Topic: applause volume for quiet show

[T.J.]

okay, the soft knee limiter sounds like a good choice for what i'm trying to accomplish. a little search yielded this info:

I use the soft knee limiter at 12db with the attack, sustain and decay values configured to something like 3ms, 5ms, and 3ms with lookahead enabled.  That will catch the impulses and drop them down with minimal effect on the surrounding music. This should apply to any editing tool that allows configuration of the limiter parameters.

The behavior I see with the usual default settings is that when an impulse is detected, the clap triggers the limiter.  The attack is too slow to apply the complete 12db attenuation to the impulse but then the limiter stays applied long past the impulse.  That affects the sound following the impulse such that the entire selection was affected and there seems to be breathing in the levels.   The default settings cause the affected channel to be lower than the other channel so the entire image shifted toward the untreated channel.  When I zoom and select just the impulse, the attack is so long that it barely affects the clap if at all. 

With the configuration values suggested above, the limiter kicks in quickly and drops away quickly so the surrounding music is hardly affected.  I found this very effective for cleaning up a show in which there was a clapper in one channel that was louder than the music.  I was able to drop her down so that she was about the same volume as the rest of the crowd noise and I operated on just that channel.  With the limiter settings I chose, I could not hear any difference in the average levels of the two channels.  

In the cases where the clapping was after the music, I was able to select a wide area and apply my settings.  However, where the clapping is coincidental with the music, you will need to selectively apply the limiter to the impulses you want to remove.  Otherwise the impulses of the music will be affected too.

To get a feel for this, zoom into the waveform and measure the width of the clap impulse.  You will find that it is a few milliseconds only.   Select about 400 ms centered around the clap and then play with the limiter settings.  You can dail in the settings with a few trials.   Then apply the limiter across a wider selection area.  

If you don't have any presets, you can make your own limiter by graphically adjusting the gain curve.  The default display is a unity gain curve with three points.  drag the endpoints to the desired position.  mouseover the gain curve and drag where you want to add a point.  to make a soft knee limiter, drag the upper right point to the limit you want.  Then drag a couple of points in the center as needed to make a gracefull trasition from the unity curve to your limit value.  

Limiter is at:

Load file
Process>Dynamics>Presets tab

then I chose Preset (-12, soft knee), Switched back to Settings tab, unchecked AUTO, and put in Lil's variables.

i guess when i get home i'll try the preset (-12, soft knee) to a small selection that contains the "clapper." if it sounds good, i'm wondering if i can just apply it to the entire file then add the gain back. i'm afraid it may affect the peaks in the music sections in negative way. if so maybe i'l just apply the limiter to the selection in between the music that has the clapper.

eidt to say: the parts in bold have me a little confused

[Brian Skalinder]

i guess when i get home i'll try the preset (-12, soft knee) to a small selection that contains the "clapper."

LKIJ's settings are a good starting point.  Fast attack, fast release, only a brief lookahead.  As for the threshold, a couple approaches I've used in the past.

Applying compression / limiting to the entire file:

• Over the entire WAV file, find the loudest music level that's adjacent to applause.
• Use the peak level of the loudest music as the threshold for the compressor - this way, no music is compressed (since it's all at or below the threshold), and only applause is impacted (since it rises above the threshold).

This is pretty easy, and takes very little time.  The downside:  it applies the same compression / limiting to all of the applause.  The settings may not be ideal for some sections of applause, but generally works fairly well.

Applying compression / limiting to each individual section of applause:

• For each individual incident of applause that rises above the level of the music, find the loudest music level adjacent (left or right) to the applause.
• Use the peak level of the loudest music as the threshold for the compressor - this way, no music is compressed (since it's all at or below the threshold), and only applause is impacted (since it rises above the threshold).

This takes more time, but provides the flexibility of customizing the compressor / limiter settings for each incident of applause, usually the threshold.

[T.J.]

i was messing around with my recording last night and can't seem to get the limiter to work. i went into dynamics > settings and put in the variables 3ms, 5ms, 3ms, selected lookahead, set my threshold to -5.49dB (the loudest volume of the music) and hit process but nothing seemed to change. I would expect to see my new peak to be -5.49dB. the peak hasn't changed at all, nor has anything on the waveform. i tried using it for both the entire file and then just a selection. the peak remains @ -2.43dB and there are numerous points above -5.49dB. the claps aren't minimized, so i think i'm doing something wrong. i tried using the soft knee limiter (-12) preset and that didn't seem to do anything either. the only thing i didn't change was the graph (which still looks like a straight line @ a 45degree angle), the ratio (1:1), and the normalize option. i tried it both with peak and average volume selected under the advanced tab.

am i missing something? the claps are very brief, so i think my variables are okay. maybe i should try to set my threshold closer to -10dB? even so, nothing was accomplished with my initial settings so i'm confused 

[T.J.]

okay going back through a couple posts on other threads i think i've found the error of my way. i need to set the ratio, right? i guess something aggresive like 20:1 will compress my burst of claps.

a couple more questions:

is it necessary to use the soft knee limiter if you are changing the variables/threshold/ratio? the preset doesn't seem necessary if i am inputing my own values. maybe i'm wrong though.

should i need to mess with the graph if i input my own values?

the peak of the music is generally around -10dB except of one "sing-a-long" spot that reaches -5.49. is it possible to set the threshold @ -10dB? i guess i'll have to try it and figure out.

in the advance settings there is a choice of: peak or average volume (something like that). the default setting has average volume selected. wouldn't i want to select peak or am i looking at this backwards?

sorry for all the questions/hijack. hopefully this is beneficial to others....

[Brian Skalinder]

i need to set the ratio, right? i guess something aggresive like 20:1 will compress my burst of claps.

Yes, you need to set the ratio.  The higher the ratio, the more aggressively the compressor reduces signal that rises above the threshold.

is it necessary to use the soft knee limiter if you are changing the variables/threshold/ratio? the preset doesn't seem necessary if i am inputing my own values. maybe i'm wrong though.

The presets are just that - presets.  Specifying your own settings overrides the presets.  In general, the presets are a nice starting point, but I've always tweaked the settings for my own purposes.  Note you can create your own, custom presets, if you wish.

should i need to mess with the graph if i input my own values?

The graph is just a visual representation of the numeric values.  You can adjust the graph and the app will change the numeric values accordingly, or you can adjust the numeric values and the app will change the graph.

the peak of the music is generally around -10dB except of one "sing-a-long" spot that reaches -5.49. is it possible to set the threshold @ -10dB? i guess i'll have to try it and figure out.

You can set the threshold to whatever value you wish.  In this instance, it may make sense to set the threshold to -10 dB and apply to the whole value.  Alternatively, it may make sense to apply the -10 dB threshold to the WAV both before and after the "sing-along" spot, and then independently apply a compression with a different threshold (-5.49 dB) to just the "sing-along" spot.

in the advance settings there is a choice of: peak or average volume (something like that). the default setting has average volume selected. wouldn't i want to select peak or am i looking at this backwards?

I don't recall this setting, but I'd set it to peak.  Maybe try average as well, to see which you like, but I think peak is what you want in this instance.

[T.J.]

excellent! i knew i was forgetting something. the 1:1 ratio i had it set must have basically compressed my peak to the same exact peak i had prior. i'm going to try it @ 20:1 for now

the peak of the music is generally around -10dB except of one "sing-a-long" spot that reaches -5.49. is it possible to set the threshold @ -10dB? i guess i'll have to try it and figure out.

You can set the threshold to whatever value you wish.  In this instance, it may make sense to set the threshold to -10 dB and apply to the whole value.  Alternatively, it may make sense to apply the -10 dB threshold to the WAV both before and after the "sing-along" spot, and then independently apply a compression with a different threshold (-5.49 dB) to just the "sing-along" spot.

great idea! i wanted to apply the compression to the entire file at once b/c for this particular recording i see no benefit in compressing each individual spot b/w songs. i'm only trying to squash one loud clapper.

i'm going to take the first part of the recording up to the -5.49dB peak and apply -10dB of compression. keep the 2-3 second peak spot free from compression. then apply the -10dB to the rest of the file. hopefully that will give me enough room to boost my levels about 10dB.

Thanks again Brian for advise! +T

i'll let you guys know how it turns out.