Sound

[higginsdj]

Hi all,

 

Just forked out a small fortune (for me) on some digitial sound recording gear. (Zoom H4n and Rode M3 mike)

 

I had planned on just using my iTouch to digitally record dialogue (seemed OK in my tests) but wouldn't be much chop for Foley/SFX. Manufacturer MRP/RRP seemed rather exhorbidant. They wanted to charge $900 for the Recorder and $250 for the mike. Well, as luck would have it, one is not expected to pay MRP/RRP for these items in the music industry and I left the store having paid less than half price - made me feel better anyway

 

Well, tests show that it's like chalk and cheese when comparing the above to my iTouch. Now I just have to learn all the features etc to ensure I can record at the right level (without invoking auto levels) - oh and I found out what a deadcat and deadkitten are - I need those for outdoors recording....... next on my list of things to buy or make!

 

Now the Zoom lets me record at up to and including 24bit/96mhz but unfortunately that .wav file was not compatable with my old version of After Affects As luck would have it, however, I purchased the Screen capture application ScreenFlow for my iMac and that will read these audio files. It turns out that ScreenFlow will allow me to do all my basic editing (joining sound to video) and it will playback in real time - woohoo

 

Anyway, I digress...... What I am after are some pointers on recording dialogue (I have 'talent' who have volunteered to voice my short) and foley/SFX. Links and any advice would be most welcome. I have joined a few forums but they seem to ignore the newbie since I am not a professional sound engineer!!! Fortunately we don't have snobs here

 

Cheers

[robcat2075]

You probably already know to record it as dry and without room ambience as possible. It's real tough to cut other wise.

 

But on recording voiceovers, if they are not clever actors I've found it useful to take a highlighter and mark the words to hit harder and put a slash where they should pause. It's a way to force some nuance into what would be a flat reading otherwise. If they are real actors you might not do this and just see what they come up with.

 

Another thing to do is get the same line read with as many variations as possible, and you give your self some options for later.

[higginsdj]

I chose my 'voice talent' (collegues at work) based on their actual character and voice matching my characters (not that I tell them that) so I am hoping that I can simply describe a work based situation that would put them in the 'state' I need to get the right level of 'emotion' out of them in the dialogue - but I like the highlighting idea.

 

One thing I want to experiment with is ambience subtraction. In my astronomy work I take a picture of an object then take a 'dark' (shutter closed). I then subtract one from the other to get a clean image (ie the Dark is the noise image). Sound is just another wavelength so I can't see why this wouldn't work (ie after the talent is gone, just record the ambient noise - it's just the faint drone of A/C and PC fans).

 

Cheers

[robcat2075]

Sound is just another wavelength so I can't see why this wouldn't work (ie after the talent is gone, just record the ambient noise - it's just the faint drone of A/C and PC fans).

 

That's basically how noise removing filters work. Some are great at it.

[Shelton]

David

 

I have been on the other side of the mike in which I was the voice talent. I have read in several different settings. The best setup for me was no ambiance noise at all and several places had a window to the director. I would blanket that as it allowed me to create the world I was reading. Comfort is the name of the game. i have also worked with two dvd productions and was in the sound studio. I can say I learned alot by being on both sides of the mike

 

Steve

[williamgaylord]

If the background noise is low enough to be masked by dialog, you can edit it out between dialog elements. A bit of work, but can be done. However, if the background noise is perceptible in the quieter elements of dialog sound, then you are out of luck.

 

Don't forget about room acoustics, though. Even if your background noise is non-existant, the resonant characteristics of small rooms can still color the sound of dialog through flutter echo and comb filtering. (This is why a bathroom or closet makes a poor sound booth.) However, this is easy enough to deal with if the room is a normal size (say 13x15): use a pop shield and keep your microphone close to the person's mouth--within a foot or so--they don't have to eat it. The M3 is a small diaphram condenser mic, but it has plenty of dynamic range. (I love Rode microphones!)

 

There are some pretty decent noise cancelling microphones/headsets at fairly reasonable prices, though. They use a pair of microphones. Up close to the pair of microphones the sound field does not cancel (much), but sound originating farther away cancels quite effectively.

 

The H4 is a really nice recorder! I use an M-Audio Microtrack 24/96, but if it had been available at the time I would have opted for the Zoom H4. It's more rugged, and has a quieter front end.

 

Check out Hal Leonard's "Guerrilla Home Recording: How to Get Great Sound from Any Studio (No Matter How Weird or Cheap Your Gear Is)". This is an excellent book on practical recording technique. Well worth the $14 or so if bought new.

[higginsdj]

I'm planning on doing the recording at work. The third floor is open plan 50metres x 20 metres with a raked ceiling to 5metres. Lots of chest high cubicles and some open areas in between. The floors are raised tiles (all the PC cabling is underfloor) and drop tile ceiling. I'm hoping that the acoustics is nearly "outdoor like" (ie no echo).

 

A recording studio would be great but I will see how things go with my trials first. If my talent knew what they were doign I would agree with you but I think I will need to do a bit of 'coaching'. I'm wondering, as I have 2 characters, if I should record both at the same time playing off each other or just do one at a time?

 

Thanks for the book recommnedation. Is the author Hal Leonard? Amazon show Karl Coryat...

 

Cheers

[HomeSlice]

I'm wondering, as I have 2 characters, if I should record both at the same time playing off each other or just do one at a time?

I've found that unless the people are very disciplined, then they will often cut into each other's dialog, making it very difficult to get a clean track for each person. Maybe you can have both read at the same time for the first couple of read-throughs, so they can get the spirit of the interaction. Then have each one read his/her dialog alone while you are recording.

[Darkwing]

the problem I've found with the one on one however is that unless you're dealing with pros, they will have a tendency to not sound like they're in conversation, but simply reading or reciting as they can't seem to stay in character without the other one present

[higginsdj]

Sounds like the best bet will be to play it by ear. Have both in attendance, do a few readings together then see how they interact. I wonder if the directionality of the mike will allow me to just record one but have both reading at the same time.... Something else to test - it is a 1 to 4 channel recorder and it was designed for bands (4 separate mikes on different instruments)

[HomeSlice]

I wonder if the directionality of the mike will allow me to just record one but have both reading at the same time.... Something else to test - it is a 1 to 4 channel recorder and it was designed for bands (4 separate mikes on different instruments)

Not an audio expert ... but it was always my impression that musicians in a studio wear headsets so they can hear what's going on, and the instruments are plugged straight into a mixer? Not sure how they handle drums though ...

[robcat2075]

Here's something. Don't tell them them they are done after you record them the first time. Make sure they know you will need them again at a later unspecified date to fix lines or record changes.

 

A problem with TWO was people who wouldn't redo a line or record a change..

[higginsdj]

Yeah, I've made them aware that there will be more than 1 sitting and I will be recording everything (including DV Camera). The 16gb SIM will let me record 7hrs of 24/96 audio

 

For the bands, looking at the configuration diagrams for 4 channel recording it shows up to 2 channels plugged into instruments (or external mikes or combination thereoff, the other 2 channels from the built in condenser mikes. I'll have to experiment.

[williamgaylord]

Thanks for the book recommnedation. Is the author Hal Leonard? Amazon show Karl Coryat...

 

Oops...yes it is Karl Coryat. (Hal Leonard--a publishing company named after Harold "Hal" Edstrom and his brother Everett "Leonard" Edstrom--if you want to learn to play guitar, piano, etc...not sure how I got the two mixed up...how embarrassing!)

[williamgaylord]

Here are some binaural recordings I made with a dummy head binaural microphone I made (listen to these with stereo headphones or at least earbuds):

 

Binaural Recordings

 

The "Dogz" recording was made with a Sony Minidisc recorder.

 

The "Train" recording was made using an Akai DPS-16 multitrack recorder. Sadly this workhorse is on it's last legs! (Had the best pre-amp stage of any portable multitrack.)

 

The rest were recorded using an M-Audio Microtrack 24/96. Your new Zoom H4 will likely outperform this, though it might be hard to tell.

[phatso]

If you've got only one mic, you can't place it optimally for two actors. You'll have to record them separately.

 

Radio announcers place the mic above the mouth and just a little to the side, aiming down, so if the actor leans forward the mic will hit the root of the eye tooth. That gets it as close as possible without being in the path of breath noises, and captures sound from the nose as well as the mouth, which is critical. If you just aim for the mouth alone you get stopped up sound.

 

Can't you set the software with a threshold to automatically gate out room noise between sentences? In an open plan office I think you're going to have trouble.

[williamgaylord]

It's amazing how much noise you don't notice until your hear it through the microphone. Human binaural hearing allows you to ignore an amazing amount of sound due to your perception of the direction of sound--the cocktail party effect. Listen to your recording environment through the microphone.

 

If you record each voice actor individually with the mic up close to their mouth so that the level of speech is well above your background noise it might be fine. You can use a noise gate to null out the noise or simply edit it out. (I prefer to edit out the noise, since a noise gate takes a finite time to turn of and on. Editing is more work, but it can be more accurate.) The key is to have the background noise low enough that the quietest speach elements can still mask it.

 

If you get your mouth too close to the microphone you get into the "near field" of the microphone which can also color the sound. However, in guerrilla recording that is usually much less of a problem compared to background noise.

[robcat2075]

If you gate out the background noise you won't have something for a noise reduction filter to sample.

[williamgaylord]

If you use a noise cancelling headset or microphone, the cancellation is accomplished by the proximity of two microphones, or two (or more) ports and one diaphram. Sound originating close to the front of the microphone will create a significant sound pressure difference, so there is a net pickup. Sound from farther away will produce much less of a difference. That is how it reduces background noise.

 

A noise gate is later in the chain, so it won't affect the noise cancelling microphone performance.

 

If it's not a live performance, like a radio broadcast, I prefer not to use automated audio processors like "duckers" or noise gates, since they take a finite amount of time to react and can affect the leading and trailing edge of what you are trying to record. I prefer to record it "dry" and edit it.

[higginsdj]

OK, just had my first recording session and it seemed fine. One Actor at a time, mike set up per Phatso's advice. Recorded background by itself but nothing registered above -48db. In Audacity it was a flat line - I had to wind the gain up +30db just to get a low frequency sound above -24db.

 

The actors voices seem 'clean' (to my ears at least) but seem to have a strong bass bias to them. As phatos suggested in one of the other sound threads I should be able to "fix" this with EQ.

 

I'm interested to know why the Open Plan room might be a problem.

 

Cheers

[higginsdj]

For those interested in Voice/Dialogue, recording methods, equipment, home based etc and production techniques then here is a great e-book on the subject:

 

Voice Actors guide to recording at home and on the road

 

I registered to get access to the download.

 

Cheers

[williamgaylord]

Sounds like you have a decent situation for recording your voice actors. I don't think and open plan office space is likely to be much of a problem as long as it's off hours so there is no other activity going on to create background noise to pick up. If your office floor plan is wide and you record in an open area, with no walls right near by, your main "early" reflections will be off the floor and ceiling. If you give them a comfortable stool to sit on so the mic is at about standing height (or about mid way from ceiling to floor) it should sound pretty decent if your voice actors are speaking within a foot of the mic.

 

In sound recording and mixing, you have two types of dynamic range. One is the dynamic range of your equipment. I usually call this "processing dynamic range". The other is the range of sound level of the recorded or mixed sound, from the quietest element to the loudest element of the sound content. This I usually call the "program dynamic range".

 

Your "program" dynamic range in recording will most likely be constrained by the clipping level of you recorder and the background noise of your environment. I guarantee your equipment is not likely to be the limiting factor for the low end, unless you are picking up hum on your mic cable or you are using a cheap mic with a relatively high electronic noise level.

 

Your recording level meters should tell you if you are approaching clipping. In digital recording, clipping is more severe than in analog recording--it's an absolute limit--so make sure you have enough headroom to cover the loudest sound you want to record. With a digital recorder you have an huge amount of processing dynamic range, so assuming you have a quite microphone, you can give yourself plenty of headroom. 16-bit recording gives you 96dB of dynamic range. At 24 bits you have 144dB of dynamic range. The pre-amp stage of your recorder is probably more restricted than that though--I can't find specs on the pre-amp of the Zoom H4. Your Rode M3 has a self noise level of 20dBA and a sensitivity that gives it a 73dB signal to noise ratio, which is pretty good. So I think background noise is your main limiting factor and that scales with your input gain so, go ahead and give yourself as much as 24dB of headroom.

 

BTW, level meters are designed for detecting overload conditions and setting gain, but are very poor at measuring relative "loudness". "Loudness" is a measure related to subjective experience, and is actually difficult to measure objectively. You deal with "loudness" when you get to the point of mixing your sound, where you have to balance the relative loudness of all your sound elements. This is best done by ear, especially when judging the loudness of speach. In the case of speech, most people will agree to withing 1dB, which is really amazing! In the case of non-speech sounds, people will at times disagree by as much as 12dB, which is a little more than a factor or 2 in terms of the perception of loudness. So...use level meters to avoid overload conditions, but don't use them to try to judge "loudness". Use your ear to judge loudness. And when you mix your sound, start with the dialog and get that aligned, then mix your other sound in once the dialog has been adjusted.

 

So how do you set dialog to a target level for production? If you can't afford a Dolby LM100 loudness meter, you can use a reference sample of recorded dialog that is at a known level. Since I have access to an LM100, I can make you a reference recording.

 

In the digital world we use "dBFS" which is the level relative to a full scale sine wave (a sine wave just shy of clipping). "Loudness" measurement is done per a new ITU-R standard, BS-1770, which I won't bother to describe other than to say it processes the sound to approximate how the human hearing perception responds to sound. It uses a unit of measure "dB LKFS" which can easily be related to dBFS when you mix your sound. In your case, I can make a reference dialog recording at a standard loudness level which you can use as a "reference by ear". You would load that into your sound editor and use it as a comparison for setting the level of your recorded dialog by ear. Once you have your dialog level adjusted properly throughout your program, you can mix in your other sound elements "by ear" and "to taste". Dialog should be kept to a fairly consistent loudness level. That goes for "whispers" as well--they should be "stage whispers". Trust your judgment--if you can clearly follow the conversation without having to adjust the volume, then the dialog level is pretty consistent. Speech has to be fairly consistent in level so that the conversation is intelligible. It's variations in speech level that most often causes people to reach for the volume control. With other sounds, you have more freedom for loudness variation, especially for artistic effect.

 

And remember what Louie Armstrong said: "If it sounds good, it is good!"

 

Hope this helps.

[williamgaylord]

A bit of trivia: A healthy human being has a hearing range that has a 1,000,000,000,000:1 ratio from sound at the threshold of pain to the smallest perceivable sound vibration. The smallest sound vibrations that can be perceived are about the same size as the diameter of a hydrogen atom!

[higginsdj]

I did a lot of 'cleaning' up and experimentation over the weekend and ended up with some good sound. Now all I need to do is get my actors to improve their pronunciation of words

 

The floor was quiet (only a couple of people) BUT a phone range and the cleaners turned up. BUT none of that was audable on the recording - not a hint. I love the Cardoid pattern of the M3 There seems to be an exponential drop off in sound pickup the further anything is from the mike. (Levels were set with the voice ~ 6" from the mike).

 

Sound recording is certainly it's own science.....

 

Cheers

[williamgaylord]

Actually it is exponential! Even for an omnidirectional mic. Think of a sphere expanding where the radius double each time. The area is four times the radius squared times Pi. Every time you multiply the radius by a factor or 2 the area increases by 2x2=4. So doubling the radius quadruples the area of the sphere. The intensity of the sound decreases as it spreads over a larger area, so it drop rapidly near the source and then ever more gradually as the distance from the source increases. So the sound is a factor of 4 (6dB) less intense at 1 inch than it was at 1/2 inch. At 2 inches it is another factor of 4 less intense. It has to expand another 2 inches for the next drop by a factor of 4, and then 4 inches and then 8 inches and then 16 inches. So already from 1/2 inch to 16 inches we've dropped by 30dB! At 32 inches 36dB. At 64 inches 42dB. And so on.

 

Reflected sound (of walls, tables, floors, ceilings, people, etc.) spoils this progression somewhat, but the basic sharp drop-off near the microphone smoothly changing to a very gradual drop of (at an already low level) is still a practical way to understand how the sound varies with distance.

 

The directional microphone still behaves according to this principle, but it also adds a dependance on the direction of the sensitivity of the microphone. So if a noise is inside the "beam" of the microphone, than distance is still your friend.

[higginsdj]

Ahhh the beauty is, if there is nothing closeby to reflect the sound then the reflections off distance ceiling and walls will be too faint to be audible So with walls 10m away and a raked ceiling at 5m overhead and 120 degree cubicle partitions (do they absorb sound or they just there for show!)

 

I think my open plan office appears to offer something quite similar to open air recording and the results after basic cleanup (Noise Removal, Normalise and Equalisation) seem to be fine.

 

Cheers

[williamgaylord]

If your cubicles are made with pressed fiberglass sound panels, they can be pretty good at absorbing mid to high frequency sound. If it has a fabric cover, and has some "give" to it, but not like foam, it is probably pressed fiberglass. If it makes a slight "crispy-scratchy" sound when you press it, that's another giveaway. ;-)

 

How many square feet of relatively open floor space do you think there is in your " recording studio". And how high are the ceilings?

[higginsdj]

4,305 square feet (400 square metres) with a 15 to 30 foot raked ceiling (15 foot at one wall leading up to 30 feet at the far wall). Concrete floor with 12" gap to a raised tiled and carpeted floor (all the power and IT cabling goes in that 12" gap rather than through the ceiling. Drop tile ceiling (metal roof)

[phatso]

The bass-heavy sound is no surprise, directional mics boost the bass when you get close to them. It's called "proximity effect." This can be an advantage; you can EQ the bass down to get it right and at the same time reduce any rumble in the background.

 

Hig - I spoze Armstrong would have agreed, but it was Ellington who said that. Not to nitpick or anything.

[williamgaylord]

Oops! My bad! It was indeed a quote from Duke Ellington.

 

Yes, directional microphones will give bass a relative boost close up.

 

If you already know this, or just don't care, ignore the following:

 

The directional microphone responds to differences in sound pressure between the front side of the microphone diaphragm and the back side of it (accessed through holes or slots behind the diaphragm). There are two ways this happens. The directivity of the microphone is based on the "phase difference" between sound hitting the front of the microphone, and the same sound arriving slightly later via the ports at the back of the diaphragm. The "phase difference" is how the peaks and troughs of the waveforms line up. As the phase diff increases the peaks and troughs start to cancel each other. The smaller the wavelength (higher the frequency) the more pronounced this effect will be as the wavelength gets closer in size to the size of the microphone. So the mic will tend to be more directional as the frequency increases (and the pattern will change), and the sensitivity of the microphone will also vary in frequency because of phase related effects as well. The location of the "ports" and the orientation of the diaphram affect how the "phase" varies depending on the direction of the sound--hence the directionality.

 

The other effect is what gives you the "proximity" effect, and that is the dependency of the sound intensity with distance I described earler--the way the sound drops of by a factor of 4 (approximately) with each doubling of the distance from the sound source (the "Square Law effect"). If you get close enough to the mic, there is enough of an intesity drop over a short distance that the back side of the microphone diaphragm gets a lower intesity than the front side. Since this effect depends only on the sound intensity variation with distance, this affect is relatively frequency independent. ("Noise canceling" microphones exploit this proximity effect to reduce background noise.)

 

The phase relationship that depends on wavelength and direction of incidence causes the mic response to drop off at low frequencies. Since the effect of distance from the source is not frequency dependent, its flat response "peeks" out from under the phase related response at the low end giving the bass a relative boost (compared to the drop-off)--IF your source is close enough. If there is already electronic compensation for the drop off with frequency, then the proximity effect exagerates the bass, instead of just reducing the roll-off.

 

Probably way more information than you want to hear, so feel free to say so. I'm just fascinated by how this stuff works, but not everybody shares that fascination.

[higginsdj]

Yep - I have been playing on that effect, and some pre-amp settings and effects inside the Zoom recorder to get that 'traditional' deep sounding voice over effect. I did a couple of requests on thefreesoundproject and now have requests coming in to do more voiceovers. I always thought my voice sounded terrible - maybe not to everyone

 

On another note. Not liking the Guerilla Home Recording book. Can't see much in there for voice/dialogue work - but maybe I'm just not getting it! I do like Rc Viers "The Sound Effects Bible" - thats going to be my bible for the audio on my movie project.

 

Cheers

[williamgaylord]

On another note. Not liking the Guerilla Home Recording book. Can't see much in there for voice/dialogue work - but maybe I'm just not getting it! I do like Rc Viers "The Sound Effects Bible" - thats going to be my bible for the audio on my movie project.

 

It is focused on music recording, which is actually more difficult in some respects than voice recording, and easier in other respects. The main advantage of voice recording is that your voice is localized basically to your head so you can really take a lot more advantage of proximity to the microphone, so you can get pretty good voice sound in relatively poor acoustic environments. In music recording, even if you only have one instrument, it typically radiates sound from a fairly wide area, especially if it is a large instrument like a grand piano or a drum set. It means the "sweet" spot is usually far enough away from the instrument that room acoustics are an unavoidable factor. If you have a whole band of instruments playing together, it is really tough to minimize the effects of room acoustics, and you usually need to mix multiple mics together, which gets into even more complicated issues.

 

I should check out "The Sound Effects Bible"!

 

There are some really good articles, FAQs, and videos on room acoustics (again, more focused on music recording) on the RealTraps web site: RealTraps Website

 

The "Acoustic Basics" is a great intoduction to the essentials of room acoustics. In voice recording you do have the advantage of being able to get up close to the microphone, but understanding room acoustics still helps. Check out the demos of the "Portable Vocal Booth": PVB

[higginsdj]

Yep, I've already checked them out - and the Portable Voice Booth. I found some instructions on creating ones own using 2" pyramid acoustic foam and a 12" x 12" x 12" collapsible box. The easiest is just hanging a doona/comforter up in an arc directly behind the mike and talk into it. It does a good job at deadening echo/reverberations in my study and the mike can be oriented to put the computers behind the mikes 9 o'clock so there is no background noise (well nothing discernible)

 

Given how important sound is in animation I'm surprised that the forum doesn't have it's own sound section!

 

Cheers

[williamgaylord]

Given how important sound is in animation I'm surprised that the forum doesn't have it's own sound section!

 

Yeah! There should be one!

[Rodney]

You mean like this one?

 

Sounds Forum

 

Its mostly for collecting links and sound bites but its what we make it.

I'll move this topic over there after once everyone has time to find and read this in the Off Topic.

[higginsdj]

Yeah, I was thinking more a discussion forum, advice, how-to's etc..

 

Cheers