|
|
|
03-27-2020, 07:28 AM
|
#1
|
Human being with feelings
Join Date: Mar 2017
Location: Berkeley, CA USA
Posts: 1,336
|
VOCALS: To Normalize or NOT to Normalize, that is the question....
I may have asked this a long, long time ago, but, I have seen videos on YouTube that advocate for normalizing vocal tracks, but, I have heard some people on this forum advocate against it and so I would appreciate a refresher of your opinions of the PROS and CONS of normalizing vocal tracks..
If it is a sometimes I should and sometimes NOT situation, is there any sort of criteria to use to make this determination ?
Thanks in advance....
__________________
Bayside Studios, Berkeley, CA - Music That Brings People Together
Steven Schuyler, Singer
|
|
|
03-27-2020, 08:20 AM
|
#2
|
Human being with feelings
Join Date: Jun 2013
Location: Krefeld, Germany
Posts: 14,686
|
Normalizing before any other activity makes things easir to look at. It does not affect the sound or the track at all.
-Michael
|
|
|
03-27-2020, 08:22 AM
|
#3
|
Human being with feelings
Join Date: Nov 2017
Posts: 1,575
|
Frankly - it doesn't matter.
Because if you normalize you don't loose information.
So you can do it - or leave it.
|
|
|
03-27-2020, 08:47 AM
|
#4
|
Human being with feelings
Join Date: Oct 2010
Location: Charleston, SC
Posts: 12,769
|
If I have a weak signal like a very light guitar player through microphones I'll normalize to try to get the gain up. Other than that I rarely normalize. If you do normalize and it's too loud just pull the item volume control down a bit.
|
|
|
03-27-2020, 08:57 AM
|
#5
|
Human being with feelings
Join Date: Mar 2017
Location: Berkeley, CA USA
Posts: 1,336
|
OK, based on all of your advice I think what I will do is copy the vocal track BEFORE normalizing so that I have it in its original form. Then normalize one of them and see the difference.
__________________
Bayside Studios, Berkeley, CA - Music That Brings People Together
Steven Schuyler, Singer
|
|
|
03-27-2020, 04:32 PM
|
#6
|
Human being with feelings
Join Date: Apr 2016
Posts: 653
|
Doesn't matter. Focus on getting the levels right from the get go.
|
|
|
03-27-2020, 04:37 PM
|
#7
|
Human being with feelings
Join Date: Jun 2019
Posts: 2,872
|
You will not notice any difference. if you level-match them and invert the phase of one they will cancel out. It's just convenience and workflow preference. I often prefer to normalize sounds so I can see the peaks well and afterwards reduce the take volume or just lower the mixer fader to get it right in the mix. Others may prefer to gain the item up until they work.
|
|
|
03-27-2020, 06:13 PM
|
#8
|
Human being with feelings
Join Date: Jul 2011
Location: Brooklyn
Posts: 2,613
|
Unless a level is way too low and you don't get what you need with the fader up there's not much point. Normalizing a track to close to 0 causes as much problems as it solves. One is likely to end up lowering the fader on it. Why not slap an eq or dynamics plugin on it and raise the output 10db?
I really dislike the idea of doing something as a matter of course before seeing how the track comes up in the mix, before anything has been worked with. If it turns out to need some major jacking up, then sure. To advise normalizing a track just because one can is pretty empty advice. It's a solution in search of a problem.
__________________
The reason rain dances work is because they don't stop dancing until it rains.
|
|
|
03-27-2020, 10:11 PM
|
#9
|
Human being with feelings
Join Date: Mar 2017
Location: Berkeley, CA USA
Posts: 1,336
|
Thanks, everyone for your input. I can pretty much hear that there is a consensus to not do it, but, that feature is there for the instances when it does become truly necessary.
__________________
Bayside Studios, Berkeley, CA - Music That Brings People Together
Steven Schuyler, Singer
|
|
|
03-28-2020, 12:59 AM
|
#10
|
Human being with feelings
Join Date: Jun 2019
Posts: 2,872
|
Quote:
Originally Posted by vdubreeze
It's a solution in search of a problem.
|
So what problems does it cause? The only one I can think of that isn’t solved with just turning the mixer fader down to the right level is feeding (some) plugins a signal which is hotter than expected but that is easily solved by reducing the item gain.
Anyway I agree, if you are able to record your tracks close to the desired level, it is easier to just boost or attenuate the gain a bit instead of normalizing it and bringing it down a lot.
I just do differently because I normally just record voice or sounds/foley and don’t bother with setting the perfect mic gain each time but if you record a musical performance it is good practise to record each instrument at the right level so that you have a good starting point for mixing.
There is no yes or no answer to normalizing like many people believe, it just comes down to knowing what you do and what you need. Just take care to feed plugins the right signal level and leave enough headroom for mastering (6-3dB usually, but that is debatable, most important is to never clip with all tracks toghethet). If in the process you do normalize or not is up to you, just keep in mind that if you normalize to 0dB you will have to lower the volume/gain afterwards to not clip when multiple sounds are playing.
|
|
|
03-28-2020, 02:09 AM
|
#11
|
Banned
Join Date: Jul 2006
Location: United Kingdom, T. Wells
Posts: 2,454
|
You can normalise to a certain -dBFS value and that is great... for organisational and visually convenient purposes.
In a pre-mixing production normalising the peaks (I repeat: the peaks) of a raw recording to a certain value (for me it is -6dbFS or -9dbFS) is helpful, depending on what you consider a "raw" recording.
Some people prefer taming down the peaks during recording either by external compressors/transformers (giving it "warmth" and that kind of sorcery) or the corresponding simulator plugins as inputFX in the DAW (compressors/tape-saturation/transformers/clippers plugins)... although latency might become an issue if you overdo it.
If you do it correctly, then you might not even need to normalise the "raw" recordings, because your levels will be right (the general rule would be: RMS at around -18dBFS (that is referred usually to 0VU), peaks should not go beyond 0dBFS or preferably not go beyond -3dBFS at max. for safety reasons to avoid digital clipping; peaks vary a lot though, when not tamed down with "gentle" compression, especially with highly dynamic instruments such as vocals and drums).
I do the latter (inputFX plugins in the DAW), because I do not have external equipment of such. For vocal recordings is good to have external tube-preamps though... or very good audio-interface (can be expensive in both cases).
If you do normalisation for the sake of it, then you also would like to perform the gain staging process ( -18dBFS RMS, peaks < 0dBFS) after the normalisation.
Most programs or processes can normalise the peaks generally (by default) to 0dBFS. If that is the case, then most likely you will need proper gain staging, described in short above ( -18dBFS RMS); this RMS value gives plenty of headroom for the peaks to live.
Anything between -23dBFS to -18dBFS for the RMS generally is considered to be healthy levels before you start mixing with plugins and sh¡t.
Then again in the mixing you would be compressing the peaks anyway in order to be able to raise the overall level of the song in the mastering/post-mixing phase of the production.
With external hardware (compressors, tape recorders, transformers, channel strips, tube pre-amps) you get the peaks tamed down (analogue "warmth") on the get go, when pushing the levels (VU is the analogue ancestor of RMS) to just above 0VU (in the "red zone").
In digital domain we do not need to push that much, because we do not get peaks compressed for free (as we get it for free with most analogue equipment mentioned above).
That's it. Hope this can help you decide what to do.
Last edited by Pashkuli; 03-28-2020 at 03:59 AM.
|
|
|
03-28-2020, 02:27 AM
|
#12
|
Human being with feelings
Join Date: Jun 2019
Posts: 2,872
|
@adXok: This! You couldn’t have given a more exact and exhaustive answer about the whole matter. Hope this clarifies the OPs doubts as well as those from people reasearching about this.
|
|
|
03-28-2020, 04:17 AM
|
#13
|
Banned
Join Date: Jul 2006
Location: United Kingdom, T. Wells
Posts: 2,454
|
Quote:
Originally Posted by sjs94704
Thanks, everyone for your input. I can pretty much hear that there is a consensus to not do it, but, that feature is there for the instances when it does become truly necessary.
|
You could also try out this extremely helpful plugin by TBProAduio.
Extremely helpful.
GainRider - Vocal level tutorial
Quote:
Originally Posted by TBProAudio
Maybe you try something like GainRider
We use it every day to compensate for the changing mic proximity.
|
|
|
|
03-28-2020, 06:41 AM
|
#14
|
Human being with feelings
Join Date: Feb 2009
Location: Stockholm, Sweden
Posts: 5,465
|
Correct me if I'm wrong, but I've heard that certain plugins work best at a certain input level. In other words, you can hit a plugin too hard (or too soft). Adjusting the channel fader won't help since the plug is pre-fader. What would help is a gain plugin inserted before the plug you want to use.
|
|
|
03-28-2020, 07:04 AM
|
#15
|
Human being with feelings
Join Date: Jun 2019
Posts: 2,872
|
Quote:
Originally Posted by Jorgen
Correct me if I'm wrong, but I've heard that certain plugins work best at a certain input level. In other words, you can hit a plugin too hard (or too soft). Adjusting the channel fader won't help since the plug is pre-fader. What would help is a gain plugin inserted before the plug you want to use.
|
True. Mainly emulations of analog gear which include emulating circuit overload. These usually expect 0VU which, as adXok said, is an RMS level of around -18dBFS. Obiously in those cases adjusting the fader won't help but you can, as you said, insert a gain plugin, or more simply just adjust the item gain to the desired level. Whether you have to bring it up to that level (or bring it down if you have normalized before) does not matter at all, as long as you are hitting the plugin with the expected input signal level.
|
|
|
03-28-2020, 07:14 AM
|
#16
|
Human being with feelings
Join Date: May 2009
Posts: 29,260
|
Quote:
Originally Posted by Jorgen
Correct me if I'm wrong, but I've heard that certain plugins work best at a certain input level. In other words, you can hit a plugin too hard (or too soft). Adjusting the channel fader won't help since the plug is pre-fader. What would help is a gain plugin inserted before the plug you want to use.
|
The only difference is you may need to adjust the input level on that plugin, which you would probably do anyway because it's a non-linear plugin and you are going to adjust based on your ears.
About the only place it is truly helpful for such a plugin, is if one is inexperienced and want's a close approximation of what a preset in that plugin was originally "supposed to sound like", that's about it.
A guitar and amp is the same thing, none of us really go online looking to find the exact right raw guitar volume to send to the amp, we just plug in, and twist knobs until it sounds good, this is no different.
__________________
Music is what feelings sound like.
|
|
|
03-28-2020, 08:29 AM
|
#17
|
Human being with feelings
Join Date: May 2009
Posts: 29,260
|
It's amazing the lengths one will go do take care of something so simple. If the recording level is nominal, it doesn't need anything done to it. If one is that worried about it, just normalize all the items to 0 dBFS then select all the items and turn the item gain down by -6 to -10 dBFS...
DONE.
I can appreciate the fad of all this peak, LUFs and RMS normalizing but that is far more for the newbie or possibly saving some additional volume/gain adjustment time, otherwise, there isn't going to be mixes that are "Wow, they must have normalized those tracks to X before they started", there just isn't. It's the same deal with that horrid "set my levels with pink noise" trick, it's just a trick when you don't have or use your ears!
Record and keep your tracks with that -12 to -6'ISH peaks which leaves room for additional processing and don't 'clip' and get on with life.
__________________
Music is what feelings sound like.
|
|
|
03-28-2020, 09:29 AM
|
#18
|
Human being with feelings
Join Date: Jul 2011
Location: Brooklyn
Posts: 2,613
|
Quote:
Originally Posted by Phazma
So what problems does it cause? The only one I can think of that isn’t solved with just turning the mixer fader down to the right level is feeding (some) plugins a signal which is hotter than expected but that is easily solved by reducing the item gain.
|
I didn't mean that it necessary causes a problem, just that it's applying a solution before one has seen that there's a problem on the track's level. There are the issues of it clipping things, so it just needs to be brought down in item db, as you mention, but to me it says more that it didn't need to be normalized than that an item can always be dealt with after normalizing.
I used to raise levels of tracks that I thought were too low, but once that was done, the master was often clipping, or peaking -2. Nothing terrible about the latter, but if letting it hit the master at -12 and adjusting its plugin thresholds and outputs to suit sounds the same, it just seemed to not be something I felt worth doing.
But I have other recording and mixing quirks that are far more pointless that I don't seem to let go of, so I get why some folks do it
Cheers
__________________
The reason rain dances work is because they don't stop dancing until it rains.
|
|
|
03-28-2020, 10:23 AM
|
#19
|
Banned
Join Date: Jul 2006
Location: United Kingdom, T. Wells
Posts: 2,454
|
Quote:
Originally Posted by karbomusic
I can appreciate the fad of all this peak, LUFs and RMS normalising but that is far more for the newbie or possibly saving some additional volume/gain adjustment time...
|
Well, in life there is not a good idea to only stick to newbie level.
And it is not some "secret society" knowledge anyway, although few people actually understand it and why it is like it is (mostly engineers whose last job would be to waste time on the forums explaining "basic sh¡t").
Gain staging also is some kind of "normalisation process" on the audio.
Also digital domain is fairly different from analogue domain in terms of physics (electricity) and as a workflow.
It would be useful, if we could use and understand both and quickly level up above beginner level.
Also, we've got a proverb: "Great preparation is half the work done." (includes both knowledge and having the right tools).
That is why in the old days this job (pre-mixing preparation) would've been delegated to the Assistant Audio Engineer. Should be the same today. Although nowadays we've got all the tools "in the box" and all the knowledge (both good or ambiguous) on the Internet.
Last edited by Pashkuli; 03-28-2020 at 10:38 AM.
|
|
|
03-28-2020, 10:43 AM
|
#20
|
Human being with feelings
Join Date: Jun 2019
Posts: 2,872
|
Quote:
Originally Posted by vdubreeze
But I have other recording and mixing quirks that are far more pointless that I don't seem to let go of, so I get why some folks do it
|
Yes I agree with you, such things like normalising often just add a second step like having to reduce gain afterwards, while it would have been fine from the get go. But I guess different working situations just lead to different working habits, that in many situations are rather compulsive than neccessary. As I am working a lot with samples which often come normalized and export my own samples normalized, I just developed the habit of normalizing so I don't have to keep the listening volume high and jump off my seat when I accidentally trigger such a normalized sound. If however you are not working with single sounds but stems that have to be mixed to a record then it makes no sense to normalize carefully gained tracks just to turn them down again to where they were before.
In the end everyone can work the way he wants, as long as there is a healthy understanding of how digital audio functions and how to deal with levels in a digital environment. But in case the OP didn't have this understanding, now it seems to me there is all necessary information in this thread.
I just cared about pointing out that normalizing is no "sin" hiding a complex algorithm underneath that destroys all audio as some people like to perceive it. I often read people reacting to a question about "normalizing" as if someone asked "Should I put a brickwall limiter with +24dB gain on each track" without realising that it is just raising volume to a certain predefined point without any alteration in sound and the question should not be to normalize or not but to raise volume or not and the answer to that question obviously depends on the situation.
|
|
|
03-28-2020, 10:46 AM
|
#21
|
Human being with feelings
Join Date: May 2009
Posts: 29,260
|
Quote:
Originally Posted by adXok
Also, we've got a proverb: "Great preparation is half the work done." (includes both knowledge and having the right tools).
|
Absolutely and I'm all for it; especially where organization and efficiency are concerned.
But threads like this aren't usually started from that foundation so the basic answer is the same as it always has been, don't clip, leave some room for additional processing. Go about that however one wants at the beginning of the chain... normalize, raise item gain whathaveyou, or at the source when you record it if you are the one recording it.
Obviously it's fine to go into the academic and explain the whys, hows, what ifs, but the actual goal is still the same.
__________________
Music is what feelings sound like.
|
|
|
03-28-2020, 12:47 PM
|
#22
|
Banned
Join Date: Jul 2006
Location: United Kingdom, T. Wells
Posts: 2,454
|
^ True indeed.
Actually in the recent years many people (home/bedroom producers) have been adopting the idea of top-down mixing.
Really interesting idea, because if it has been prepared and done properly, could save so much hassle. What do I mean.
Maybe a video will spare me the details...
Basically it describes the usual suspects that would go in an analogue world to accomplish a recording/mixing (mastering is dependant on the final application); recording - phase I:
1. Signal goes to a transformer, then goes to a
2. mic pre-amp (saturation), then
3. compressor, then to a
4. recording channel and finally to a
5. gets actually recorded on recording (magnetic) tape
Then the second mixing - phase II, would be roughly:
1. Mixer channel (saturation of some sort)
2. FX, compressors, etc.
3. Bus FX (tube or transistor based)
4. Master channel (compression/limiter)
5. the mix gets recorded on a Master (magnetic) Tape or also on DAT (digital audio tape)
-----------------------
So, the idea is to have a simplified version of these two chain sets in the DAW, in the Master track or some Mix Bus as plugins in this order:
1. Transformer
2. Compressor
3. Saturator/"analogue warmth" branded sh¡t
4. Clipper (soft limiter)
5. Tape simulator/Saturator
Limiter6 Limiter usually has some of these combined in one interface. Very convenient for the job, extremely recommended (if done with vigilance and caution/understanding).
Limiter6 - basic video demonstration
Needs maybe "tape" warmth simulator as a second plugin after it and True Iron (by Kazrog) before it. Plenty of choice nowadays...
The idea is not to worry too much about the peaks in most scenarios (like they didn't in the analogue era) and leave it to your ears for both the levels and the actual mix with spacial FX and EQs. Of course this setup should be performed with chain presets and compared before using some of them depending on the genre and material.
Of course that could be performed in the DAW at the two separate stages:
(recording) - stage 1 chain preset I - export all tracks
(mixing) - stage 2 chain preset II - export all stems
Save them as templates (start/blank) projects and forget about worrying too much.
Then mastering to suit the need.
Last edited by Pashkuli; 04-01-2020 at 03:40 PM.
|
|
|
03-28-2020, 10:15 PM
|
#23
|
Human being with feelings
Join Date: Jun 2013
Location: Krefeld, Germany
Posts: 14,686
|
Quote:
Originally Posted by Jorgen
Correct me if I'm wrong, but I've heard that certain plugins work best at a certain input level.
|
That is certainly correct. But I would consider designs that relay on input volume and don't provide an input level fader to be buggy.
Anyway it's recommended to do the normalizing before tying any plugins.
-Michael
|
|
|
Thread Tools |
|
Display Modes |
Linear Mode
|
Posting Rules
|
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts
HTML code is Off
|
|
|
All times are GMT -7. The time now is 03:41 AM.
|