garage band/imovie

wow now this is a bit vague.

Is it sound in between talking, hiss or general rowdiness.
The things is either of these will need different tools to do this.

In between talking you can just cut and snip bits out. Hiss requires a noise reduction utility which ithink is only available with logic or buy as a seperate plug in. and general rowdiness requires a similar thing but a compressor can also help.

Can you give us a bit more information on your audio file.

A terrible business

I know a number of serious film makers who do use iMovie (at least iMovie HD) for small projects like low-budget TV documentaries. It really seems to be a good-enough tool up to semi-pro level where perfection and absolute control are not required.

Audio is another matter. Garage band is great for chucking together musical ideas. I love it for this purpose. However, doing editing (imho) is a nightmare due to the lack of sample-level precision and a fairly small tool set in terms of effects and filters. The best you could do with garage band would be some sort of noise gate. For hiss - and this is pretty crude - you could duplicate the audio and split it into a bass and mid channel, and a high frequency channel using low pass and high pass filters respectively (included in your core audio effect set). You could then use a noise gate on the high frequency channel to avoid the full-spectrum cut-in cut-out effect of a noise gate. It will still sound dreadful.

Staying with free tools, there is the noise-reduction algorithm in Audacity (a much better editor btw) but I've never got satisfactory results from this.

The previous post is right pointing you in the direction of proper audio tools, and Logic Studio is good value. You get Soundtrack Pro which, although a little unstable, is far more powerful than any free tool and is designed for this very task. Logic provides a very tidy set of gates, compressors and noise reduction tools.

All this said, the adage "poo in, poo out" is very true. Developing the skills to record your source material with as little background noise as possible is key. Using a good wind baffle on a windy day sure beats trying to extract the broad-spectrum Braughhhhh in Soundtrack. Similarly, I'm guessing doco makers would sometimes use near-field or directional dynamic mics for interviews to leave out most background noise in the first place (I don't actually know this for sure, but it's what I'd do). Remember, in TV shows, all ambient sound is mixed in later. A Party for example, is silent (freaky to watch people dancing in silence during a shoot) except for the dialogue parts being captured on condenser mics. The music and hubbub is dubbed in later for atmosphere. Try recording the dialogue in a real party and, well, you'd probably be in real trouble. You'd certainly have more hope with a dynamic mic, but it would probably be easier to step out onto the balcony for the interview. Good luck.

I'll need some input from other people here!

1. If you are using iMovie 08, don't. As an iMovie 08 owner, you can get the previous version, iMovie HD, as a free download from Apple.
2. For this kind of work, you will need to export your audio track for processing. From memory this is quite straight forward in iMovie HD.
3. I have only ever played for any length of time with 3 tool sets:

• Free tool and plug-in I've found on the net
• Logic Studio's included plug-ins
• Waves plug-ins

This does, however, cover the spectrum. Now, trying to remove the kind of noise you are talking about will require pretty mean tools, and probably a lot of your time a patience. Regardless of plug-ins, I suspect that we're not looking at applying one or two plug-ins to the file and hitting process. You will probably need to crawl over this audio doing serious, semi-manual second by second massaging, using the best tools for each block of different acoustic messiness. Wind-gusts, for example, will require different treatment to background conversation will require different treatment to white or pink noise. Background hubbub will be hardest because it's obviously of the same spectrum and acoustic properties of the human speech you are wanting to differentiate it from.

As for choice of tools, I'm calling out for help here from people with broader experience. I've never had much joy with free tools on this kind of problem. The tools included in Logic Studio ($500) are okay, so using Soundtrack pro (included) I'd experiment with noise-print noise reduction, the other noise-reduction plug-in included that just has a series of sliders, side-chained frequency targeted gates, and manually drawing an amplitude envelope. Expect disappointment. You have a big "poo in" problem here. Now Waves plug-ins change the definition of expensive (I used someone else's). I think a full suit of plug-ins will set you back about $7000, so you'd want to know they'd do the job. Waves have some specific restoration/ noise reduction packages, costing in the thousands, that may be of particular use here (that I have not used). Broadly, Waves are superior to logic's plug-ins, both in terms of sonic quality and (probably) results. I'm still guessing it would require second-to-second examination of the audio choosing the best set of tools for each problem that is occurring. What do others think?