electronic hiss when boosting low sound

B

birnbaum

Member
hi folks,

am editing a piece of film with some footage with good sound (after buying the me66 :-)), and a lot of "pre-me66"- footage with miserable sound (much too low levels, but it is people lecturing in a workshop, so they should be understandable).

now: when I boost the low level footage sound from it's about -15 to -19db peak level up to 0db peak level, as should be, I get an awful electronic hiss. is there any method to get rid of this hiss and still get good understandable speech?

thanx in advance,

marion
 
If the noise/hiss is consistent, then this might be a candidate for the Sony Media Software (ex Sonic Foundry) Noise Reduction plug-in.

Other than that you can experiment with equalizers and see if they do much for you, but try downloading Noise Reduction -- it may do what you want.
 
You have had the pleasure of meeting the noise floor. It's nasty and is suppose to stay in the basement where it belongs. :)

As said, there is little you can do to make the sound good. But you can remove some of the noise with Sony's Noise Reducion. Problem is, it will pull out the hiss between words but will only be able to "hide" it during the words. Depending on how close to the floor your dialog is, this might not do anything.

If you can't go back and reshoot, try some ADR or as a last resort throw in some subtitles.

Also, you should not be pulling the audio to 0dbfs peak, are you speaking in analog scale?

Good luck
 
Thanx to both of you.

Had a go with premiere's parametric equalizer, which seems to improve the sound in setup's audio preview, but then on the timeline and compared to the original it's as bad as it was...

And had a go with a downloaded trial version of noise reduction, but again can't see (hear) much result, I guess it's simply that I don't understand the program's working well enough.

after all, especially after reading a recent thread about how impossible it is to work sound as a one person makes-it-all-wonnabe filmmaker, I will most probably find myself someone knowledgable to look over my audio.

wabbit, I did not understand your question about analog scales? I learned on a starter's workshop about sound that I should always record sound as close as posible to 0db peak level, but never above. so I concluded that i would also boost low audio to 0db peak level, and sound forge has such a function, too. is this not right?

and what is adr? would this be that someone repeats the words said and this new audio would be put over the footage?

my best greetings, and I am so glad to have this forum!

marion
 
disclaimer: not a sound expert or anything.

1) 0db should be reserved for the loudest things, like if during the interview, someone shoots the interviewee.
Even then 0 might be too loud.
I usually normalize my peaks to -3 or -6, depending on my mood.

2) if you have some "room" ambience recorded (or you can create some afterward), you could try a noise gate that will completely cut off any sounds below a certain level...the level of the hiss, in this case. Then mix in the room sound to refill the silence that creates.

3) if this is just people talking, you can cut out a lot of the upper and lower end of the freq range. This could also cut down on the hiss.

But when it comes right down to it, unless there is a lot of room between what you want to hear and the noise, you're pretty much screwed.

Oh. Audacity (free program) comes with a noise reduction tool. You sample the noise, then use that sample as the filter. Maybe give it a try and see if it doesn't screw up the stuff you want to keep. (It will eliminate the noise from the good sound, too...but it may distort it too much).

Dan
 
ADR (automatic dialog replacement), otherwise known as looping: you get the speaker to re-record the dialog in a studio to replace the noisy footage. It's tricky to do well, but you can take a run at it (if for no other reason as a learning experience). Get Jay Rose's 2 books on audio, they will help you immensely.

Th confusion of db is very common. You have your dbfs (digital), your dbV (prosumer scale), and your dbu (professional scale). Without going into harsh detail (you can google if you like...probably make your head explode though), bascially ~0db in the analog world is generally calibrated to -12dbfs in the semi-pro (DVX level of equipment) digital audio world (professional digital would be -20dbfs).

So editing on a computer you want to keep the levels averaging at -12dbfs with peaks hitting no higher than -2 but even that is pushing it. I keep the peaks around -4 for dialog. If your meters are actually hitting 0dbfs then I guarantee you are creating distortion.

The idea behind pushing as close to 0dbfs during production is to get the most dynamic range in your audio above the noise floor. -12dbfs gives you some breathing room before your audio is ruined by hitting the distortion "ceiling". In post, as you pump up the db for your dialog you bring up the noise floor too. Since the floor is mixed in across the voice frequencies, there really is little you can do to create any more distance from your dialog to the floor. Try pulling up the frequencies in the speakers voice range, you might get a slight improvment.

If this is a narrative piece I would suggest you try and reshoot scene or ADR.

Good luck
 
Also, check out Bias' sound soap. I am having noise problems on the feature I'm working on, and it's really helping.

-Dave
 
thanx to all of you, i think now i can manage my sound. will definitely get the jay rose books, too.

wow, its the international documentary festival now here in town, and its great! loads of amazing documentaries.

marion
 
You must log in or register to reply here.
Back
Top