When we went to put a final polish on the music we’d just recorded for our third CD, to our horror we saw we’d been deluding ourselves. The pieces we’d so boldly posted on our Work In Progress blog were still full of problems. Somehow we’d managed to not hear a witches brew of playing noises, balance problems, distorted sounds, and basic obnoxious hiss.
So we clapped editing caps back on our heads and got to work.
Good thing we didn’t understand the mind numbing effort this would actually involve, or that in the end it would be 5 months before we’d tracked down and eliminated all the discrete noises, until we’d brought the hiss under control.
Going methodically through just a single piece took us between 15 and 25 hours of editing time, while during this final re-editing we went through each and every one of our 9 cuts at least 3 times.
Such intensive editing was possible only because we did it all ourselves. The hundreds of editing hours needed to create Work In Progress cost only bucketfuls of nasty computer sweat and long days glued to our screen. If we’d had to pay real money for sound engineers and studio time, it would have been out of the question.
But we persevered and in the end brought each file to the point where there was nothing we knew to do which would make a significant improvement. In line with our Neo-Confucian world view, only then did we feel it OK to call our work “finished”.
There were other less ethically exalted reasons for continuing to edit until our pieces were preternaturally clean. Since our music takes listeners to a more noble prelapsarian world, this uncanny cleanliness suits it. Also as musical heretics, we knew it wise to head off bad tempered defenders of the status quo before they could attack our work as crude, unfinished, or unprofessional.
During this intensive re-editing we came to more fully understand the power of our new multitrack recording system. We’d set it up primarily to experiment with overdubbed instrumental and vocal tracks. Now we found it also gave us the power to deal with noises which before we’d had to accept.
The first hint of this came when we were rough carving raw sound files down into shorter semi-finished pieces for our blog.
With our old system and WaveLab Lite, cutting in the middle of a continuous sound always made a pop, which meant every deleted section had to start and end in the total silence between two sounds. But even edits in sections which appeared to be silent often produced pops. It was mysterious and very frustrating. Sometimes we could move a cut a few tenths of a second and set things right. More often we had no choice but to give up our musical idea and try something completely different.
Only the peculiar shape and slowness of our music made possible this simple minded approach, only the way it unfolds as “sound islands” separated by silences. Otherwise we never could have created our first two CDs.
So you can imagine our delight to discover with our updated kit and Tracktion 2 we could cut just about anywhere. Sure if we did it in the middle of a continuous sound, there’d be a pop, but then fading in and out for a few hundredths of a second around the noise made it disappear.
Of course more of our edits naturally fall in the quiet after the last note of a musical idea. However in our music and with our instruments these seemingly quiet intervals are not even close to silent. All our notes are followed by long “tails” of ring, and when we make a cut in the middle of one of these, there’s usually sufficient sound to cause a pop.
This screen grab shows such a situation. Though before the cut the decaying ring’s signal is almost invisible, it was loud enough that without microfades it popped.
Unfortunately when we make a cut in the middle of a resonant tail, even after removing any pop if the tonal color and/or volume of the ring is too different on the two sides of the cut, our work still stands out as an artifact, an unacceptable sore thumb which disturbs the flow the music.
So microfades are not an almighty silver bullet. However they remain a very powerful weapon in our war against noises, and we’d feel crippled if now we had to edit without them.
In most music pops and noises are less serious. Usually there’s enough other stuff going on to cover them. But since we’ve gone for a powerful rich simplicity, in our music there’s less place to hide. Our files must be super clean or noises and editing artifacts will be brutally obvious.
Figuring out things like this it helped that Tracktion, our new sound editor, does everything by applying “filters”. Even basic operations such as deleting, adjusting volume, and equalizing are treated as filters. In recording jargon this is described as non-destructive editing. Setting aside the fancy words, all that means is whatever we did, the original recorded signal stayed the same (though of course what we heard changed as we edited.)
This was perfect for ignorant folks like us. We could just push the virtual sliders around without fear of causing irreversible damage. When we didn’t quite understand what a filter was doing, trial and error taught us how to use it.
But back to those noises…..
Our new software also helped us tame “explosive sounds” which start with attacks so loud and sudden they distort. We’d simply put the problem note into a separate clip and fade in its “nose”. Then if that didn’t do it, we’d add high and/or low pass filters.
These filters were equally valuable for dealing with boomy sections. In fact in general we found them much more useful than equalizers. They seemed to do what we wanted them to do without producing as much change in the overall feeling of the sound. Of course equalizers might work if we knew something about the frequency distribution of our troublesome sounds. But we don’t. So rather than wasting time fiddling, more often we just went with high and/or low pass filters.
Since finger noises are just a type of initial explosive sound, we could treat them the same way. Immediately before the start of the noise we’d cut the file and then fade in the new clip right up to the nose of the actual note. Breathing noises were even less trouble. Since between words a vocal track can be totally silent, we could put our huffs and puffs into their own clips and then silence them.
Lip noises were trickier. Those before or after words we silenced like breathing noises, but others were rude enough to occur right in the middle of a word. For these guys we’ve yet to find a general solution, but usually with enough sweat and a combination of very short silences and clever fades, we could make them less audible.
Of course these editing moves are easier and quicker to describe than to do. It can take some serious head scratching before we find the best spot to split a clip. And even after it’s obvious where it should be done, actually doing it takes a few seconds. When there are long sections where for editing purposes every note must be put into a separate clip, these seconds start adding up. Similarly saying “fade it in” is very easy, but in reality it usually demands quite a bit of fussing to make a fade the correct length and shape. When there are thousands of fades in a single piece, here too the seconds accumulate…..
But discovering one more time that doing something well requires a shitload of work doesn’t really surprise us. Indeed we’ve seldom seen a “quick” and “effortless” approach result in much worthwhile. Most shortcuts seem to lead to dead ends. As a beaten up Zen friend once observed, “the easy way often turns out to be the hard way.”
For example cleaning up hiss we checked out d-essers, gates, and other filters which claimed to do the job with just a few clicks. But ultimately we were able to eliminate more hiss by going with our own labor intensive version of a gate filter.
We realized most of the hiss was in tracks recorded with our bottom of the line condenser microphones, and though we still wanted the detail these mics captured, we didn’t need their help catching the important between note resonance. There was enough of that in the ring rich tracks recorded with the dynamic mics placed in direct contact with our instruments. So in the noisy condenser tracks we were free to fade in sharply before the notes, to fade out more gradually after them, and then to put the rest of the hissy tails into separate clips which we silenced. Sure this cost us the little bit of resonance in the condenser tracks, but by greatly improving the signal/noise ratio of the remaining ring, we made it more clearly audible and powerful.
Now tricks like these must be taught to every would be sound engineer. But since we figured them out for ourselves we ended up doing them our own way, and our editing became yet another reason our music is so profoundly different.