There seems to be a recurring question about how much a good mic placed on top your camera will improve your audio. The short answer is it will most likely improve it but it still will sound like crap. Here's a short video showing how much difference it can make.

Sanken CS-3e example - YouTube (http://youtu.be/TrkCTBkAOsI)

The on-camera mic in the video really had a lot of echo. And I mean, a LOT. Can't say that I ever remember getting that much. For just ordinary family videos and the like I use a Rode Stereo Video Mic on the hot shoe because it's easy and fast to set up. So far, I haven't noticed THAT much difference but maybe I've been lucky.

If there's any consolation, at least the audio from the on-board SVM is better than the built-in camera mic.

It would be interesting to hear what innovative solutions there are for one-man shows. Using a light stand with a boom that was posted earlier is a good idea. Lavaliers are an option. Hidden mic with a cable (but not practical for most family-type gatherings). At this point I'm running out of ideas.

That particular room was super live. Yes, it did have that much echo. No matter what mic you used on camera there would be that much echo. We actually had three cameras running and the other on camera mic had the same amount of echo.

For this interview we did boom in the mic on a light stand that had a boom arm. It allowed us to place it without having a shadow.

For Round-Table discussions I've used this £25 puppy. I connect to an Senni TX and I can then "circle" the subjects, filming away. I don't know HOW it does it, mainly it's because it uses the table as a gathering and sound board, but the quality is quite acceptable indeed: Audio-Technica - Products - Microphones by Application - Fixed Installation - ATR4697 (http://eu.audio-technica.com/en/products/product.asp?catID=1&subID=4&prodID=3702)

Sometimes "simple" does work. It also underlines that old rule: Get the mic as near to the audio source as is visually possible.

Grazie

Yeah, the amount of echo depends on the room and mic placement most. After that, the off-axis performance of the mic will affect the characteristics of the recorded echo.

One tip is to get some stands and blankets and place them strategically. Last year, I did a recording in a very large room with large glass windows. It didn't sound all that live to the ear, but even with a well-placed hyper-cardioid mic, the echo was unfortunate. To this day, I wish I had hang a single blanket between the interviewee and the window. That simple, inexpensive fix would have changed the audio from not-bad to totally pro. And I wanted totally pro!

I've had stands and blankets on every shoot since!

Not long ago, friends shot some green screen video in a barn using a so-so shotgun. Though there were lots of items in the barn to break up echos, we were on a cement floor and the actor was facing the door. We put blankets under his feet and opened the door. (Brrr.) Fortunately, it wasn't windy and we were far from traffic. Echo problem solved. (And, yes, a Sanken CS-3e would have helped!)

Interesting video, Garrett. Camera mounted mics should indeed be discouraged.

Of course, the video also illustrates the little publicised effects of using phantom powered condenser mics in close proximity to newly-washed hair.

Not quite as extreme as the other example below where a phantom supply of 48000V was accidentally supplied to the boom mic:

I fully agree that proper mic placement is something that everybody needs to try.

As a wildlife filmmaker, I have decided to place my Sennheiser MKH 416 directly on the C300. I am able to get many useful sounds. Last month I was filming without the Sennheiser on my C300 and the desert fox came pretty close and started calling. And worse, my assistant thought that I am filming and recording sound, so didn't even move to pick up the microphone for fear of disturbing the fox which was directly staring at us and also for fearing that the onboard sound will get ruined. And then later both of us realised that we didn't record any sound. As humans, we mess up. So for certain situations, better to have some sound rather than nothing.

On another occasion, I was waiting for a leopard. Dawn was about to break in. Light was too low. A bird (Shikra) flew and sat on the branch over my head and started calling. I just pointed my camera up and recorded the sound. So I feel while one should place the microphones as closer to the subject, an on camera mic also helps, at least for my type of filming.

I fully agree that proper mic placement is something that everybody needs to try.

As a wildlife filmmaker, I have decided to place my Sennheiser MKH 416 directly on the C300. I am able to get many useful sounds. Last month I was filming without the Sennheiser on my C300 and the desert fox came pretty close and started calling. And worse, my assistant thought that I am filming and recording sound, so didn't even move to pick up the microphone for fear of disturbing the fox which was directly staring at us and also for fearing that the onboard sound will get ruined. And then later both of us realised that we didn't record any sound. As humans, we mess up. So for certain situations, better to have some sound rather than nothing.

On another occasion, I was waiting for a leopard. Dawn was about to break in. Light was too low. A bird (Shikra) flew and sat on the branch over my head and started calling. I just pointed my camera up and recorded the sound. So I feel while one should place the microphones as closer to the subject, an on camera mic also helps, at least for my type of filming.

That is fine as a back-up but sadly you can not change the laws of physics so a mic placed on a camera 500 yards away will still sound like a mic 500 yards away, so to start off with just a camera mic mindset will end up giving you more un-useable sound than valid decent recordings that match the pictures.

Most of the sound on the BBC wildlife doco's is post produced with separate sound recordists working independent of the camera crew to get the sound they need, slapping a mic on the camera may get your something but it is still not the best way to do it!

I completely agree with you. I prefer to get the sound recording while I am shooting, by directly recording to the camera as well as simultaneous separate sound recording. My onboard sound is helps to sync. I was talking about some rare situations which are not easily encountered by a separate sound recordist later. Ofcourse one an try FOLEY. Animals vocalise differently in different situations, and we simply don't have the understanding of that. In my case, the onboard mic serves as a record of understanding the behaviour. Most of the times when I am shooting very early in the morning or late in the evening, the ambient noise levels are pretty low and I am able to increase the gain without it sounding bad. I was just giving an alternate point of view as to why a microphone on top of a professional camera is not a sin. I am trying alternate ways of recording sound by remote placement of microphones. Some of the knowledge acquired from people here definitely helps.

I know this is a video forum, but please - can we have some proper use of language here? How many people would refer to focus as zoom, or talk about pan when they mean zoom.

The clip you are listening to has NO ECHO WHATSOEVER!

The room is live because what you can hear is reverberation. The arrival at the microphone or ear of sound that has reflected from the various surfaces, creating arrival time differences. Until you can actually hear two or more repeats, it's reverb, not echo. It may seem a small point but echo and reverb are primary phenomena in audio and not being able to tell the difference is pretty bad. Up until recently, reverb was not able to be removed or reduced by software, but now it is - to a degree and subject to a bit of uncertainty.

The physics is so simple - camera mics give great general sound in mono or stereo, but if we look at distant sources of sound it's wanted v unwanted sound, and while narrower polar patterns help, distance reduction and narrower patterns win most times!

Please guys - can we just be accurate. Shouting hello at a mountain and hearing your own voice saying hello back, is echo. What you hear in a cavern or large cathedral is reverb.

Echo is greater than 30 milliseconds.

Reverb is less than 30 milliseconds.

Echo is returned off of one or more distinct distant far away surfaces and don't necessarily create their own audio profile.

Reverb is returned from a multitude of surfaces, close by and which impinge on each other and tend to create their own audio profile.

Our Brains DO NOT and can't adjust for Echo - ie we hear distinct separation/s.

Our Brains DO adjust for reverb - ie we don't hear distinct separations. But we CAN train ourselves to hear it.

Now, all of this may seem nitpicking, but when we are dealing with echo it often wont be the same as how we deal with reverb.

The only thing, IMO, one can say is that they both contain sound energy in the form of longitudinal waves, and when they hit a relatively hard surfaces a lot of their energy is not absorbed but is reflected back. And that's about as far one can observe as a convergence of similarity. Further away from this description they are very different critters.

Grazie

I completely agree with you. I prefer to get the sound recording while I am shooting, by directly recording to the camera as well as simultaneous separate sound recording. My onboard sound is helps to sync.

Same here. Not only for syncing but for backup insurance. There's nothing like a little bit of redundancy because "Stuff" happens.

Animal sounds are quite unique and can't easily be replicated in post. If something happened when recording and nothing was captured on the camera you could be totally out of luck.

Have you ever used a parabolic reflector to aid capturing audio? Except for the bulkiness aspect, it seems this would be good for wildlife sounds. It's not like one can put a lavalier on them.

Speaking of wildlife recordings, there was a great video (not great from a videographers standpoint, but great from a wildlife standpoint) of a dolphin that requested help from a diver to remove some wire or line from a fishing net that was wrapped around it's fin. The divers in the boat recognized the dolphin was trying to tell them something and finally figured out it was a call for help.

Editorial Comment:
This was not only a call to help a lone dolphin, but a call to help a species, the animal kingdom, and actually, the whole environment.

A local salmon stream fish count has gone from ~125 ten years ago down to 0 last year. A large shopping center was built at it's headwaters and runoff from the development goes into the stream. Most people are totally clueless about the environment, and that includes the Poulsbo City Planning Commission and the Poulsbo Council that approved it. Ten years: From a Salmon stream to a Dead stream.

I know this is a video forum, but please - can we have some proper use of language here? How many people would refer to focus as zoom, or talk about pan when they mean zoom.

The clip you are listening to has NO ECHO WHATSOEVER!

The room is live because what you can hear is reverberation. The arrival at the microphone or ear of sound that has reflected from the various surfaces, creating arrival time differences. Until you can actually hear two or more repeats, it's reverb, not echo. It may seem a small point but echo and reverb are primary phenomena in audio and not being able to tell the difference is pretty bad. Up until recently, reverb was not able to be removed or reduced by software, but now it is - to a degree and subject to a bit of uncertainty.

The physics is so simple - camera mics give great general sound in mono or stereo, but if we look at distant sources of sound it's wanted v unwanted sound, and while narrower polar patterns help, distance reduction and narrower patterns win most times!

Please guys - can we just be accurate. Shouting hello at a mountain and hearing your own voice saying hello back, is echo. What you hear in a cavern or large cathedral is reverb.

Don't want to start an argument but Paul, you are incorrect. From a physics standpoint (yes I started my college life out as a general physicist), reverberation is a series of echoes closely spaced so that the brain perceives it as a continuation of the original sound source. This usually occurs when the first arrival of reflected energy less than .03 seconds (3 ms) after the initial instance. That works out to having the reflective surface approximately 16 ft (5m) from the source. Reverb is a phenomenon created by echoes.

What you are hearing the the clip are echoes that exhibit properties of reverb. Again, as you say it may seem nitpicking but if we are to be truly accurate you cannot have reverb without echoes. In terms of physics and how sound works, the properties are the same between echo and reverb. The things you have to analyze to determine the effects of them and how you'd handle them have to deal with how the energy from the sound waves interact with each other. In other words how the effects of constructive and destructive interference affect the energy reaching the ear or mic.

I suppose it is a matter of degree? How much is needed to be corrected, and by what method. Maybe it is the latter that determines just what of the 2 examples dictates the how.

Cheers

Grazie

reverberation is a series of echoes closely spaced so that the brain perceives it as a continuation of the original sound source
I would agree with that if you replace "...series of echoes" with "...series of reflections."

According to the Audio Cyclopedia (copyright 1959, 1969), which I consider to the the granddaddy of audio reference books, Section 2.79 states: "What is an echo? -- The repetition of a sound caused by reflection from a surface. To be an echo, the reflected sound must be 1/20 of a second or longer behind the original sound."

Also, the Audio Cyclopedia, in a lengthy Section 2.34, states (in part): "Reverberation is the persistence of sound within an enclosure after the original sound has ceased."

Everything seems clear so far: an echo is one distinct reflected sound delayed by at least 1/20 second; reverberation is persistence.

But then the Section 2.34 goes on to say: "Reverberation may also be considered as a series of multiple echoes, decreasing in intensity, so closely spaced in time as to merge into a single continuous sound and eventually be completely absorbed..." So even the Audio Cyclopedia appears to contradict itself! (Again, if you replace "... multiple echoes" with "... multiple reflections" then it's entirely clear and I personally agree with that definition.)

A different set of definitions comes from Audio Postproduction for Digital Video, copyright 2002 by Jay Rose. He says:

"Reverberation is the collection of thousands of random reflections that real-world spaces contribute to a sound..."

"Echo is a series of evenly spaced repeats that get softer and softer. It doesn't sound like natural reverb."

But Jay doesn't make the distinction about minimum spacing between repeats. Could they be 20ms apart? 10ms? 1ms?

Personally, I like Audio Cyclopedia's definition of echo, and Jay's definition of reverberation. And what we've been talking about in this thread are neither... they are "reflections" making the room sound very "live."

But this discussion could go on... and on... and on... and on... and on... and on.......

Personally, I look at the terms more from a physical or Physics point of view.

Echo: this would be like throwing a rock into a body of water (with no existing waves) and where the wave hits, say, a wall, and is reflected back to it's origin, that reflected wave would be an echo. On a pool/billiard table an echo would be the cue ball bouncing off something and returning back to where the cue hit it.

Reverberation: That would be when the reflected wave hits another wall but this time on the opposite side, bounces off the wall and returns. The returned wave would be the reverberation. On a pool table the cue ball would be bouncing off one bank and then another. But it would have to go past the point where it was originally (maybe more than once?) otherwise it would be just a ricochet.

Time: I don't know why there should be a time limit, whether it be seconds or what ever. If one is working in an atomic scale a wave, or particle, could travel at the speed of light. On a cosmic scale there is sunlight that bounces off the moon and goes back to the Earth, and I'm sure some percentage of it is bounced back again to the moon. Earthquakes cause tsunamis that create waves that can reverberate throughout the ocean. I don't understand why a time limit should be distinguishing factor. Surely there must be some reason behind it.

That's my 2c worth.

John, I like your echo analogy. To my mind, an echo is a discrete reflection that the brain can perceive as such. If I yell "hello" at a cliff, and the sound returns to me with enough delay that I can hear the word "hello" again, that's an echo. If I'm standing on stage in an empty auditorium, and clap my hands, and a fraction of a second later I hear my clap bouncing off the back wall, that's an echo.

And if the stage curtains are open and the clap bounces back and forth between the rear house wall, and the upstage wall and back and forth a few times, that's a special case which I've heard referred to as a "slap echo."

Anyway, I think the issue about minimum delay (20ms, 33ms, 50ms, depending on whom you ask) is because your brain needs a certain amount of delay to perceive the reflection as a separate echo, and not just coloration of the original sound. In fact, if you make the time short enough and you create a comb filter.

Now if I'm singing in the shower, I hear a lot of first-order and multiple-order reflections of my voice. But they occur so soon in time that I don't hear them as individual echoes... I just perceive that the room is very "warm" and resonant. (In fact there's some reinforcement, and probably some destructive subtraction too, at various frequencies... related to the room dimensions and wavelength of different frequencies.) I don't think I would quite call this "reverberation" because it dies out very quickly. It's just coloration of the original sound.

But now get in a big room like a masonry cathedral, with lots of hard, reflective surfaces, and say or sing something. You might, indeed, hear a distinct echo (depending on the room and your location in it), but you will also hear a countless number of first-order and multiple-order reflections as the sound bounces around from the walls, floor, ceiling, pillars, wooden pews (if any), etc. They are so many in number and so diverse in time that you can't distinguish individual reflections... they take a while to build up to their loudest, and take longer to die away completely. That is certainly reverberation.

Interestingly, though, what used to be called an "echo chamber" really produced artificial reverberation. Oh well...

It seems to me that this is a point of view issue.

From the point of view of human perception, an echo is very different than reverberation.

From a physical world point of view, an echo is simply a reflection. For instance, for a radio signal echo from a building or mountain, the 20ms threshold is meaningless.

And when describing audio situations to a potential novice, precise language is often sacrificed in order to communicate technical concepts (reverberation) using familiar terms (echo). Unfortunately, looser language usage sometimes sticks.

Back on topic, I agree with Sabyasachi. An on-camera mic can be helpful in providing a reference track to synchronize an external recording or as a basis for ADR or other audio replacement. When not intending to use in-camera audio in the final product, I turn automatic-gain-control ON. I do this as insurance so I will get a useful reference track without monitoring, adjusting gain, or otherwise distracting from shooting the image. And yes, in some documentary situations, I've ended up using this audio when things didn't go as planned.

One place where I've used on-camera mics on purpose is at tradeshows. Yes, I've done the lav thing, but it's a hassle, it takes time and effort to hide the wire, and it can be the extra hurdle that can make somebody refuse the interview. For a tradeshow video, I'd rather get the shot with so-so audio than no shot at all. The trick is to use a wide lens (28-35mm on full frame), and to get the camera close to the subject. It's not ideal, as people aren't as attractive up close, but it keeps the crowd from walking into the interview, provides a certain intimacy to the audience, captures a wide view of the tradeshow eyecandy, and gets the on-camera mic close to the talent. Yes, it delivers an imperfect image and sound recording, but this approach with a simple monopod is fast and light for a one-man crew.

With a 3-man crew, you can add a boom and have a person hold an LED light. Those keep people from accidentally walking into the scene when the camera is at a distance. Alternately, the "audio person" can do the on-camera interview with a handheld mic. One can roll the on-camera mic if desired, but from six feet or more at a tradeshow, the audio would be worthless as a backup.

If I can make a serious comment for once (apologies to Garret for my earlier frivolity - I didn't think there was anything to argue about in what he said): on the subject of echo vs. reverb, I can understand from Paul's point of view as a marker in Sound Engineering examinations that the difference between echo and reverb is made clear to students, and the fault/effect demonstrated in the camera audio in question was clearly reverb and not echo - "echo" given as an answer here would be marked wrong, end of story.

I am not sure of the precise definition used by the various Examination Boards in England and Wales, but the Scottish Qualifications Authority's Sound Engineering Glossary has this to say:

reverb – The natural series of very short and dense echoes of a sound
that occur in a confined space such as a room or a hall. While echoes
with a longer delay would be discernible, in reverb the echoes
happen so fast and are so dense, it is impossible for the listener to
hear individual repeats. Reverb is the essence of natural sound.
Listening to a close miked instrument is like having the instrument
play in your ear in a very small room. The addition of reverb to a
sound makes it appear as if the instrument is being played in a real
acoustic. Nowadays reverb can be emulated digitally very easily and
nearly all effects processors have a wide range of reverb types for
different applications. See also gated reverb.

Reverberation and echo are varieties of reflections, if we want to be pedantic. Echo is not a appropriate word to use to describe what we're talking about. A series of discrete echoes, would not be reverberation. As has been said, the usual key discriminator between echo and reverb is the 30ms rule - coupled with the fact that reverberation does not have clearly defined individual returns. We can detect what the audio people term as flutter when a number of closely spaced (in time) returns arrive at the ear. It's still not reverb.

I'm only interested in correct identification, and a live room, such as the one we hear in the clip contains no content identifiable as echo, as all returns are too close. All they do is colour the sound.

If I can make a serious comment for once (apologies to Garret for my earlier frivolity - I didn't think there was anything to argue about in what he said):

No need to apologies Colin. I actually found it quite amusing and would have made additional comments about the speaker's hair but I didn't want to seem irreverent about a Stanford University professor.

I'm only interested in correct identification, and a live room, such as the one we hear in the clip contains no content identifiable as echo, as all returns are too close. All they do is colour the sound.

My apologies to all regarding the use of "echo". While those who are audio specialists would not classify what they heard in the clip as echos but instead as reverberation, those of us who have studied physics and had years of energy and wave theory crammed into our somewhat limited brains would tell you that you are completely insane (please note the sarcasm in my use of words to keep the discussion light). In the world of physics you cannot, yes, absolutely cannot, have reverberation without the presence of echoes. As Jon had noted, in physics, time has no bearing on an echo or the definition of one. It is simply the reflection of a wave. In fact, there is no requirement that an echo needs to be within the audible range. It is simply the reflection of energy, in the form of a wave, off of a surface. The energy contained in the reflection must be equal to or less than that of the incidental wave. All of this discussion has caused me to go back and once again, spend needless hours looking through my old physics texts which has caused me to come to the same conclusion I did 20 some odd years ago, there are some very smart people figuring out things that have no real bearing on the majority of those living in the real world. And, as one of my old materials science professors use to say, "It just doesn't matter" (said with a very distinct English accent).

Sorry for the sidetrack. I am the easily distracted type. So, to get back to the original discussion.... and to respond to Sabyasachi's post, I agree that there are situations where getting a mic close to the source is not possible. In those instances you make do with what you can. The video was meant as a response to one of the questions I get asked a lot, "how can I improve the sound on my videos?" I usually discover that they are using a mic placed on top of their camera or the internal mic on their camera. My response is usually, "get the mic closer to the speaker."

It's OK, I understand. However - none of my gizmos, and I have a lot, has a knob labelled 'echo' - we have reverbs, delays and modulation.

Coincidences abound - Myself, like Colin have been involved with education. I was Principal Examiner for Music Technology when it first arrived on the scene in the UK - so use of precise language is pretty important.

Probably things are different in the US, where language does get 'adjusted'. The misuse is pretty widespread, we even had a few proper echo units years ago, didn't we - Roland Space Echo and Watkins Copycat, with tape and drum echo, and then spring reverbs and even the EMT plate echoes, which were not echo.

Here in the UK, in electronic engineering and physics we've been using reflections since the radar days, although just to muddy the waters, people often called the blip on the screen the echo return - but I guess I can see this one.

I'll stick with reverb.

For what it's worth, the new removal plugins are actually very good, and although people generally use them on music, they might be interesting to give a boost to video work where the room is tricky?

Hi John,
I am using a Telinga Parabolic microphone to record distant sounds. It is bulky and carrying it in the field becomes a challenge along with all the other equipment. I find that placing the parabolic microphone in a tripod gives better results else there is some handling noise. The other challenge is it catches dust due to its shape.

I am also recording directly to my camera through the Sennheiser MKH 416. I am also trying to remotely place microphones as well as take the microphone as close to the animal as possible, place it and then back off. I am getting the best sounds early in the morning or late in the evening and in night when human activity is minimal and the ambient noise is less. During those times, I am also able to increase the gain without any noticeable problem.

The dolphin story in incredibly touching. Thanks for sharing.

Cheers,
Sabyasachi

Same here. Not only for syncing but for backup insurance. There's nothing like a little bit of redundancy because "Stuff" happens.

Animal sounds are quite unique and can't easily be replicated in post. If something happened when recording and nothing was captured on the camera you could be totally out of luck.

Have you ever used a parabolic reflector to aid capturing audio? Except for the bulkiness aspect, it seems this would be good for wildlife sounds. It's not like one can put a lavalier on them.

Speaking of wildlife recordings, there was a great video (not great from a videographers standpoint, but great from a wildlife standpoint) of a dolphin that requested help from a diver to remove some wire or line from a fishing net that was wrapped around it's fin. The divers in the boat recognized the dolphin was trying to tell them something and finally figured out it was a call for help.

Editorial Comment:
This was not only a call to help a lone dolphin, but a call to help a species, the animal kingdom, and actually, the whole environment.

A local salmon stream fish count has gone from ~125 ten years ago down to 0 last year. A large shopping center was built at it's headwaters and runoff from the development goes into the stream. Most people are totally clueless about the environment, and that includes the Poulsbo City Planning Commission and the Poulsbo Council that approved it. Ten years: From a Salmon stream to a Dead stream.

I'll stick with reverb.

My spell checkers don't recognize the term, "reverb". Not sure why. The term has been printed on guitar amps made since the '50s, so it's not exactly the latest slang. "Reverberation", and "echo" are both recognized, but until this thread, I don't know that I'd ever spelled out the word "reverberation".