emphasis as the mic gets close in. Less gain is needed, and noise goes down.

It is a beginners guide so it is aimed at people with little or no understanding of sound but the wording could be misleading so I have changed the section to

"It seems logical that the closer you get to the source of the sound the better the sound (i.e. there will be less interference from background noise). This is indeed true but is even more important than you may first think. As with lighting, the M2 rule applies for sound. Halving the distance between mic and the sound source quadruples the usability of the sound, and the inverse is true. As you move away from the source the sound becomes exponentially less usable. By usability I mean it does not pick up unwanted background noise. The closer you get to the sound source the less background noise you get."

Originally Posted by Paul R Johnson

A singers vocal mic may well be omni if it is a boom mounted, over the ear, musical theatre typical mic, but the Britney, Madonna and of course hand held mics are invariably cardioid, or even hyper-cardioid. Mentioning singer could confuse?

Yes, I have been struggling with this a bit, I have been trying to make the guide pragmatic and short without misleading people. I original have laveliers as an example of a omni directional mic but have been told some are directional (not sure which, are cos 11, rode and Sennheiser Lav directional). Does anyone have an good example of an omni microphone or is it that generally mikes are either very directional or slightly directional ?

Originally Posted by Paul R Johnson

"I don't like the stereo section, X/Y is usually explained as being at 90 degrees to each other, not normally 45 degrees each. M/S hasn't been mentioned which is more common in video than in normal recording because of mono compatibility.

OK, changed 45 to 90. Looks like this section needs rewriting. So Far I have

"Stereo mics are can be thought of as a set of directional mono mics pointing is different directions. There are 3 main configurations.

* X-Y where there are two directional microphones one pointing left and one right (generally 90 degrees) that are close together. This does allow the two elements of the microphones to be close together or often in the same unit so is convenient but is generally less "spacey" and has less depth compared to recordings employing an below A-B setup.
* A-B Where two onmi directional microphones are used spaced some distance apart (i.e. 50cm). The left right separation is created through the time it takes sound waves to reach the microphones rather than the direction they are pointing. The 'stereo width' can be altered by moving the microphones further apart.
* M/S or mid-side. This configuration used This coincident technique employs a bidirectional microphone facing sideways and a cardioid pointing forward. In this configuration the stereo width can be manipulated in post so it is popular in film and video.

For X-Y and A-B two separate microphones can be used. The Rode NT5s can be purchased as a matched pair and mounted on a bar creating a very flexible stereo configuration (or can be used separately). "

Looks like I need to add bidirectional microphones to omni and directional. I don't quite understand the practicalities of MS, I guess I need to get my head around bidirectional. Is it always a bidirectional and a cardioid or are there other setups?

Originally Posted by Paul R Johnson

I'm afraid I disagree completely with your statement that minidisc is not good enough for music. The fact that they are on the way out as solid state is more reliable and convenient is very accurate, but having DAT and MD available on proper pro machines, I proved to myself a very long time ago that MD quality is perfectly sufficient for broadcast and theatrical playout - the early ATRAC compression was a little rough in the cheap machines, but the pro machines from Sony, Tascam, Denon and a few others were actually very transparent. Not as good as 48K DAT, but very, very close. In blind testing between DAT and the Tascam 801, although there was a difference detectable in the sound, our tests indicated that it was impossible to state that one was better without resorting to hi-fi technosnakeoil babble.

In fact, the MD machines produced in the past 5 years or so have amazing capability to handle over level signals, far better than Solid state devices recording uncompressed to cards.

OK, trying to work out exacly how to reword this. I guess I could say that top end MD players are good for music but cheaper ones are lacking. I am guessing the second hand £20-£30 ones on eBay are inferior and should be avoided?

Originally Posted by Paul R Johnson

I think the style and photos really help, but my comments above suggest you're wavering when it comes to explanation - some sections are quite high level, others over simplified so the meaning gets blurred. This is so easy to do, so you level wavers up and down - a good example is where you need to explain a complex technical area - like phantom power. I'd suggest that perhaps explaining it as the power superimposed on top of the audio might help understanding - because you use the analogy of the audio being the difference between the two cores, it is difficult to find quite the words to explain it to beginners. Perhaps a simple diagram, showing the circuit could work better - one 'sharing the screen', the other with the separate screen, and then maybe a dotted line with arrows going back to the mic for power. It's a good job you've done, but really green people could misunderstand. It's common when trying to write this kind of material to envisage the 'average reader' - for us, this is why we moan so much about manuals - they have to cope with people who find fitting the battery difficult, so take ten pages to get to the state when a product is turned on!.

You are talking about recordists and 'proper' video stuff, so maybe trying to dumb down too much wrecks the professional end at the expense of real unskilled, non-technical beginners?

I spent an hour today, trying to decide if I could use certain words. If you specify "understand" as a requirement, it is fine - it can be tested. "Know" however can't! So you can say that somebody understands how to select an audio recorder after reading the article, but you will never be able to guarantee they know about something. Language is damn difficult. There's a notion nowadays that our tendency to oversimplify to guarantee understanding, is actually doing them a disservice, as they have learned it 'wrong'..

In lighting, for years we've always told students Amps x Volts = Watts, but strictly speaking this has never been quite correct in this context because we're dealing with AC and sometimes reactive loads. Let's not go there!

Thanks for this, I guess I am going for a pragmatic guide for beginners and want to add a lot more links for further reading. The idea is not to scare people off with a book but have 10 or so pages of useful stuff to help them vastly improve the sound they get. In terms of getting the level right it is a tricky one. The phantom power description was added my a friend and although it is interesting may of gone into too much detail. May reword it.

Thanks for sharing your time and experience.
Ben