Drum kit recording/mixing tips
This is my way of drum recording. Others may have different opinion.
Last Update: Dec. 7, 2014
1. Recording
Note: Margin for peak -6dBFS can
be between -6dBFS and -18dBFS, or lower. It is up to you.
- Format: Use 24bit if possible, sampling rate 44kHz for CD, 48kHz for video
- Recording(input) level adjustment
-- peak at -6dBFS(0.5 in linear ratio) or lower.
If your recorder has peak LED, read the manual and find
the threshold dbFS for the peak LED. Common setting is
"Set the input gain at peak LED ocasinally and turn down
the gain until LED never blinks."
- Miking, Tuning drum kit
- Recording Test: verify the peak level of each track is just
below -6dBFS(margin) for entire duration.
If possible, with sox command. (adjust input level if necessary.)
- Sample mix with PAN, but without any EQ, effect
Panning -- drumers view or audience view
If the sample mix is not satisfactory, check micing and tuning.
If strange vibration is heard, remove the head and tighten screws.
2. Mixing Drums
Usually gate and comp., eq. are used in this order, but
other orders can be possible
and it may be recording engineers secret.
In my system, hardware gate and comp are inserted before
the recorder, and, therefore, applied in recording.
- Gate -- optionally used for better separation
in multi-track recording
Gate threshold level depends on input gain. Do not overdo.
Applied: toms, kick
snare top and bottom (2ch paired)
Not applied: overhead stereo mic
- Compression -- start with 2:1(natural) (5:1(medium)、8:1(high))
The threshold is lower, the effect is stronger with average output reduction.
Do not overdo.
Right threshold level depends on the recording level(input gain).
Trust your ears.
- EQ -- roughly adjust for drum kit
Overhead: almost cut low at 100--150Hz (low freq is not used)
Kick, F. Tom: reduce low at 40Hz (My bass drum has KickPort)
Other toms: reduce low at 80Hz--100Hz (when boomy with large speaker)
Snare top: reduce low at 80Hz -- 150Hz (with your choice)
Snare bottom: almost cut low freq at 100Hz -- 200Hz (low freq is not used)
(My rock kit is tuned quite low: Toms 13,14,16, bass 22, 6.5 depth snare
with very thick heads.)
In general, drum sound has large power in low frequency range
and it may occupy the dynamic range of a stereo mix.
So, I reduce low freq to make the final mix louder.
- Pan(LR) -- Drummer's view or audience's view
Audience's view for right-handed drummer
Left Center Right
Overhead(L) Overhead(R)
F.tom M.tom Kick H.tom(or far right) -- it is up to you
Snare
In this settings, probably hi-hat is rather from right and
ride is from left. It depends on the hight and angle of overhead stereo mics.
NOTE: In old commercial record, all drum sound is coming from left or right.
But It sounds good. An example is Super Stition by Beck, Bogart and Apice.
Gate threshold adjustment (in my case)
If gate can be adjusted after multi-track recording, it is easy and quick.
In my recording system, gate is inserted before the recorder and
I have to adjust the threshold at recording.
- (Tune each drum.)
- (Adjust the input gain of each mic for your drum kit.)
- Record sample tracks:
Use all tracks but play slowly each drum separately for 30 sec or more
with weak to strongest hits.
- Bass drum(easiest): Playback bass drum track only, and
find a nice threshold value where sound from other drums can be hardly heard
AND weakest bass drum hit can be heard.
- Floor tom: Similary playback floor tom sound only.
Floor tom is rather close to bass drum.
- Mid-tom: Close to hi-tom. Do not try to completely eliminate
hi-tom sound leak.
- Hi-tom: Close to mid-tom and snare. Do not try to completely
eliminate snare sound. Muffling the bottom head will
reduce unplesant resonance of snare sound.
- Snare(important): Similar to Hi-tom but be careful to not lose sustain.
Set paired gate for snare top and bottom if the functionality is available.
- Record usual drumming play and fine tune the thresholds.
Note: Sustain will be lost in some extent with gate effect. It can be
compensated by simulating sustain with reverb effect.
Manual gating (or mute) can be possible for toms with DTM software.
I mean you can mute tom sounds in the DTM when not used.
3. Mixing with other tracks
Level balance, EQ adjustment, reverb
Trial and error: listen by(with?) big speakers, car stereo, head phone etc.
My theory: if it sounds good by a decent car stereo, it is good by most
music players.
Pan(LR), volume balance, and EQ fine adjustment
- kick + bass guitar -- reference level (usally no reverb)
- +all drums (reverb)
- +guitars etc.
- +vocal -- minimum level is at hearing lyrics clearly
- +guitar solos -- compare with vocal
4. Maximize and 16bit output
You may omit this process or do it at the same time with Mixing.
If you want to make a wav file for CD, you need to convert the mix
into a 16bit, 44.1kHz wav file.
Maximizer may be included in your DTM or add-on(plug-in).
Also limiter can be used after maximizer as an insurance
or aggresively raise loudness with a little bit of distortion.
Sox command has multiband-compression but it is difficult to use
because there are so many parameters to set and I should understand
the theory completely.
I use ZOOM R16(R24) built-in maximizer with custom parameters.
Also maximizer in ZOOM HandyShare (free software) is very easy
to use, nothing to be adjusted.
Just choose "maximize"(decent) or "ultra maximize"(quite strong).
On Youtube, 48kHz sound in an uploaded video file seems
automatically converted to 44.1kHz.
24 bit samples was not accepted in my case.
Links
- Lesson/Guide
- Miking:
- Input level
- Gating, compression, EQ, effects, mixing
- Low freq. (Kick/Bass)
- Maximizing
- Monitors (Test tones)
Theoretical background for dBFS
0dBFS -- absolute max for digital recording (FS: Full Scale)
Absolute value of 16bit signed integer 215 - 1 = 32767(max)
for 24bit 223 - 1 = 8388607(max)
(Note: 1bit is for sign (+ or -))
24bit recording is recommended in most case to not lose sound wave
details when large margin is applied.
dB = 20 log10 (abs.value/max)
-3dB(FS) 10-3/20 = 0.7079... (71% of 0dBFS)
safe only if it is the peak in whole track
but we do not know the peak until when we stop the recorder
-6dB(FS) 10-6/20 = 0.5012... (50%) safe
-12dB(FS) 10-12/20 = 0.2512..(25%) very safe
-18dB(FS) 10-18/20 = 0.1259..(13%) very very safe
Analog meter shows average: 0VU = -18dBFS
(appropreate average level for digital)
To view the peak level of a track sox command can be used
(available for Windows, Linux, MacOS etc.)
sox input.wav -n stat
Example:
prompt % sox AmericanGirl.wav -n stat
The Last of American Girl/Green Day -- very loud music
Another loud example in my CD collection is Scream/Ozzy Osbourne.
Samples read: 20376552
Length (seconds): 231.026667
Scaled by: 2147483647.0
Maximum amplitude: 0.999969 --- peak is almost 1.0
Minimum amplitude: -1.000000 --- absolute value is same as Max
Midline amplitude: -0.000015
Mean norm: 0.216091
Mean amplitude: 0.000139
RMS amplitude: 0.289337
NOTE: In a final mix of commercial pop/rock CD, the peak is almost 0dBFS
using equalizer, multi-band compressor, peak limiter, analog equipments etc.
at the mastering stage. Even distortion can be heard.
goto at kmgoto.jp
Feel free to use/modify the contents.