Soundbite gain question and suggestion

[FMiguelez]

Hello.
I used to draw volume automation to tame annoying "ssss" and "shhhs" and "ppps" from singers. It's all fine and works, but lately I've been trying doing it directly with the Set Bite Gain command. This method is faster, and I am liking it better (make a time range selection and use my QKs shortcut, which it gets me command-Y to split the SBs and control-esc to invoke the SB Gain command. Then I type a number and press control-F to crossfade the edits. All this takes no longer than 1.5 seconds per edit).

Question:
Suppose I reduced a SB's gain by 4 dB. The day after, I think it still needs a little more gain reduction. If I go ahead and reprocess the previously processed SB, will it do it from the original SB or from the processed one?
I seem to recall it's the former (maybe from the manual, but I'm nowhere near it ATM). Can someone please confirm?
I hope DP is smart enough to do this on its own, otherwise it would be a waste of time hunting for the original SBs and replacing them, etc.


Suggestion:
After going through a whole song like this, I may end up with dozens of little SBs all over the place. If I want to re-tweak one, DP does this gain change relative to the current SB gain setting. Problem is, I can't remember the original setting of each SB.
So if I want to reduce the gain by 2 dB and then later I want to reduce it further, I need to remember its current value... not very realistic.

MOTU, it would be so useful to set soundbites gain with ABSOLUTE values, so I don't need to keep track of previous values (just like the Trim plug-in does).

I realize I can use the SB info window to see its current value and add or subtract from there, but I think it would be much faster and effective to simply incorporate my suggested feature for this command. Who needs to look in there when the computer can figure it out on its own?

What do you guys think?

Does anyone have a faster or better method to accomplish this?

Thanks!

[stubbsonic]

Agreed. Even if the Soundbite volume dialog had a slider and showed the current value, it would be easier.

[Phil O]

Hello.
I used to draw volume automation to tame annoying "ssss" and "shhhs" and "ppps" from singers. It's all fine and works, but lately I've been trying doing it directly with the Set Bite Gain command.[/i]

How were you drawing in your automation? I find that the easiest way to do it for short times like s's etc. is to use the reshape tool (instead of the pencil tool) with the reshape mode set to flat. Simply drag horizontally above or below the current automation line in the Bite Gain layer and it adds four points at the appropriate places. Takes no longer than making a range selection and you're done.

Phil

[Dan Worley]

The bite gain stays with the new soundbite, until you merge. So if you set it to -4 it will show -4 in the event tab when you select the soundbite.

[Dan Worley]

Hello.
I used to draw volume automation to tame annoying "ssss" and "shhhs" and "ppps" from singers. It's all fine and works, but lately I've been trying doing it directly with the Set Bite Gain command.[/i]

How were you drawing in your automation? I find that the easiest way to do it for short times like s's etc. is to use the reshape tool (instead of the pencil tool) with the reshape mode set to flat. Simply drag horizontally above or below the current automation line in the Bite Gain layer and it adds four points at the appropriate places. Takes no longer than making a range selection and you're done.

Phil

Yeah, this is good and I work that way too, except many times I like to take care of things like this before any processing, and in that case I will use bite gain exactly how FM described. It's a great tool.

[FMiguelez]

How were you drawing in your automation? I find that the easiest way to do it for short times like s's etc. is to use the reshape tool (instead of the pencil tool) with the reshape mode set to flat. Simply drag horizontally above or below the current automation line in the Bite Gain layer and it adds four points at the appropriate places. Takes no longer than making a range selection and you're done.

Hi, Phil.
Now you got me thinking that all this might be a retarded misconception on my part...

I used to do it exactly like you suggest. But I always suspected that sudden drops of volume of 6 or more dB would result in an audible glitch. Since the flat reshape tool draws these sudden jumps (and I probably heard an artifact a couple of times), then I started adjusting the reshaped volume dots so they had a very small slant. I started dragging these "corrected" volume shapes so they got copied where needed (and readjusted if needed). As you can imagine, I quickly got tired of doing that.


I was also experimenting with the workflow I wrote about in the OP... I saw a Groove3 video-tutorial where the presenter used an EQ (along with a frequency analyzer) to tame the "sss" by frequency alone (instead of simply reducing the total gain). It works, but is more work, though.

The very nice thing about the reshape method you mentioned is that it's so easy to readjust a volume gesture once it's there. Hmmmm... I should probably go back to doing it like that.

So Phil, I take it you're not bothered by the squared volume shapes when fixing "sss's"? If you think it's safe doing it like that, then I shall go back to that faster method.

Never the less, for other things, the Adjust SB gain suggestion I made should still hold.

Thanks!

[FMiguelez]

The bite gain stays with the new soundbite, until you merge. So if you set it to -4 it will show -4 in the event tab when you select the soundbite.

Ok.
But if you process the same SB multiple times, will DP do it from the original SB, or keep processing the processed SBs? I'm worried about that, since it may degrade the sound otherwise after a few passes.
Not that I'm too worried of ruining the "beautiful sound of an ssss", but I mean of other things as well.

[Phil O]

So Phil, I take it you're not bothered by the squared volume shapes when fixing "sss's"? If you think it's safe doing it like that, then I shall go back to that faster method.

Actually, if you zoom in on one of those "squared" shapes, there's about a 1ms ramp - at least there is on my system. I don't know if any settings would change that. So yes, it would bother me if that shape was inappropriate for what I was doing, but most of the time it's not a problem. If you try to put those ramps on the zero crossings of the waveform, there's less of a chance of clicks.

As far as sss's are concerned, I find that automating the MW EQ has become my weapon of choice when I'm being surgical. When it's a budget job, I'll use a de-esser (don't like it, but hey, time is money). I usually use the bite gain trick for things like breath noises, page turn noises, smacking of lips, sneezes, farts, etc. In general though, I don't do a lot of de-essing. I think some engineers go overboard with that. YMMV.

Same with plosives - automated low freq cut using MW EQ.

Phil

[Dan Worley]

The bite gain stays with the new soundbite, until you merge. So if you set it to -4 it will show -4 in the event tab when you select the soundbite.

Ok.
But if you process the same SB multiple times, will DP do it from the original SB, or keep processing the processed SBs? I'm worried about that, since it may degrade the sound otherwise after a few passes.
Not that I'm too worried of ruining the "beautiful sound of an ssss", but I mean of other things as well.

When you set it to -4 and later on want it at -6, it will only be attenuated by -6, not -10. and there is no degradation.

[Prime Mover]

Going to have to try that reshape method. I always avoided it because I thought as FMiguelez did, that it was a single sample jump. 1ms will probably take care of all clicks, and it's pretty fast. I would be concerned that a 1ms, though might not creating pops, might make an audible volume jump.

However! I would seriously explore using an EQ to do this instead of gain. You should be able to use the same basic method, just using filter gain automation instead of bite gain. It sounds so much more natural. De-essing isn't all about volume. It's about controlling the harshness that Ssses create. Furthermore, volume automation in these kind of jumps can become obvious. It really sounds like a lowered volume S, instead of a softer more pleasing S.

All I do is throw on an EQ and loop a few words with Ssses in them. Play with one of the high frequency bands until you find a notch that sounds good on the Ssses (try to ignore the rest of the vocals) and a dB drop that makes it pleasing but doesn't lisp. Then create an automation on that filter's gain, starting at 0. In the past, I drew a notch, then used copy/paste to throw it over every S. Not all Ssses are created equal, so you may still want to adjust the notch level on every one, after half-dozen or so, you'll start to get a feel for how much drop you need from the shape of the waveform. If you end up doing it a lot of projects with some of the same voices, just create some EQ presets, and you'll be ready to go next time.

The reshape tool method sounds pretty slick too, but now that I think about it, I'm not sure it'll be any faster or less work than a simple copy/paste. One hand on the mouse, and one hand on "Cmd-V", and you're off!

For an average song, I can probably De-ess everything with this method in about 15 minutes, including finding the EQ setting.

Careful not to lisp!

[Phil O]

One hand on the mouse, and one hand on "Cmd-V", and you're off!

Also, don't forget that this data can be saved as a clipping.

Phil

[waitsongs]

MOTU, it would be so useful to set soundbites gain with ABSOLUTE values, so I don't need to keep track of previous values (just like the Trim plug-in does).

+1000!

[Radiogal]

I have the SB Gain shown in the upper short list in SE window under Selection Information.
When you select a SB the current Bite Gain is shown there.
No need for guessing.
However, I do not remember where I selected and activated this option.
I can not find where to click for this option right now.

EDIT: Found it! It´s in the EVENT LIST.

[Phil O]

Just to be clear (and make sure everybody's talking about the same thing), everyone understands that bite volume and bite gain are two different things. Right? I made an error in an earlier post referring to the Bite Gain layer, and no one called me on it. I should have said Bite Volume layer.

Phil

[Dan Worley]

Just to be clear (and make sure everybody's talking about the same thing), everyone understands that bite volume and bite gain are two different things. Right? I made an error in an earlier post referring to the Bite Gain layer, and no one called me on it. I should have said Bite Volume layer.

I've been talking about Bite Gain. You're right, I didn't catch your layer blunder, but you should be kicked off the list for making it.