If I create a drum sample pack, what volume level should the samples be ?

[MasterTuner]

I'm creating a drum sample pack.
Right now I am normalizing all the samples, kick, snare etc., to about -1 db.

Is this Ok, or should they be lower?

[Scrubbing Monkeys]

I would think lower around -6db peak maybe push it to -3db.
My thought behind this is that on a 12 plus channel mix if I gain stage my instrument tracks to -18db peak and my percussion to - 6db the faders can all rest at unity without clipping the master. Then I can mix by reducing or boosting slightly at the point where the fader is the most detailed. If my drum sample is peaking near 0db I simply use clip gain to turn it down 6db. Headroom is a good thing. YMMV

But it does seem every one is slamming stuff to the ceiling...Nearly all vst synth patches come in at near .0db.....you can't mix that...first move....turn it down......annoying.

[yellowmix]

As long as they are consistent with each other. Important part is having as much true dynamic range as possible, then normalize factoring in loudness. So when people audition the pack one sample after the other it doesn't vary too much. And if there are similar instruments like all kicks, they can be A/B fairly. The thing is, people can be dragging them directly onto a timeline or loading them into a sampler/drum rack so there's compromise. Loudness is hard to do with such short samples so do it by ear.

And yeah, when synth presets are slammed to 0 db it's annoying. I generally put my own presets at -18db since I can leave the track at unity without gain staging. But if I share them with others then I look at the factory patches and try to match that. Consistency is more important than anything else in this regard. Very few synth factory patches are inconsistent and it's annoying when you come across it.

[enroe]

I'm creating a drum sample pack.
Right now I am normalizing all the samples, kick, snare etc., to about -1 db.

Is this Ok, or should they be lower?

Why not normalize to -0.3 dB?

The wavs should all be normalized to maximum.
The volume-adaption is usually done in the playback definition.

[Scrubbing Monkeys]

Yeah after rethinking this .....a sample is bit different than a Preset. I can see where maximum dynamic range would be better. So...I like the 0.3db also.

[imrae]

If the samples are 24-bit you really don't need to use the full bit-depth, it's ok to leave some headroom.

Being able to flip through samples in the same pack without adjusting the loudness for each individual sample is good for workflow. Peaks are not a good indicator of loudness, especially for drum/percussion samples.

[MasterTuner]

thanks all
I just had a look at a drum sample pack I bought a little while ago. The kicks, snares and even hats etc. seem to be normalized at 0 db.

[imrae]

Do you find that useful or are the perceived-volume inconsistencies annoying?

[Scrubbing Monkeys]

Do you find that useful or are the perceived-volume inconsistencies annoying?

I would think that a sample pack normalized to 0db peak would be full of inconsistency. A deep snare with lots of body would seem louder than a piccolo rimshot for example.....I can only imagine the kick.

Some one above mentioned using your ears....maybe the best approach.

[jancivil]

"A deep snare with lots of body would seem louder than a piccolo rimshot for example"
This is very true. But: The expectation of consistency may be a thing in "samples packs", recordings of/production of drums is another whole ballgame, with problem-solving needed at a level an EDM person may not be into.

So you can canvass users for the answer, but consensus gon' be hard to come by. Some apparently don't want to bring faders down, some may find things too weak-sounding out of the box.

[Scrubbing Monkeys]

"A deep snare with lots of body would seem louder than a piccolo rimshot for example"
This is very true. But: The expectation of consistency may be a thing in "samples packs", recordings of/production of drums is another whole ballgame, with problem-solving needed at a level an EDM person may not be into.

So you can canvass users for the answer, but consensus gon' be hard to come by. Some apparently don't want to bring faders down, some may find things too weak-sounding out of the box.

Kkkkkk ..... All so true.

[imrae]

You could lowpass ambient train tunnel noises at 100Hz and there would still be someone here complaining that they sound "thin".

[jancivil]

arf

Kids are so spoiled these days

[JeffLearman]

Consistency is the key, but as said above, volume is more important than peak, yet volume is hard to judge.

If using peaks, the difference between -1dB and -0.3dB is insignificant, so take your pick. (Ditto if using measured volume, somehow. If you find a good way to measure volume on drums, please share!)

Peaking at -6dB you're only sacrificing 1 of your 20 significant bits, and then only if your recordings are literally perfect. In reality, you're unlikely to have more than 16-18 significant bits, so you lose no fidelity, though the waveforms won't look as pretty.

Reserving "headroom" for the end user assumes you have a clue what the end user wants. I think it's a hopeless exercise. Note that it's not really headroom, it's just planning where you think they want most of their faders to sit in some imaginary ideal mix. Good luck with that. Frankly, there are controls for that anyway.

[lpr]

Ideal