Correct order of operations for post processing of dialogues
What is the correct order of operations for the post processing of vocals- dialogues & VO (for a feature film); i.e, normalisation, Noise reduction, compression, EQ... I know it depends on the requirement and from case to case, but generally speaking what is the correct sequence in which these operations should be performed..?
What i figured out and concluded after some research and the way I am planning to go about it is:
1.) First to normalize the dialogue track,
2.) Then remove the DC offset,
3.) Then remove all the clicks and pops,
4.) Then remove the background noise by doing the noise reduction process,
5.) Then compress the dialogue track,
6.) Then do the EQ,
7.) Then add some reverb to it,
8.) And then finally applying the hard limiter.
Is this alright or am I missing something here or anything else in general...
There are many different opinions on this on various forums and websites... Some say to do the EQ first, then the NR, then compression and finally normalisation... Some have other order...
I am in a big fix here guys... Any help would be greatly appreciated. I wanna be sure before I go for it.
Hello Steve and welcome to the Cow Audio Forum.
Hmm, let's see....
I have never normalized a dialog track.
If you're getting DC offset then something's technically wrong with the recording.
Which clicks and pops?
Noise reduction is frequently the result of a bad location decision. While some may be used, if it's that bad ADR is the solution. Exterior scenes in "Nurse Jackie" are ADR, for example, and stunningly good.
If you're mixing with automation, you may not need to compress anything but the two mix (or 5.1).
I'd probably EQ earlier, if needed.
Adding reverb or other effects to ADR to make it sound more like the location dialog ambience may help, but just adding it to add it is not called for,
I usually have a compressor followed by at least one limiter in the two mix, but your mileage may vary.
It sounds like you have been handed some poorly recorded dialog tracks. Each may have their own problems (or this one worse in one area than another) . Sorry to hear it, but I don't think there's a gold standard recipe to salvaging bad audio. For me, it's about making whatever I have sound as though it's never been touched.
Cow Audio Forum Leader
Want better production audio?: Ty Ford's Audio Bootcamp Field Guide
Ty Ford Blog: Ty Ford's Blog
Like Ty - i think its too much.
Id err away from normalising and compressing and limiting. Sounds like youre trying to make a brick wall of words here.
If its a film, and therefore for Cinema then the sound platform your'e making for is designed for light and shade, highs and lows. youll have none after all that
Its all about texture. Letting him be loud when he shouts and quiet when its soft.
And as fpr reverb. If you really want or need that, it should be done in the final mix only - in fact id suggest the only processing you do to a VO reccording before the mixing stage is
1) De click and pop
2) Possibly adjust levels if certain parts are way over or under, but not normalising.
Batch processing is your enemy in the creative mixing space as it doens know how your emotions are running right now!
Post Production Dubbing Mixer
@ Peter Groom
Thnx a tonn for reverting back...appreciate it!
I'm not trying to make any brick wall or anything, it's just that the tracks I have they have some problems and they need to be taken care of. I know that cinema is all about highs and lows, lights and shades. I'm not suppressing the shouts either, it's just that the audio is not as rich as it's supposed to be so thats y i was looking for normalisation etc... I know nothing beats a well studio recorded pristine quality dialogue track but if there comes a situation where you can't have one, I was just wondering if there are little problems in the track then can they be taken care of and is there any particular procedure or order or sequence to approach it...
Thnx a ton Ty for such a warm welcome!
Well I am not getting any DC offset as of now but its advisable to make it null before going for noise reduction/removal...
Its good to knw that u nvr had the need to normalize the dialogue track, u must be working with good equipment. But I would be needing it that I know... After hearing it...
Anyways I was asking in general that if normalization is required then when should it be performed..? At the end after working with the dynamics of the audio, or at the very beginning before retouching anything else...u knw...
By clicks and pops I mean little "tick" kinda sounds which are sometimes present.
I don't have that bad noise..thank Almighty... But little hiss noise is there...
"If you're mixing with automation, you may not need to compress anything but the two mix (or 5.1)."-- I currently don't know about automation, its in my pipeline n have to learn about it. But I wud appreciate if u wanna throw some light on it... And by two mix, u mean stereo mix..?
Eq earlier, means before touching any dynamics..? I mean, I din get it...
U knw that's wats boggling me... I mean, I'm eager to know a procedure, like wen handed over a dialogue track then if there are problems in it then how to approach towards fixing them... I know nothing can beat a well studio recorded dialogue track; but I mean if there are some issues then how to approach them towards fixing them, like;
First of all check if there is some noise present in it then first the noise should be removed, if there's no noise then proceed further n check for the next thing.. Like for example, check if the track requires any normalization, if yes then do it & if not then leave it and move further ahead and check if there r any problems in the EQ, then check if the track requires any sort of compression... U knw...that's what I would love to know and learn... Its surprising to know there's no solution for this on the entire internet...! I searched a lot..some have a nearly similar question as of mine but...no solution, as in everyone says there's no gold standard recipe...
Anyways, we all aim for making whatever sounds we have, to sound as though it's never been touched...
I can't thank u enough for reverting back. But still looking forward for your help on my question...