I arranged and mixed the songs while making the demos. I'd decided
that I wanted to do full demo recordings because it wasn't much extra
work to pull the loops off of the looper and sequence them in the
DAW. I wrote Little Gwaii before I had the RC-300, so I couldn't work
this way before while keeping the looped instruments separate (the
RC-300 has three separate tracks).
For
someone who barely uses any outboard gear to make records, this
project was dominated by two new toys: the Boss RC-300, and a 15”
HP Envy laptop I bought that I ended up using for the final
recordings and mixing. I needed it so that I could keep working on
the recordings while in Toronto music-directing my musical, since I
couldn't travel with my desktop. I picked it because it was the only
sub-$1000 laptop I could find that had a significantly better
processor than my desktop's; I didn't want to spend $600+ on
something that'd be relegated to second string when I got back home.
At the same time I got a slightly newer version of Sonar, X2 (which
was an unstable nightmare until I got the X2.1 patch, which calmed it right down).
I never actually got around to doing demo vocals, I was still just singing live overtop of my demos when I started recording the final versions. I worked out most of my ideas about how many verses and choruses there should be at that time, keeping the song structures fairly traditional, my goal being more to see how I could replicate normal pop while using a looper than creating super experimental things. Three of the 16 songs still needed lyrics when I started doing final drums. I'm not sure why those three (ESTJ, INTJ, and INFP) took longer than the rest, which were mostly written in short succession over a two week period, but I managed to finish them in a couple of days in Toronto.
For the final versions of the drums, I decided to just use my practice drumset rather than renting things to augment it, as I usually do. So that's the same no-name Taiwanese kit I've had since I started playing, although I've learned a bit about tuning since then. Similarly, I used whatever mics I happened to have on hand, apart from an SM7b for the top of the snare. That meant CAD drum pack mics, SM57s and SM58s.
But instead of letting those mics sound like themselves, I decided to do some extensive EQ modelling on them as an experiment. I picked mics that I wished they were, like MD421s for the toms, KM141s for overheads, and then compared the frequency response diagrams provided by the manufacturers of those mics with the ones for my mics. Next, I determined what modifications I could make to bring the one curve into line with the other, the locations of the boosts and the cuts I needed to make and their Qs, which I expressed as the length of the frequency spectrum it had to affect, like this:
+2 dB @ 500 Hz ± 250 Hz
I put these EQ changes into Ozone 5's EQ module, which then gave me a single slider to control the total amount of these modifications to use, from 0% to 200%, and chose the spot where I thought it sounded best. I'm aware that the results obtained from this method will differ from using the actual mic desired, because different mics behaved differently off-axis and depending on proximity, and many other reasons I don't understand. But I didn't mind the results I got in the end, so I think it is a useful method, even if a better method for next time might be to rent the desired mic for a day and record it side by side with my substitute, so that I can compare the waveforms directly and use Ozone's matching EQ feature to calculate the changes required automatically.
I edited the drums in spare moments I had in Toronto while working on the play. Some came together quite quickly, depending on the length of the drum loop and how well I'd played it. A few were as simple as finding the cleanest two-bar section and making sure it looped tightly. Other tracks required more extensive cutting up, and I told myself I was being zealous now so that I'd have a solid base to overdub everything else onto. It was still way easier than recording non-looped drums.
Next I added keys, when I was using MIDI and synths, and any keyboard loops that I was salvaging from the demos. I rescued some of the original DX7s keyboard tracks because they sounded fine – appropriately gritty – and I couldn't find a soft synth that I preferred. I recorded MIDI with an unweighted KeyRig 25 controller, editing the results extensively for timing and velocity. This was about the same time that I was working on the cast recordings of my musical, where I decided that I'd just import the MIDI piano from the score writing program and “humanize” it a bit by adding timing and velocity irregularities, so I was aware of how far I should go in the opposite direction for it to sound believably tight as opposed to robotically quantized. Not as good or fast as being able to play the piano well, but still faster than learning how.
Other odds and ends like the melodica bits I was able to stick right in from the demos; I didn't change any of the tempos, which might not have been best. I should have been performing the songs for people at this point, making sure they didn't feel like they were dragging or going too fast. I guess I was busy though, between the musical and the rest of the recordings.
I never actually got around to doing demo vocals, I was still just singing live overtop of my demos when I started recording the final versions. I worked out most of my ideas about how many verses and choruses there should be at that time, keeping the song structures fairly traditional, my goal being more to see how I could replicate normal pop while using a looper than creating super experimental things. Three of the 16 songs still needed lyrics when I started doing final drums. I'm not sure why those three (ESTJ, INTJ, and INFP) took longer than the rest, which were mostly written in short succession over a two week period, but I managed to finish them in a couple of days in Toronto.
For the final versions of the drums, I decided to just use my practice drumset rather than renting things to augment it, as I usually do. So that's the same no-name Taiwanese kit I've had since I started playing, although I've learned a bit about tuning since then. Similarly, I used whatever mics I happened to have on hand, apart from an SM7b for the top of the snare. That meant CAD drum pack mics, SM57s and SM58s.
But instead of letting those mics sound like themselves, I decided to do some extensive EQ modelling on them as an experiment. I picked mics that I wished they were, like MD421s for the toms, KM141s for overheads, and then compared the frequency response diagrams provided by the manufacturers of those mics with the ones for my mics. Next, I determined what modifications I could make to bring the one curve into line with the other, the locations of the boosts and the cuts I needed to make and their Qs, which I expressed as the length of the frequency spectrum it had to affect, like this:
+2 dB @ 500 Hz ± 250 Hz
I put these EQ changes into Ozone 5's EQ module, which then gave me a single slider to control the total amount of these modifications to use, from 0% to 200%, and chose the spot where I thought it sounded best. I'm aware that the results obtained from this method will differ from using the actual mic desired, because different mics behaved differently off-axis and depending on proximity, and many other reasons I don't understand. But I didn't mind the results I got in the end, so I think it is a useful method, even if a better method for next time might be to rent the desired mic for a day and record it side by side with my substitute, so that I can compare the waveforms directly and use Ozone's matching EQ feature to calculate the changes required automatically.
I edited the drums in spare moments I had in Toronto while working on the play. Some came together quite quickly, depending on the length of the drum loop and how well I'd played it. A few were as simple as finding the cleanest two-bar section and making sure it looped tightly. Other tracks required more extensive cutting up, and I told myself I was being zealous now so that I'd have a solid base to overdub everything else onto. It was still way easier than recording non-looped drums.
Next I added keys, when I was using MIDI and synths, and any keyboard loops that I was salvaging from the demos. I rescued some of the original DX7s keyboard tracks because they sounded fine – appropriately gritty – and I couldn't find a soft synth that I preferred. I recorded MIDI with an unweighted KeyRig 25 controller, editing the results extensively for timing and velocity. This was about the same time that I was working on the cast recordings of my musical, where I decided that I'd just import the MIDI piano from the score writing program and “humanize” it a bit by adding timing and velocity irregularities, so I was aware of how far I should go in the opposite direction for it to sound believably tight as opposed to robotically quantized. Not as good or fast as being able to play the piano well, but still faster than learning how.
Other odds and ends like the melodica bits I was able to stick right in from the demos; I didn't change any of the tempos, which might not have been best. I should have been performing the songs for people at this point, making sure they didn't feel like they were dragging or going too fast. I guess I was busy though, between the musical and the rest of the recordings.
No comments:
Post a Comment