gate sidechain question

[vtone]

I remember reading on gearslutz (couldn't find it in a search) where one could copy a track, nudge it a bit earlier, and put a gate on the original track-- sidechained to the copy. the idea was that it would trigger the gate to open a bit earlier so as not to clip the original.

my questions:

how is this better than just setting a fairly fast attack, slow release?

how much do you nudge the copy?

and what do you set the release to?

is this an effective way to cut out bleed?

I love using the chandler tg1 compressor on vocals, I often compress lightly on one channel and then into the limiter for a crisp paul mcartney-style compression. but the backround noise is so high, it's difficult to deal with.

it's tedious to go in and create fades before and after every line so I'm hoping a technique like this will speed up the workflow.

[Wayne]

I'll take a stab. Generally, it allows the option of a slower opening rate which might sound more natural, or to possibly stay will a higher threshold if you're going for a bit more of a signal that might be up ahead of where the peak is.
The software gates seem to get away with 'Zero attack' (on a drum a least), but analog gates have been known to 'click at their fast settings.

I love using the chandler tg1 compressor on vocals, I often compress lightly on one channel and then into the limiter for a crisp paul mcartney-style compression. but the backround noise is so high, it's difficult to deal with.

Depending of course on a million different things, go after the noise source first.
Wayne

[Kiwiburger]

how is this better than just setting a fairly fast attack, slow release?

It's better to open the door before you run through it, rather than to smash into the door, and then open it - no matter how fast you can open it after you've smashed into it.

The gate has a threshold - and even if it opens immediately the threshold has been reached - you will lose whatever was before that threshold. How much that is, depends entirely on the source.

how much do you nudge the copy?

50 ms should be a good starting place - experiment.

and what do you set the release to?

100 ms should be a good starting place - experiment. Remember that the sidechain track is delayed by the amount you select, so you have to add this before you even start considering the release time.

is this an effective way to cut out bleed?

Absolutely not. I'm not a big fan of gates for anything other than an obvious gated sound. As a noise or bleed fix, they are usless. For the simple reason that they let all the noise or bleed through when they are open. When the sound isn't there, they shut - but otherwise you have achieved nothing.

If you have phase problems because of bleed, those problems are still there when it matters - when the gate is open.

You might like the gated sound anyway, for what it is. But it's not a cure for noise or bleed. IMO.

[Paul Frindle]

Originally Posted by Wayne

I'll take a stab. Generally, it allows the option of a slower opening rate which might sound more natural, or to possibly stay will a higher threshold if you're going for a bit more of a signal that might be up ahead of where the peak is.
The software gates seem to get away with 'Zero attack' (on a drum a least), but analog gates have been known to 'click at their fast settings.


Depending of course on a million different things, go after the noise source first.
Wayne

There are two reasons for reducing attack times; the most obvious is to reduce the 'clicking' associated with switching signal on and off - the slower rise produces a gentler 'edge' which is more musical - particularly on LF programme where the clicking would not be masked by the music. The other reason is to prevent the gate opening on noises and events that would be too fast to hear anyway.

Look ahead (which is what you are describing) is employed to compensate for the extra time it takes the gate to open with the extra attack law - so you don't actually end up losing the attacks.

There is an optimum lookahead time for music programme (which more or less sets your maximum attack time - if you want to retain action on peaks), which allows just enough attack slew to soften the clicks - without it audibly beginning to open noticeably before the peak (which would sound awful of course). BTW this is much less than 50mS!

The software dynamics process I designed has this lookahead built in and a suitable range of useable attack time control to allow a good range which retains action on peaks - or slower attack times to produce late action effects.

I would suggest that my design is not the only one to incorporate look ahead - and therefore it's wrong to suggest that S/W gates mostly get away with zero attack time. A zero attack time in a digital gate would sound just as bad as that on an analogue gate - if it were not for the signal timing manipulation possible in the digital domain :-)

[Wayne]

Originally Posted by Paul Frindle

I would suggest that my design is not the only one to incorporate look ahead - and therefore it's wrong to suggest that S/W gates mostly get away with zero attack time. A zero attack time in a digital gate would sound just as bad as that on an analogue gate - if it were not for the signal timing manipulation possible in the digital domain :-)

Aye, thank you. I completely forgot about lookahead.
Wayne