Comments

Chienworks wrote on 12/5/2002, 4:47 PM
First off, take a few deep breaths and calm down :)

Sadly, there isn't much you can do. Once something like reverb or echo has been mixed in with an audio signal, it's for all practical purposes impossible to eliminate it or even reduce it. That being said, there are a couple of things you can try. Acoustic Mirror probably won't help much because it's purpose is to simulate various room echos and reverbs, not counteract them.

You can use a noise gate to cut out the sound during quieter parts. This can help because the echo is usually a lot quieter than the main signal. However, you end up with a recording that sounds like parts have been unnaturally sliced out. It's worth a try to see if it helps. Of course, it won't eliminate the echoing sound during the speaking.

Another thing that might help some is EQ. The important part of the speaking is mostly contained in a relatively narrow frequency band, probably 300Hz to 3KHz. The echos and reverb will tend to spread out through more of the frequency spectrum. Using EQ to eliminate frequencies outside the normal vocal range will tend to cut down on echoes and reverb more than it will cut down on the main voice.

Neither of these will perform miracles. With enough experimentation they might help just a little bit though.
Erk wrote on 12/5/2002, 6:09 PM
Doormill,

Chien's right, try EQ, cut some of the low end, maybe that will take some of the boominess that echo/room reverb adds to the sound.

Also, just a thought.. depending on what your clip is and its context in the overall video, could you add some sound FX, ambience, or music to sort of cover over the less than desirable room sound? Maybe not, if you're trying to get clarity on the people speaking instead of covering it up.

G
craftech wrote on 12/5/2002, 6:19 PM
Try posting here:
http://www.dv.com/community/forums.jhtml;jsessionid=2OXGONKF24BZMQSNDBCSKHSCJUMEIJVN
on the Audio Solutions forum. Ask Jay Rose (the moderator)the question. You will have to sign up for the forums first. Be sure to tell him that you don't want to spend hundreds of dollars for software to fix the problem first.

John
musicvid10 wrote on 12/5/2002, 11:47 PM
doormill,
Here's a technique (not a fix) that has been causing quite a stir since I posted it on the Studio webboard. Hope it gives you some help, too.
---------------------------------------------------------------
Here's an old trick (from and old dog) dating back to analog recording days that usually gives some improvement (you need a good wave editor like SoundForge):

1. Select All, then Copy the audio.
2. Paste to a new track, invert the waveform, and apply moderate compression.
3. Reduce the new track volume so that a preselected "quiet" area is about 50% of the level of its corresponding area in the original.
4. Paste Mix the new track into the old. Renormalize if necessary.

There should be a noticeable improvement in clarity and echo reduction because you have applied negative feedback to the areas where the echo is most objectionable. Too high a compression or too high level of the feedback track will give a "pumping" effect, however.
---------------------------------------------------------------
ibliss wrote on 12/5/2002, 11:54 PM
musicvid - nice tip. The key to that solution is the compression - you'd have to be fairly precise with it to get the best possible result. No compression would of course result in total cancellation of the original audio track!

Mike K
musicvid10 wrote on 12/6/2002, 12:15 AM
Mike K,
You got the gist of it right away. Your comment about precision should not be taken lightly, I apply it quite conservatively, then back off as soon as it starts to sound unnatural or I lose subtleties in the audio.

We used to do this realtime in the studio using a compressor and phase shifting to mix down "bar band" tapes. Using the technique in Sound Forge is a snap by comparison.
TorS wrote on 12/6/2002, 1:36 AM
Nothing in a video is sacred to me, but if there is talk - and the talk is neccessary - I want it to be clear.
You're getting good advise (meaning a lot of ground has been covered in just a few replies). Try various combinations of those suggestions.
Also, if clarity and communication is your main issue, you can supply subtitles to whatever is being said. It's a striking effect (possibly even more so in a country where subtitles are uncommon) but it does get the message across.
If the overall quality of the soundtrack is as important as the words being spoken (and there is no way to re-record the lines and post-sync) try to camouflage unnatural gaps etc with a sound carpet (music or else) with frequencies mainly unlike those of the spoken words.

If it is a short clip in a longer production, you will want to introduce the sound carpet earlier on - to not draw attention to the problem clip.

Tor