Recording Dhols
Originally published at The Null Device Blog. You can comment here or there.
This sunday, I made what turned out to be a pretty respectable set of attempts at recording some dhols. Not perfect, but it was certainly a step in the right direction. There’s scant little information on dealing with what is, to most western engineers, a fairly esoteric instrument, and what info there is is mostly british kids making forum posts of varying accuracy in a mix of panjabi, south london slang and 1337-speke. There are one or two sites that have great information about micing a dhol for live use - a few major instrument retailers will go so far as to mount an SM-57 inside the drum - although that strikes me as a method that wouldn’t apply well to studio situations.
Attempt 1:
Mics: 2 SM57’s
Setting them up on two stands, I recorded a stereo track with an SM-57 on each end. I set the bass side a bit further away than the treble, to allow the bass waveform to build.
Result: Decent levels, but a bit of muddiness on the treble head. Also, while I caught the initial attack of the bass head, the level dropped off quickly enough that it fell through the 57’s noise floor at the distance. It was more of a thump than a boom.
Attempt 2:
Mics: SM57, Rode NT2
Pulling out my trusty and well-loved NT2 large diaphragm condensor, I used that on the treble head instead of the 57.
Result: The sound was clean and ringy - the NT2 has a bit of a “hyped” high midrange, specifically for vocals, but given the tonal characteristics of the treble side, I figured it’d work. The NT2 is a MUCH hotter mic than the 57, though, and required an awful lot of gain adjustment. It’s also sensitive enough that it picked up a fair chunk of crosstalk from the bass head. I would imagine any decent “vocal” condenser mic would work well for this - something like an SM-78 might be better, in fact, given it has a tighter cardioid pattern than a large-diaphragm mic, and would likely be less susceptable to crosstalk.
Attempt 3:
Mics: SM57, Rode NT2
For giggles, I swapped configurations from attempt 2.
Result: nice resonant bass boom, muddy treble side, and a lot of crosstalk on the bass side from the treble side. The bass boom was actually quite resonant and fairly impressive, but was so strong that it became clear quickly that I would never get this to sit in a mix.
So, attempt 2 was the clear winner. I may try this experiment again with cleaner preamps at some point - my 828’s pres are workable but not great, and the dbx pre I’ve got has great, transparent sound (unless I crank the tube up) but is only a single channel and thus doesn’t work so well for stereo recording.
Recording was done in a pretty dry room (namely, my studio) so I had very little room sound. This I consider to be entirely acceptable, as I don’t have a good room for recording room sound anyway, and prefer to go with a decent external reverb. I mixed down the stereo width (having recorded this as a stereo pair) to something a little more realistic - the dhol heads are not split that far apart, after all - then applied a little light compression. The overall effect was pretty solid, and quite usable, although it still wasn’t quite right. My next trick was to split the stereo pair into individual mono tracks and process the EQ separately. This allowed me to control some of the boxiness on the bass side and pull out the low-end crosstalk on the treble side, and with a little gating, control some of the ringy tail of the treble head for purposes of mixing.
Of course, after doing this a few times, I came to the conclusion that the wall-of-dhols didn’t realy work especially well in this track, so it was essentially all-for-nought from a musical standpoint. It did give me some good insight into recording dual-headed drums, though.
It also proved that I’m a pretty lousy dholi at this point. I had maybe 4 salvageable measures of various beats in each take, despite recording about 32 bars at a pass. I also discovered that even with headphones on, the dhol tends to drown out my click track rather impressively.