VST only vocal channel setup possible?

[garyb]

[dawman]

The last PrimmaDonna I worked with kept taking the windscreen off of my rented U80 and Chandler Mic Pre as if it had an effect on the volume of her voice. After putting it back on twice I had to unplug the mic and show her that the nasty pulp she was emmiting came from her Screwdrivers she was chugging all night.....The fresh squeezed Orange Juice was a major hassle with this foul swine.
If she didn't have such a fine backside and chi-chi's, there would have been a bounty on her head.

[Cochise]

IMHO, if vocal dynamics extend from whispers to screams, some compression could be needed.
It even depends on mic proximity effect kind and amount and on the ability of the singer using it.

[garyb]

yeah, very artistic.

to every rule, there is an exception. in recording there are very few rules...

[Cochise]

Music and audio practise (...and theory of course, above all), lately have ended up occupying a very little amount of my time.
It's since a couple of weeks I feel the need of recall back and reinforce that some of basic knowledge I had about, with the aim of disperse the thin mist in which It looks like I move, when it comes the time of making music.
Actually, I started putting some order in physical and virtual stuff in my room and hdds; then I went on revising basic concepts like equations and logarithms, electrotechnical priciples, transfer functions... At now I'm back on Audio Technical basics; in these days i was on the specific subjects of Operating Levels and Dynamics. While on these subjects, my mind was drawn back to this thread on PlanetZ, since here it was the last time I discussed these topics.
Ken, I'm sure at now you solved the issue, with the advices here, gathering other informations, ...with some trial even, and above all maybe
Nevertheless, in order to rectify something I said here, as well as in order to share my attempt to increase confidence with these things and better fix them in memory, (even in order to take a breath, a break, from formulas, theories and experimentation), I'm gonna post a speadsheet I just made about dynamic units, values, ranges and compression.
The second-last table calculates the out values for a compressor at given threeshold/ratio; it doesn't include the make-up gain ...but it should be a simple sum to the out values.
The last table is kind of a calculation tool for the compression/limiting of dynamic ranges: at a given amount of attenuation it returns compression ratios at relative threshold levels ...to experiment with.

dynamics.zip

(27.88 KiB) Downloaded 57 times

In your specific case,

oops, sorry, this is for live. Ya, for recording I wouldn't mess with the signal on the way in.. Just for live I want to do a bit of compressing to keep the vocals from bopping in and out of the mix when the vocalist's alignment changes.

the second-last table should had been more usefull in order to display the corresponding level difference between quiet and loud peaks, at in and out; I'm thinking about kind of a specific algorithm for this, since various values of ratio have to be tried to find the right values at the output... but...it was at first intended just to better familiarize with basic compressors behaviour... numerically, of course; the audible result of different compression ratios and threesholds is another subject: transients and attack time, coloration, release time...

At Garyb:
I think the need of compression for vocals should further be discussed.

DynamicRanges.jpg (34.85 KiB) Viewed 943 times

Besides, the performance itself can be quite different in presence of processing...

[garyb]

pop singers aren't that dynamic, and if they are, they should know how to use a mic. it all depends. however 24bit digital recordings have plenty of headroom for most sources even though one will want to limit dynamics for the purpose of playback(the mic pre has plenty of dynamic range and so does a good 24bit ad/da). the best place to limit is in mixdown, not recording, but if you need a compressor because you're recording a major dramatic event or because the singer sucks, go ahead and use one.

[astroman]

it just entered my mind while looking at the last table...
As dB is a logarithmic scale the proportions are rather unfortunate, but that's just a minor.
The funny thing is that we're pleased to be fooled time and again by this huge 'processing' range of the ear.
Actually it's complete BS or does anyone actually hear a 12dB whisper right after standing next to an aircraft engine
We need time to recover or 'adjustment', just like any amp

cheers, Tom
(strange that I never noticed this before)

[erminardi]

http://www.izotope.com/products/audio/nectar/

this is new and seems to be very interesting: all in one and with "mixing" or "tracking" switche that solves (I hope) the problem of delay in realtime.
The quality is iZotope... so seems promising.

[garyb]

if you NEED dynamics control while tracking, you need it BEFORE the ad/da....

[garyb]

it just entered my mind while looking at the last table...
As dB is a logarithmic scale the proportions are rather unfortunate, but that's just a minor.
The funny thing is that we're pleased to be fooled time and again by this huge 'processing' range of the ear.
Actually it's complete BS or does anyone actually hear a 12dB whisper right after standing next to an aircraft engine
We need time to recover or 'adjustment', just like any amp

cheers, Tom
(strange that I never noticed this before)

all excellent points, yes that graph should look crazier, and no, no transducer is expect to use 140bd of dynamic range in any single event, not even the human ear. obviously, the noise floor in the head is very low and theres some kinda noise cancellation network involved for bloodflow...

[erminardi]

if you NEED dynamics control while tracking, you need it BEFORE the ad/da....

obviously
every little studio should invest over a decent mic-pre-eq-compressor vocal/instrument rig.

[tgstgs]

@cochise

i see lots of work in it_
why dont you make the same you did with excel in sdk?
makes more fun to listen to the results than to watch it;

good vibes

[Cochise]

pop singers aren't that dynamic, and if they are, they should know how to use a mic. it all depends. however 24bit digital recordings have plenty of headroom for most sources even though one will want to limit dynamics for the purpose of playback(the mic pre has plenty of dynamic range and so does a good 24bit ad/da). the best place to limit is in mixdown, not recording, but if you need a compressor because you're recording a major dramatic event or because the singer sucks, go ahead and use one.

Input levels should be set across the S.O.L. to work in the best response zone of analog input devices.
-10dBV devices usually have got 10 dB headroom only, above S.O.L.

DynamicRanges(electrical).JPG (261.98 KiB) Viewed 884 times

As for compression used with the aim of improve intilligiblity of sung words, or to correct misuse of (probably kind of too hypercardiod?) mics, or whatever: post-processing can bring a flattening of musical accents.
In pre-processing the singer could adapt the performance to the response of signal chain as every musician does playing different instruments or different processing of the same instrument...

[Cochise]

it just entered my mind while looking at the last table...
As dB is a logarithmic scale the proportions are rather unfortunate, but that's just a minor.
The funny thing is that we're pleased to be fooled time and again by this huge 'processing' range of the ear.
Actually it's complete BS or does anyone actually hear a 12dB whisper right after standing next to an aircraft engine
We need time to recover or 'adjustment', just like any amp

cheers, Tom
(strange that I never noticed this before)

Of course, that's right.
Besides, a classical music "fortissimo" should be around 95 dBspl. Almost every 24 bit AD converter could handle that as peak value, even setting dBspl=dBfs...

That table is just taken from the "basic knowledge" section of the user manual of a device from a manufacturer of audio equipment producing hardware compressors, limiters, expanders, gates, among the others...
I could probably post the link to the whole document...

[Cochise]

@cochise

i see lots of work in it_
why dont you make the same you did with excel in sdk?
makes more fun to listen to the results than to watch it;

good vibes

Thanks for the encouragment

Probably i'll try, sooner or later.

Maybe there's a reason why I made it in excell though. Probably it's why when I listen at sounds, I usually become too focused on them, losing sight of values

[garyb]

if you NEED dynamics control while tracking, you need it BEFORE the ad/da....

obviously
every little studio should invest over a decent mic-pre-eq-compressor vocal/instrument rig.

so how does a vst plugin help when tracking?

do what works for you, but remember, whatever processing is done in tracking can't be undone EVER. one should always use the least amount of gear in tracking as possible, the right mic, through the right pre adding only gain, through a quality ad converter or straight to tape is always the best solution, now as well as 50 years ago. if a limiter is needed because a vocalist doesn't know how to control him/herself or use a mic, then so be it, but i wouldn't just compress automatically.

[Cochise]

@cochise


why dont you make the same you did with excel in sdk?

If I should really be good enough for software development, I should try to realize a peak meter
with minimum values retention, besides max values. It could be useful, sometimes...

ADDED:
Ok, I'm messing with it now. Actually it makes no sense to get minimum values (unless you're setting the threshold for a noise gate), and average values are already displayed by virtual leds in Scope meters...

[Cochise]

...whatever processing is done in tracking can't be undone EVER...

Unless two channels/tracks are used, one for the compressed signal, one for the dry one...
This could further mess the things, though.

[valis]

Methinks you're overthinking. If you need compression to tame levels while recording then adjust mic placement & signal chain (impedence mismatches between gear can do funny things to 'reactance' to AC levels...sometimes) or get an inexpensive compressor and set it to just tame the worst peaks. FMR 1137 or JoeMeek (classic, not new) mono 'green' units are affordable and work fine.

Software can be useful for a 'software monitoring' recording chain, but what GaryB means is that if you're compressing post-conversion you're not gaining *anything* while tracking for the actual signal you're recording since any damage that occurs will be at the point of conversion. If the issue is simply having something 'hot enough' for the singer/vocalist/etc to monitor themselves then a software monitoring chain is the solution. If it's live sound the same might be said but I wouldn't trust a plugin stack for live use on a vocalist over a moderate investment in a vocal chain (gate/comp/eq or etc) in hardware.

[Cochise]

Right. Facing real level ranges of a vocal signal at AD input, it's possible to avoid clipping without setting the average level too much below the S.O.L., even if the S.O.L. for the AD interface is -10 dBV and the max input is 0 dBV; then compression isn't strictly needed for tracking vocals; nor for live apps... not with limiting purpose at least.
Sometimes, listening to Skunk Anansie music, Skin's voice passing from murmur to scream make me think she's could be hard to track without compression, but I'm not in confidence with real levels of a voice signal...
As a musician I shouldn't fear the take though. Probably i'm in mistake twice: having few live experience and not so much in recording, I anyway fear the take as a musician, but I don't fear it as an audio technician...