Solution for Peak limiting BEFORE sound card - do you?

[CreamWare4Ever]

Hi people!

Just to be short - I'm exploring different solutions for killing transients over 0dB that occurs on irregular basis when recording into A16. Mostly for vocals.

As I have decent mic preamps it is necessary to give them some gain so that they can deliver that 'mmmm...' sound. So I keep the input levels at aprx. -10dB <> -6dB into A16U. But despite that, a couple times per session the red lights flashes on A16U and that awful digital clipping sound gets recorded.

I know that some people use just a compressor with 0dB as threshold and infinity as compress ratio to 'kill' all peaks over 0dB on the way to A16U. I'm also interested in limiter units as such.

I read somewhere that DBX 166XL can successfully be used for limiting as they have dedicated 'PeakStop Limiter' that can be set at 0dBu.

I will appreciate any inputs.

[the19thbear]

There is no need for running your levels hot. thats a leftover from tape days, when you had to drive the tape pretty hard... wel not had too.. but just to keep noise levels down.

In digital you dont have to be near the red all the time - i never get clipping, becuase i just keep my levels low.
Before i start tracking i have the artist sing/play the loudest passage and adjust for that. another option is two mics/mic pres. one for low input one for high input...

[CreamWare4Ever]

There is no need for running your levels hot. thats a leftover from tape days, when you had to drive the tape pretty hard... wel not had too.. but just to keep noise levels down.

Thanks for reply. OK what is your suggested levels? For me personally my 'average' load in wave files images is about -10dB to -6dB. But still there is some peaks. I don't think that it is a good solution to lower the input gains so that signals drop to between -20dB and -14dB just because of irregular digital clipping. After all, I want to take the benefit of using hi end pro gear

But maybe that is 'standard' levels that people are using!?

[spacef]

my suggestion is to determine what is causing the peak, and to lower its volume by any mean.
If you have such differences in levels that your sound is good but there is a sudden peak, it means the mix should be improved and that it cannot be mastered in the best conditions.
you could record a wave file to see this graphically and not miss any peaks.
moreover, i would not peak at 0.0 dB, but at -0.1 dB (as seen on modular mixer's peak meters, when using optimaster for example). For a master stereo file, I would in fact use -0.01 dB normalization, which translates to 0.0 dB in most displays, but we know that it never goes by 0.0dB because it is digitally peaked at -0.01dB .
when playing a stereo file through scope, levels are generally at 0.0dB gain everywhere unless there is an equalizer without a gain control (but even in that case, you lower the volume in your DAW/Wave editor, rather than in Scope). With all faders at 0.0dB in scope, if it peaks in/out Scope, it is the player that is peaking (daw/wave editor/other), unless you use plugins that change the levels of course (in such case, levels should be ajusted inside the plugin if possible).

[garyb]

If you record in 24bit, halfway up the meter is about 12 bits. Since 16bit is a cd maximum bit depth, 12bit can be seen as a pretty good level. I like to get about half to a third up the meter average. Peacks are usually within tolerances and I get the sound of the source. The digital noise floor is quite low. I always hear about how "analog" my mixes sound. The trick is not a hot input level. It's true that the technique is a tape trick. It's much better to add compression and distortion afterwards when you have control in the mix with digital. If your levels are as I describe, then yes, a 166 might work well without being obvious,but you likely won't need it.

[spacef]

i'm recording at -12 / -6 db (max) thru a joemeek compressor/eq, for vocals it works pretty well with a tube mic. or you lower the volumes even more if you don't like compressors on recordings. you add gain afterward and cut some peaks in the wave to lower some levels individually if needed (yeah sorry for the other post lol) - but i also keep my faders at 0.0dB gain when recording personnally, gains are adjusted on the preamps (inputs and outputs, with compression on high thresholds (like a limiter).

[CreamWare4Ever]

my suggestion is to determine what is causing the peak, and to lower its volume by any mean.

The peak is causing by the singer, the signal that comes out from mic preamp. True, I can lower output from mic pre, but there is some 'wall' in the head that I have which makes me very uncomfortable knowing that I can record vocals around -20dB to -16dB!?

If you have such differences in levels that your sound is good but there is a sudden peak, it means the mix should be improved and that it cannot be mastered in the best conditions.

My 'problem' is related to recording process, during recording sessions, not the mixing & mastering.

you could record a wave file to see this graphically and not miss any peaks.

Done that. Beside that I clearly hear every time A16U leds flashes red.

moreover, i would not peak at 0.0 dB, but at -0.1 dB (as seen on modular mixer's peak meters, when using optimaster for example). For a master stereo file, I would in fact use -0.01 dB normalization, which translates to 0.0 dB in most displays, but we know that it never goes by 0.0dB because it is digitally peaked at -0.01dB .
when playing a stereo file through scope, levels are generally at 0.0dB gain everywhere unless there is an equalizer without a gain control (but even in that case, you lower the volume in your DAW/Wave editor, rather than in Scope). With all faders at 0.0dB in scope, if it peaks in/out Scope, it is the player that is peaking (daw/wave editor/other), unless you use plugins that change the levels of course (in such case, levels should be ajusted inside the plugin if possible).

Killer good tip! This should be sticky level thread! Thanks!

[CreamWare4Ever]

If you record in 24bit, halfway up the meter is about 12 bits. Since 16bit is a cd maximum bit depth, 12bit can be seen as a pretty good level. I like to get about half to a third up the meter average.

Well isn't that around -12dB to -8dB?

Peacks are usually within tolerances and I get the sound of the source. The digital noise floor is quite low. I always hear about how "analog" my mixes sound. The trick is not a hot input level.

I often hear that statement, I'm just wondering how much (little) do you mean exactly by 'not SO hot' input level? Please give me some hint there, I'm sure that community will appreciate that.

It's true that the technique is a tape trick. It's much better to add compression and distortion afterwards when you have control in the mix with digital. If your levels are as I describe, then yes, a 166 might work well without being obvious,but you likely won't need it.

Agree 100%. I newer use destructing sound processing during recording.

[CreamWare4Ever]

Before i start tracking i have the artist sing/play the loudest passage and adjust for that. another option is two mics/mic pres. one for low input one for high input...

Well I want to (try) to keep things simple. The goal is to set and forget input settings so that they can be used for whole project with a specific artist. It is much more easier to mix afterwards because I know that I have a same source gain on all songs for that project (...).