I wrote the bulk of this text a couple of months ago after a day of not producing as much as I'd liked to. There are no real suggestions of how to improve or learn, but I thought I'd describe my thought processes so others may learn from my experiences.
Yesterday
I had one of those days yesterday where inspiration was lacking. Partly it may have been the brief I'd been given since it didn't have the depth of information I'd have liked. Mainly, though I think it was just me, I couldn't really find much to do, not that sounded right anyway. But I still managed to deliver something of sufficient standard.
How?
I tried creating music for the part from scratch. I wrote some interesting ideas. But they didn't really fit the brief as far as I was concerned. They're now exported to mp3 and the relevant project files archived for future reference. You never know what may be useful later on - I revisit unfinished files on a regular basis and often manage to complete them another time.
I then reviewed my collection of part-finished works hoping there was something nearly finished in there. Unfortunately there wasn't anything near finished. So I took two that really needed a lot of work and moved them forwards. One was almost finished from a composition point of view, but required a lot of mixing and processing, the other required more composition and almost no mixing.
Both featured a repeating rhythmic phrase and some background drums; one had a few bits of melody as well. The originals were rough versions of ideas that I thought worth keeping but didn't have the inspiration to finish when I first recorded them. They were a mixture of midi-triggered samples and audio recordings.
First Track
The first track was a sad and slow bluesy-jazz and had the following tracks:
For this one, it was mainly a case of arranging and mixing rather than composing. The original idea was over 3 minutes, the brief was for 1 min 50 seconds. So I worked at taking the best bits and arranging them into something musically attractive.
Before Arranging
Before I could arrange the work, I had to sort out the audio. It just didn't sound right. The original was recorded in one take per track, i.e. place some drums out, record the rhythm guitar in one go, then record the bass, then the lead guitar. Any of the audio processing was done very roughly just to convey the idea.
Rather than tweak and tweak and tweak some more, I removed all the processing and started again. After having listened to the track a few times, I knew where I wanted to go with it.
Clear some space
First of all I had to clear some space, the instruments were overlapping way too much. This involved some careful eq mainly focussed around making room for the bass guitar to be heard. Rolling off the low-end on the guitars and nudging the drums down in the same area helped bring the bass through considerably.
Guitar Tone and Amp Placement
I actually liked the tone I'd originally recorded with but it didn't fit in my vision of the mix. I changed the speaker emulation for a more familiar British 2x12 and turned the gain down to reduce the distortion. I also removed the reverb from the plug-in so I could add a different one later to fit in with the rest of the instruments. I'd considered re-recording with guitar amp and microphones but decided against this since the tone was sufficient for the background rhythm guitar.
Delay for lead
The lead guitar sings and needs to be heard above and beyond its accompaniment. I used a stereo delay on a aux bus for this and set the send so that it was audible. I did think about backing it off as I would do for a reverb, but I liked the extra character the delay gave the guitar.
Compression for drums
The drum loops were from BetaMonkey. They're good quality loops and are usually very dry or, in the case of the early ones, in a good-sounding room. In this particular song, the drums didn't cut through as I'd like. I could hear the snare too loud in the mix compared to the other drums. So I used a multiband compression on the drums with the gain mainly in the lower two bands and stepped down in the higher two respectively. Although setting the compressor up in this way is probably the opposite of what I'm more used to (i.e. heavier treble, lighter bass), this changed the character nicely.
Reverb for drums
To make the snare and cymbals live more in the mix, I used a reverb on an aux channel with a high-end pass. This added a nice, but barely audible quality to the drum track. The compression used in the previous step had brought the other drums up to a good level so this action balanced that out.
Mixbuss
Although nicely balanced, the instruments almost sounded like they were recorded in different rooms and the overall recording was slightly flat. A small eq alteration helped. I added the slightest amount of reverb to bond it together. I'm usually happy with a convolution reverb of a nice-sounding studio to get the feel for a track and then choose from there when I've got an idea of where I want it to be. For this particular track, I used a jazz club reverb for a more live sound. It fitted the bluesy-jazz feel of the whole track.
The final element in the chain was a limiter. I usually have two exports, one with the compressor in the chain and one with out. Then I've got the flexibility to hear what it would sound like compressed like many other tracks but still give clients the more dynamically accurate version without the compression.
The second track
This was a lot simpler in scope, but I spent a lot more time on it due to having to compose more parts and phrases to glue it together. The original idea was 50 seconds long and consisted of a midi-triggered samples of a piano, acoustic double-bass and an acoustic drum-kit. That's not that long.
Timing
I'd originally played all the parts on a midi keyboard. So the first step was to ensure all the notes were on time. This was a combination of quantising and manually adjusting the placement of notes. Some of the notes were also wrong, just not fitting in with the key. I manually adjusted these as well, rather than relying on any logical processing.
Drums
Oddly for me, the drums were samples rather than loops. I tend to find that loops fit in a lot better than any samples ever can. For this song I'd only used 4 drum sounds. I'd envisaged a drummer with an older style kick, snare and hihat, rather than the mammoth rock kits of the 80s. This simplicity, specifically the lack of cymbals, helped make the sound more realistic.
To add further realism, I patched in a light drum room reverb, just light enough to hear.
Piano
The piano sample sounded too artificial when solo'ed. I auditioned a few other piano samples, including some that were sampled from much better pianos and with much better recording. Yet I decided on the original but with two major changes:
Structure
The original was 50 seconds long. I needed an extra minute. By thinking through the structure I wanted to introduce, I was able to create the 1 minute 50 required. It was just on the edge of having to introduce a different chord progression, theme or motif. Any longer and I would have had to. As it is, the different instruments coming in and out make it interesting enough for the short time. I looped the piano, bass and drum parts. Cutting out a few loops to bring attention to other instruments. I then added the guitar over the top and then the jazz organ.
Guitar
The guitar was troublesome. I knew it needed another instrument or two to keep it going to 1 minutes 50 seconds, but every time I played along to the song, I was left short of ideas and just had a tangled mess. I also sat down with the chords and tried working out a melody. Again nothing worked. It was accurate harmonically, just not interesting. I tried changing guitar sounds a few times - I find the inspiration from that can help a lot, even from a totally inappropriate guitar tone. In the end, I switched on the cycle function and hit record. About 45 minutes, later I had enough pieces I could use. These were pasted into the arrangement and unused pieces deleted.
The tone was deliberately kept mellow.
Jazz Organ
Again another instrument to fill in the gap from the guitar and bring the listener out of the emptier phrase in which it starts. I played this, then quantised/shifted notes manually. I was careful to use the correct quantisation for each phrase (mainly 16ths and 24ths, i.e, triplet 16ths).
Finally
There was almost no eq of the individual tracks, they sat well enough together without much alteration. I added a simple jazz reverb on the mix buss and compression similar to the first track. That worked well.
Results
I'm happy that I finished two pieces both of the required length meeting the brief.
Was it a good day though?
Probably not.
I wouldn't call either piece usual for me. That's not necessarily a bad thing. It wasn't so much that it was a different type of music to what I'm used to (they're not, I'm quite happy with that style). Instead it was more than I felt uncomfortable with the end results. They're good enough for background music which is what they're intended to be, but don't have the same melodic focus I'd usually want.
I was nowhere near as creative in the whole day as I was in the first hour of the following day. On that day, I wrote three good ideas in less than an hour. I spent an hour tidying up one of the tracks, bear in mind it was 7 minutes long. Like the above tracks, they'll take longer to finish, but that can wait for another day when I need to progress them further.