need to reduce echo in post

[Noiz2]

Just curious are you suggesting that this be done in production or in post?

I've heard this suggested as a post "fix" for noise. Though it is usually suggested that you loop a noise sample and invert it. I have never heard it work. Not saying it doesn't, just that I have never found anyone who got it to work. In this case what he's talking about should if done correctly completely cancel the signal.

But when you have this kind of problem go for it. What do you have to loose? Some time but not as much as Foleying the whole thing.

 

[Cedric Akins]

You are right the Foley idea is the best solution. He shouldn't have to reshoot it just get the dancer to agree to preform the dance in a "good" sound environment.

 

[Petros Kolyvas]

Ok although I am all for recording new footwork. I thought I would take 30 minutes to see what would happen if a client came to me with audio like this and that wasn't an option. What could I do?

Here's what I got.

Used:
a) Izotope RX to first run a high-pass filter, then a few passes of noise reduction using samples of just the response of the room. In this case it was nice that the footwork resulted in such sharp transients and a very clear room response independent of the impulses. You can even clearly hear the first reflections!


b) Sonnox Trans Mod. I've had this tool for about 4 years. I don't use it often but this seemed like the perfect time to use it. It's a bit like a reverse compressor in that it can amplify transients (or attenuate them).

I'm not saying the result is any good but I'm always up for a puzzle/playing/testing my limits. I still don't think this would come close to cutting it though. Maybe if it was over some music to hide the rest of the room response. Maybe.

If I was being really critical, I'd say the footwork gets a little too snappy and looses some of the body, but hey, half a dozen of one, six of the other right?

 

[Cedric Akins]

Ok although I am all for recording new footwork. I thought I would take 30 minutes to see what would happen if a client came to me with audio like this and that wasn't an option. What could I do?

Here's what I got.

Used:
a) Izotope RX to first run a high-pass filter, then a few passes of noise reduction using samples of just the response of the room. In this case it was nice that the footwork resulted in such sharp transients and a very clear room response independent of the impulses. You can even clearly hear the first reflections!


b) Sonnox Trans Mod. I've had this tool for about 4 years. I don't use it often but this seemed like the perfect time to use it. It's a bit like a reverse compressor in that it can amplify transients (or attenuate them).

I'm not saying the result is any good but I'm always up for a puzzle/playing/testing my limits. I still don't think this would come close to cutting it though. Maybe if it was over some music to hide the rest of the room response. Maybe.

If I was being really critical, I'd say the footwork gets a little too snappy and looses some of the body, but hey, half a dozen of one, six of the other right?

That was actually impressive. And yes you are right the body of the footsteps has been lost but nothing a little eq couldnot fix.

 

[j1clark@ucsd.edu]

You could use a little compression on a duplicate track, reverse phase on the duplicate and mix them. Phase cancellation should bring down the echo a noticeable bit

Sort of... The problem is that the 'echo' does not just have one reflective surface, and so, for each instant in time, one has a multiplicity of the original signal being received due to the varying paths, with attendant variation in phase.

The 'echoed' signal needs to be deconvolved with the impulse response of the 'room', and at the moment, the techniques to do that, tend to 'add noise' from the mathematical operations... to the point where it's not worth it...

 

[Noiz2]

As a total off the wall option. You can use a noise gate to eliminate all but the main hits. It will then sound like crap. You then use what's left as a trigger for either a sampler or some dedicated "drum replacement" software. I have done something like this before and it works about 80% and the rest you do by hand. The better way is to do it all by hand. There are clean tap samples around, I even have some. Take those and slice them up then place them one at a time to sync with the steps you want. You will probably have to do some switching around till you get ones that fit but it's surprising how much generally drops in. This will take a bunch of time, depending on how many steps you need to replace. It is important to remember that it's not really cheating. Even Fred Astair's taps were Foleyed in after the shoot. It's very difficult to get good sounding taps on a set.

 

[Northainan]

Ok although I am all for recording new footwork. I thought I would take 30 minutes to see what would happen if a client came to me with audio like this and that wasn't an option. What could I do?

Here's what I got.

Used:
a) Izotope RX to first run a high-pass filter, then a few passes of noise reduction using samples of just the response of the room. In this case it was nice that the footwork resulted in such sharp transients and a very clear room response independent of the impulses. You can even clearly hear the first reflections!


b) Sonnox Trans Mod. I've had this tool for about 4 years. I don't use it often but this seemed like the perfect time to use it. It's a bit like a reverse compressor in that it can amplify transients (or attenuate them).

I'm not saying the result is any good but I'm always up for a puzzle/playing/testing my limits. I still don't think this would come close to cutting it though. Maybe if it was over some music to hide the rest of the room response. Maybe.

If I was being really critical, I'd say the footwork gets a little too snappy and looses some of the body, but hey, half a dozen of one, six of the other right?

Wow!! That was impressive. It will pass for the average viewer!