Sound off about sound and audio

[Cail Young]

Edit: I would expect that there is sound processing software that can remove regular hums like this, although I'm not sufficiently familiar with what's out there to know. If the software can predict the hum, it can simply subtract it from the signal. Anyone know if there is software that works this way?

Any parametric EQ worth its salt (i.e. with a sufficiently low bandwidth) will be able to kill a 50 or 60Hz hum; it's the harmonics that can screw you.

 

[mdo]

Any parametric EQ worth its salt (i.e. with a sufficiently low bandwidth) will be able to kill a 50 or 60Hz hum; it's the harmonics that can screw you.

Definitely the harmonics are the problem, which is to say the non-sinusoidal aspects of the wave. The phase inversion technique fully addresses those.

My question is whether there is software that is capable of performing a Fourier analysis or in some other way calculating the mathematical characteristics of the hum and subtracting that from the signal. If that can be done, it would provide the most surgical removal possible.

Simple EQ has never worked for me, although I have only attempted it on tracks that have had music in the same frequency range. I guess using an EQ on a track that only has voice would probably have better results than I was getting on music tracks.

 

[Nick Shaw]

Soundtrack Pro does a pretty good job of noise removal by analysing the noise. Don't know if it does it by Fourier analysis or not, but that would make sense.

 

[Cail Young]

My question is whether there is software that is capable of performing a Fourier analysis or in some other way calculating the mathematical characteristics of the hum and subtracting that from the signal. If that can be done, it would provide the most surgical removal possible.

As mentioned above, Soundtrack Pro has an analysis tool for hum, and also has specific 50/60hz hum removal filters with selectable harmonics. So, the software's out there...

 

[Jochen Schmidt-Hambrock]

BTW, somewhere in this thread somebody said like "Given enough time, you can salvage about any audio" or so.

This is not true. Its usually garbage in - garbage out even for the best of us.

Fatal mistakes are:
ununderstandable dialog by actors mumbling, too much athmo, mike too far away, too much broadband ambient noise.
Wrong audio level (digital clipping, analog distortion, overenthusiastic compression an / or limiting, audio level so low that the signal is buried in the noisefloor.

Too many different ambiences, not enough tracks to choose from, not enough time....

Oh, damn. Sorry, I have to leave.....

Bye,
Jochen

 

[number6]

Just was wondering if those loop-based compositions are clean enough to use in low-budget feature work?

 

[Bruce Allen]

Just was wondering if those loop-based compositions are clean enough to use in low-budget feature work?

The recordings are very clean, of course. But they are soulless.

It's like using clipart in Illustrator to make an animated film. I'm sure if you used it in an interesting way it might work, but if you try to imitate a traditional film score using loops, it might come off as cheesy.

Bruce Allen
www.boacinema.com

 

[number6]

The recordings are very clean, of course. But they are soulless.

It's like using clipart in Illustrator to make an animated film. I'm sure if you used it in an interesting way it might work, but if you try to imitate a traditional film score using loops, it might come off as cheesy.

Bruce Allen
www.boacinema.com

Yeah, I think you are probably right. But it seems so much of what is being produced in today's music industry is so synthetic sounding, I'm wondering if audiences aren't accustomed to that?

 

[Zakaree Sandberg]

#6,
Yes you can fix certain things in bad sound.. hum.. a little wind..a crackle here and there.. hand slip on the boom...but, I have also spent 8 hours trying to fix a 20sec piece of dialog that had crazyyyyy stuff goin on.. (I boomed the wrong way in a certain location.. boomed from the top instead of bottom) I had tons of car noise.. wind.. and just bad sound in general. I spent hours eq'ing, de-noising, and using alll my tools in the waves diamond bundle as well as in some ozone programs...and sometimes it just doesnt work. Ended up doing adr at this point I have no idea what the original question was.. but i guess the moral of my story is.. capture good sound on set!

 

[number6]

#6,
at this point I have no idea what the original question was.. but i guess the moral of my story is.. capture good sound on set!

The original question was "Sound?", so you are on topic.

 

[Jochen Schmidt-Hambrock]

Just was wondering if those loop-based compositions are clean enough to use in low-budget feature work?

Today all audio has so called high production value. The problem is the flexibility of the material. Suppose at 10.04.08.21 the mood changes 180 degree. Can you do that with loops without your score sounding like somebody just switched stations?

Its like painting by numbers. Theoretically you could rembrandt your way to a fine painting. Might take a lot of numbers though.
But something like a "Mondrian on a very bad day" is within reach. And this is what I hear on a lot of tv soundtracks today.

Jochen

 

[Adam C Lubkin]

And hey Zakaree, what mic setup would you recommend for recording him blowing up his house? I know he could do library sounds, but if he is really crazy enough to do it (and we've pleaded with him not to...), he should at least mic it right, eh?

Maybe you could lend him your Schoeps? hehe

 

[Adam C Lubkin]

Lav mics:
a) don't capture the space that the actors are in - everything feels too dead and dimensionless
b) are capturing audio from a weird place - like, a dude's chest. How many times do you listen to people from the perspective of their chests?
c) are expensive relative to their overall sound quality

EDIT: oh yes, and don't forget the problems they cause with clothing rustles. It's also hard to shoot nude scenes with 'em. Basically whether your actors have their clothes on or off, they suck. The only time I'd use them would be if I were specifically trying to do something in the Altman style. Altman and his sound team are the only ones who seem to make them work. Any other good lav sound examples you guys know?

Until then, I'll use a hypercardiod like a Schoeps MK41 for controlled indoor environments, a Sanken shotgun for more problematic ones, and a nice shotgun for outdoors, please.

Bruce Allen
www.boacinema.com

From what I've read, Altman used a boom whenever possible along with the lavs. Plus the overlapping layers of audio help fill out and disguise the lav sound.

 

[Bruce Allen]

From what I've read, Altman used a boom whenever possible along with the lavs. Plus the overlapping layers of audio help fill out and disguise the lav sound.

Ahh! The secret is revealed! Thank you! Makes sense...

Maybe you could lend him your Schoeps? hehe

Awesome... Now THAT'd a good way to settle that whole Schoeps fragility issue once and for all

Bruce Allen
www.boacinema.com

 

[Adam C Lubkin]

I do recall Jeff Wexler (jwsound.net) mentioning that he used octavas in Mission Impossible 3 during scenes with explosions so as not to damage his Schoeps. BTW, he is one of the minority who prefer to use the mk41 even outdoors.

 

[Cail Young]

I have a sub-question. Is the highly annoying practice of camera dept (usually the DP) telling sound that rolling sync won't be necessary for given setups as prevalent in the professional world as it seems to be in my world of student and low budget? I'm more than happy for a 1st AD or the director or the recordist to make that call, but camera?!

 

[Joel Kaye]

.

Oh, and joelnet, before doing all of the above steps, try editing your audio as much as possible. Not only macro-scale editing (you were asking about this when talking about cutting out the silent portions in foley tracks?), but micro

Thanks Josch, Bruce, MDO - those are some great tips. Yeah Bruce, I'd like to take it as far as I can at home so I understand the process. The micro waveform editing to patch up syllables is a really good idea.

So for dialogue before the mastering pass -
1) Strip out dead zone and apply noise reduction techniques
2) apply multiband compression - (maximize loudness here?)
3) EQ to suit taste
4) Add reverb or other effects if desired.

Is that right?

A friend of mine has a full blown recording studio loaded with good stuff. He's pretty much an audio perfectionist so I think I'll have him give it the final master with fresh ears and try to learn some stuff from him.

I'm going to give MDO's hum removal thing a try too. I do have some hum plugins - but I'm curious to see if the results of that technique are any different.

 

[Bruce Allen]

Glad to help. Although I should say I'm not an expert in this - I've just done a few mixes and this is what I've learned from doing them.

I think you are getting more on track, although once again I'll say it's best to become expert at step 0: learning to tightly edit, to organize your tracks nicely, etc. It makes mixing so much easier.

Also, the steps 2-4 that you described MUST be done non-destructively. The mixer may well find that your eq / compression / etc. just doesn't work in the mix. Similarly, if you are going to strip out the noise, you want to do it "just right" and make sure the original is still accessible in case the re-recording mixer wants to go back to it.

Even something as "simple" as those volume rubberbands that we draw will often be deleted by the mixer - the mixer probably will want to ride the faders manually. You get a more organic feel and a good mixer can essentially do compression manually with clever fader twiddling.

How complex is your mix, by the way? We should also talk about stems, etc here.

By the way, I think you should watch your friend mix a bit in order to get an idea of the practice.

Also, technically, what I'm describing is the job of a re-recording mixer - it is not mastering (where sometimes a final sweetening stage, etc is done...)

Bruce Allen
www.boacinema.com

 

[Jochen Schmidt-Hambrock]

2) apply multiband compression - (maximize loudness here?)

Oh, be extra carefull. I,ve watched a short where the dubber was a music mixer.
Compressed / limited it like a HipHop record.

That film was unwatchable because you literally could not stand the sound after 5 minutes.

With jingles and commercials its obviously ok to compress untill your ears bleed.

Jochen

 

[vanguy]

I've just finished working on a doc where some of the sound was really bad, but irreplaceable (the person had passed on to a better world), so it MUST BE RESCUED. Probably the worst was sound recorded with the camera mic which was beside an idling car, and the radiator fan was ramping in and out every few seconds.

One of the first problems was hum, and line (60 Hz) hum is really easy to remove. I noticed that the hum reduction filters with the harmonics settings enabled affected the dialogue, so I used stacked parametric EQ's instead, set to 60, 120, 180, 240, et cetera, and as high a Q as possible. Worked quite well.

It was flourescent hum that was a problem, since the frequency seemed to be different. I got a piece of freeware called ToneGen to find the frequency of the offending noise (this also helps with things like fans). Once you find that, you can remove the hum or buzz.

I also found that you can roll off low frequency at about 170 Hz for most speech, to help kill wind noise and traffic. A little higher for female speech. But don't do this unless you're already in trouble. Clean is better.

Once you've done the hum and LF reduction, Soundtrack Pro's noise reduction works a lot better. Just use it for broadband noise like traffic or white noise, not narrowband noise, like hum.

Finally, I tried noise-gating, and then mixing back a little of the noise to make it sound normal.

With the radiator fan, I actually automated the frequency of a notch filter to try and match the speed change of the fan. But it made the dialogue sound kinda flanged.

As for music, I use loops as the skeleton (bass and drums), then add some melody to make it unique from a small keyboard. I do a sketch in Garage Band (which now has a video track) and port it out to Logic to finish. But even though I'm a fairly experienced musician, I still call in a composer for finals. I mainly just do temp tracks. But sometimes, those temps are good enough. Definitely not all the time.

Hope this helps.