Using compression for color when mixing Dialogue.

[sailor]

Depends on the material. If it's well-recorded & well-edited material for a feature (which also may imply having a larger time budget for surgical mixing), going without compressors makes plenty of sense. If it's TV or doc stuff with tighter turnaround times and more variable audio quality, putting comps on the dialog helps tame things and eliminates at least some of the fader gymnastics. When I use them, I'm not really looking for "color," just control -- mostly just snaring the occasional peak.

[Tom Fleischman]

Unless the signal is weak and the dialogue is buried under a lot of BG noise, I use compression on just about everything. Depending on the script, performance, and what's going on behind the dialogue I sometimes use quite a bit. I find it helps a lot with intelligibility. I do not like to squash it, just to smooth out those annoying bursts of energy that actors often generate which seem to go right through to the back of my skull...

The compressor is first in the chain, on the input. -24 at 3:1 with a 10db knee, fast attack, 60 - 100ms release is not unusual. I do not use any compression on the busses.

[Andrew Mottl]

Tom, thank you for sharing your knowledge!!
I would like to ask, do you have the comp before any eqs (notch) and denoisers against cam noise, hums etc. or has this already been taken care of by the time it hits your (desk?) inputs and compressors?

Cheers!

[Tom Fleischman]

Originally Posted by Andrew Mottl

Tom, thank you for sharing your knowledge!!
I would like to ask, do you have the comp before any eqs (notch) and denoisers against cam noise, hums etc. or has this already been taken care of by the time it hits your (desk?) inputs and compressors?

Cheers!

I put the compressor on the input to the channel strip so it's the first thing the signal hits coming into the console. Any plugins or signal processing that's happening in ProTools is ahead of that (usually very little, sometimes some pitch shifting, denoising or declicking). Any EQ, notching, or processing that happens in the desk (which is where the lion's share of the processing is done) is after the compressor.

I use no dynamics on the busses during predubbing, but I will put a soft limiter on the main stem busses, ahead of the recorder, during final mixing.

[sailor]

Originally Posted by Tom Fleischman

I put the compressor on the input to the channel strip so it's the first thing the signal hits coming into the console. Any plugins or signal processing that's happening in ProTools is ahead of that (usually very little, sometimes some pitch shifting, denoising or declicking). Any EQ, notching, or processing that happens in the desk (which is where the lion's share of the processing is done) is after the compressor.

I use no dynamics on the busses during predubbing, but I will put a soft limiter on the main stem busses, ahead of the recorder, during final mixing.

Tom, I sometimes find that my channel compressors, if they're first in the chain, end up working harder than I want them to, due to the low-end rumble that exists in a lot of in-the-field dialog recording. So I usually have the EQ, or at least a HPF, in line before it. With the comp first in your chain, do you commonly encounter the same situation? Is it just dealt with it as needed?

[ggegan]

If the compressor is in a reassign chain, you can use EQ or a filter on the source track feeding the chain in order to get rid of the rumble, but you don't necessarily even need to have a stand-alone EQ or filter prior to the compressor to deal with this. There are many compressors that have built-in filters which can be placed either in line or keyed that you can use to control which frequencies the compressor sees. They can usually be set to HP, LP or even band pass. Besides rumble and bass bumps, sibilance can also cause comp issues, so if you wanted to, you could also filter out highs or high-mids and leave the sibilance to be dealt with by a de-esser, which is specifically designed for the job.

[cananball]

Tom thanks for the insight. Just saw Hugo yesterday, dialogue sounded great!

[sailor]

Originally Posted by ggegan

If the compressor is in a reassign chain, you can use EQ or a filter on the source track feeding the chain in order to get rid of the rumble, but you don't necessarily even need to have a stand-alone EQ or filter prior to the compressor to deal with this. There are many compressors that have built-in filters which can be placed either in line or keyed that you can use to control which frequencies the compressor sees. They can usually be set to HP, LP or even band pass. Besides rumble and bass bumps, sibilance can also cause comp issues, so if you wanted to, you could also filter out highs or high-mids and leave the sibilance to be dealt with by a de-esser, which is specifically designed for the job.

I suppose a broader question is what are the relative advantages or disadvantages to applying compression pre- or post-EQ on dialog tracks. I tend to put the EQ first for the reasons I stated (and seldom end up creating the same problem for myself with too-aggressive EQ-ing), but am open to considering other ways of doing it. Thoughts?

[mattiasnyc]

Interesting topic.

Personally I find that with the compressor in the beginning of the chain it ends up working too hard at times. I get much better results by doing at least basic rides and then having the compressor post fader and post eq. That way I get it to work pretty much equally hard throughout the program material.

I can also do several tracks first and just bus them to the same DG bus with that compressor on it.

[Brian B]

Originally Posted by Tom Fleischman

Unless the signal is weak and the dialogue is buried under a lot of BG noise, I use compression on just about everything. Depending on the script, performance, and what's going on behind the dialogue I sometimes use quite a bit. I find it helps a lot with intelligibility. I do not like to squash it, just to smooth out those annoying bursts of energy that actors often generate which seem to go right through to the back of my skull...

The compressor is first in the chain, on the input. -24 at 3:1 with a 10db knee, fast attack, 60 - 100ms release is not unusual. I do not use any compression on the busses.

Good luck tonight!

[TheGolfer]

Congratulations Tom for your Oscar. Amazing!

[cebolao]

I put compressor on my dialog stem. When it works too hard i automate the threshold.

[DEAN HUMPHREYS]

I have only recently discovered 'gearslutz' and this is my first posting. I feel like someone about to go into a confessional box - vaguely apprehensive but in need of purging myself.
I don't use compression on dialogue.
Phew. I feel better already!

[ggegan]

Originally Posted by DEAN HUMPHREYS

I have only recently discovered 'gearslutz' and this is my first posting. I feel like someone about to go into a confessional box - vaguely apprehensive but in need of purging myself.
I don't use compression on dialogue.
Phew. I feel better already!

That'll be three Hail Mary's and three Our Father's.

A compressor or nimble fingers can basically accomplish the same thing.

[Steven1145]

I think there are lot's of different approaches to compression for dialog mixing.
I love using the channel comps on the S5, while also keeping a multiband comp on the main dialog bus with a high threshold to keep any leftover peaks under control.
I use the channel comps as needed, and most of the time they are bypassed. But when I feel a line needs help to get through, or that an actor is a bit all over the place in terms of dynamics, I'll be using them.

May this post also be the opportunity to welcome Dean Humphreys to Gearslutz. I hope you will share plenty of your extensive knowledge and experience with us!

[Dragomir]

Originally Posted by Tom Fleischman

Any EQ, notching, or processing that happens in the desk (which is where the lion's share of the processing is done) is after the compressor.

Could you get the same results ITB or the desk (what desk btw?) is that much superior?
Do you use any analog gear these days?

[ThisIsSka]

Originally Posted by Dragomir

Could you get the same results ITB or the desk (what desk btw?) is that much superior?
Do you use any analog gear these days?

I work for a small independent post studio so I'm not 100% sure what the larger guys will do. But for me, working ITB means I can jump from project to project a hell of a lot faster than having to deal with outboard gear, let alone a console that needs to be adjusted from session to session. And really, plug-in performance is close enough to physical gear that my clients certainly don't notice. But what they do notice is that I can throw up a session from 3 months ago, tweak the bit they want changed, and re-export very quickly, which I imagine could be very difficult to do with a lot of outboard gear.

[minister]

Originally Posted by ThisIsSka

I work for a small independent post studio so I'm not 100% sure what the larger guys will do. But for me, working ITB means I can jump from project to project a hell of a lot faster than having to deal with outboard gear, let alone a console that needs to be adjusted from session to session. And really, plug-in performance is close enough to physical gear that my clients certainly don't notice. But what they do notice is that I can throw up a session from 3 months ago, tweak the bit they want changed, and re-export very quickly, which I imagine could be very difficult to do with a lot of outboard gear.

It really isn't all that hard. Some easy methods : Recall sheets, recall notes in the comments field of the track in question, Masseys Recall Notes RTAS plugin. Less convenient are polaroids or digital pictures stored with the session.

smassey.com

Scroll down for the recall plugin.

[Steven1145]

Originally Posted by ThisIsSka

... let alone a console that needs to be adjusted from session to session....

I understand what you mean, but bear in mind that proper post oriented digital consoles (like the Euphonix/AVID System5) have total recall of all project parameters, including multiple automation versions and passes. I still by far prefer mixing on the S5 than completely ITB, except for very short projects like film trailers or commercials.

[cinetj]

Originally Posted by Steven1145

I think there are lot's of different approaches to compression for dialog mixing.
I love using the channel comps on the S5, while also keeping a multiband comp on the main dialog bus with a high threshold to keep any leftover peaks under control.
I use the channel comps as needed, and most of the time they are bypassed. But when I feel a line needs help to get through, or that an actor is a bit all over the place in terms of dynamics, I'll be using them.

May this post also be the opportunity to welcome Dean Humphreys to Gearslutz. I hope you will share plenty of your extensive knowledge and experience with us!

Hi, Steven

Could you tell me why do you prefer to use a multi-band compressor for the dialog bus instead of a "regular" one?

Thanks

[BIGBANGBUZZ]

From my experience in reading and my own experience on the tools...

Movies a little bit, TV a fair bit, spots and promos for TV a shit load.


Sent from my GT-N7000 using Tapatalk

[Steven1145]

Originally Posted by cinetj

Hi, Steven

Could you tell me why do you prefer to use a multi-band compressor for the dialog bus instead of a "regular" one?

Thanks

Even though a multiband compressor will colour the sound more if used too much (it is in fact a dynamic EQ), it can be quite transparent as a "catch all" device. As I said, I try to tame dialog dynamics as much as possible at the track level with either faders or some comp, so the MB comp at the end of the chain is only there to round the edges a bit, and help (within reason) to deliver a coherent dialog predub.
When I am actually working on the dialog, the MB comp is bypassed, so that I can really hear what is going on. I switch it on when I add all the other elements for the main mix (AMB - FX - MUSIC), and have the returns of the MB Comp on a fader within my Main mix session.

[nzl62]

You can enhance the character of an actors voice with a bit of compression but you obviously need to be careful and it affects bg noise etc. My dialogue chain always includes some sort of compression near the beginning to mainly protect the rest of the process chain, then a shaping compressor at the end, mostly on a channel but sometimes buss.
I struggle with multi-bands as they can really shift the tone and character of dialogue and voice overs. For a longtime I have Summit DCl200's as part of the analogue stage on my dialogue chain and they were superb. I often find that nae sayers to compression often simply do not know how to use them or haven't used a good one

[Henchman]

I have a compressor at the end of my dialog processing chain, but it only hits with very loud dialog.

[ggegan]

Originally Posted by nzl62

I often find that nae sayers to compression often simply do not know how to use them or haven't used a good one

Either that or they just don't like hearing compressed dialog. I'm not opposed to using compressors, I just don't like to hear them working.

[Steven1145]

Originally Posted by nzl62

You can enhance the character of an actors voice with a bit of compression but you obviously need to be careful and it affects bg noise etc. My dialogue chain always includes some sort of compression near the beginning to mainly protect the rest of the process chain, then a shaping compressor at the end, mostly on a channel but sometimes buss.
I struggle with multi-bands as they can really shift the tone and character of dialogue and voice overs. For a longtime I have Summit DCl200's as part of the analogue stage on my dialogue chain and they were superb. I often find that nae sayers to compression often simply do not know how to use them or haven't used a good one

My dialog chain involves audio hitting the S5 channels, and then being routed to a DNS One in PT, then coming back into the console on separate strips that are routed to a mono Dia bus (in turn routed to the main 5.0 DIA stem). This allows me to compress (when needed only) before or after the noise reduction, and that flexibility is important to me.
The MB only catches a few stray dynamics of loud dialog.

Sent from my XT910 using Tapatalk

[cinetj]

Originally Posted by Steven1145

My dialog chain involves audio hitting the S5 channels, and then being routed to a DNS One in PT, then coming back into the console on separate strips that are routed to a mono Dia bus (in turn routed to the main 5.0 DIA stem). This allows me to compress (when needed only) before or after the noise reduction, and that flexibility is important to me.
The MB only catches a few stray dynamics of loud dialog.

Sent from my XT910 using Tapatalk

Thank you for answering Steven, but I would still like to know if you're using the multi-band compression as such as an"compressor-eq", focusing on certain parts of the spectrum, or if it's just because you like the plug-in or piece of hardware that you use for this task, that could otherwise be replaced by a regular compressor.

[Steven1145]

Originally Posted by cinetj

Thank you for answering Steven, but I would still like to know if you're using the multi-band compression as such as an"compressor-eq", focusing on certain parts of the spectrum, or if it's just because you like the plug-in or piece of hardware that you use for this task, that could otherwise be replaced by a regular compressor.

I dont ride it hard enough most of the time for it to be considered EQ-esque (which would also involve having different comp ratios or Make up gains for different frquency bands). But as a multi-band it can be effective without the audio sounding compressed, so I think that I would not replace with a regular comp as the result is clearly not the same.