Whenever we listen to music we learn from it. This happens whether we're consciously aware of it or not. Some pieces of music won't appeal to us and (subconsciously) we'll have an awareness of why they don't appeal. Likewise with songs which we love as well as the many songs we neither openly love nor detest. We're always learning what works and what doesn't, from our perspective.
We may open up our sequencer and start with a blank slate, but our minds aren't blank slates. When it comes to crafting a pleasing mix, we'll be guided by the music we've listened to. If our guitars don't sound chunky enough, we only know that because we've heard nice chunky guitars. If our kick drum doesn't have enough impact, we recognize it because we've heard satisfying kick drums in songs we like the sound of.
Given that we're always learning, I'd recommend doing it consciously. If we hear a beautiful lush pad in a song, we shouldn't rely on serendipity to bring it forth in a year's time, when a similar pad may fit perfectly into our current song. Focus on the pad and try to describe it as fully as possible - how wide is it? Is it only represented in the mid range, or does it have decent low and/or high end as well? Is there any modulation - panning, filters, delay effects? How does it interact with the other sounds - are there particular other sounds which help to frame the pad and make it sound the way it does? The more detail we can use to describe what we're hearing, the better we'll be able to recreate and incorporate the sound into our own sonic palette.
This leads me to the main topic of this post. It's great to describe a sound in detail, but in my experience the next step is more important - actually attempting to recreate the sound. I've addressed the issue in some of my posts already regarding the "copying vs originality" aspect of this. In short, as I've already said, we're learning all the time whether we're doing so consciously or not. Recreating sounds won't lead to any less original music than starting with a blank slate. It will however give us more techniques to use when we're making music.
Here's how I go about recreating songs.
I like to build songs up from the foundation - the kick and bass. The first thing I do is look for sections in the song/ album where these are as exposed as possible. When I find a song which is a good candidate, I'll import it into my sequencer and adjust the project's tempo to the tempo of the song. Some sequencers (such as Ableton Live) do this automatically, but mine doesn't, so I just loop a four bar section and adjust the tempo until the section loops cleanly.
Then I set the loop markers so a two bar section with a decently exposed kick drum is looping. The first bar plays the original audio, the second bar is silent (either via volume automation or simply cutting the original audio and moving the section after the first bar out of range of the loop). Then I go through my kick samples, in whichever sample library is closest to the style of sample I'm looking for. I'll hear four of the original kick, then four of the kick I'm previewing. I make note of kicks which sound close to the attack transient, kicks which sound close to the body and kicks which sound close overall. It's great when I find a kick which is close on its own (even better when I find the exact kick the original artist used!), but I have no problems splicing together a great sounding kick using the attack portion of one and the body of another. The thing to keep in mind when cutting out the parts of each kick that aren't required is to enable the "snap to zero crossing" option. Otherwise the audio will pop every time it plays the uncleanly spliced kick. Over the course of a song that's a heap of pops! Once I've gone through my kick library I may have noted down 20 kicks. I'll go through these again and keep narrowing it down until I find the closest one(s).
Once I have a kick which is pretty close I compare it to the original by running a frequency analyzer over both channels and seeing what differences that shows up. Obviously if the original version has other sounds playing I make allowances for these. I'll make a few EQ adjustments to get my kick even closer to the original.
Now that I have a decent version of the original kick it will be easier to hear and audition the other sounds relative to it. Ideally every sound would be isolated at some point during the song/album. When that doesn't happen it's helpful to have a very similar kick sound and listen to where the other sounds are sitting relative to it - level-wise, frequency-wise, their stereo width, panning and so on. Sometimes I'll look at a frequency analyzer, in which case it's definitely handy having a similar kick so I can focus on what the other sound is adding, rather than trying to judge which frequencies belong to the kick and which to the other sounds.
Kicks are often dry, mono and in the middle, whereas other sounds are more likely to be further processed. I note the reverb and delay that have been added to sounds, as well as any other treatment and do my best to replicate these.
After the kick I'll either proceed to the hihat and snare, or the bass. It depends on which element is most exposed. In this example I'll do the bass first.
I have a number of samplers and synthesizers I'm reasonably familiar with, so I'll generally know which one to turn to for a sound similar to the original bass sound. As before, I loop a section where the bass is exposed (if available), or where it's predominantly the kick and bass playing. I'll work out the notes the bass is playing and create a MIDI version to send to my sound source. Then I repeat the sound selection process - going through bass sounds, making notes of the sounds which are closest to the original, modifying parameters to push the sound closer to the original. Again, I may have noted down 10 to 20 sounds, which I'll progressively narrow down to the one I'll use.
The hihat and snare/clap are generally quite straightforward. Sometimes I'll adjust the pitch or length of a sample but generally I'll be content with something which fulfills the same role as the original sample rather than seeking an exact match.
I find a lot of mid/ higher bass sounds in trance very difficult to recreate. They often have filter modulation, phasers, delays or a myriad of other effects applied to them. Often this will be where I have to concede defeat for the time being. On a few occasions I've inserted a different mid bass sound which essentially fulfills the same function, which resulted in a very solid foundation to build pad and lead sounds onto for my future original productions.
I won't go into detail for the other sounds, since the same principles apply. I have a number of guitar amp emulators when I'm trying to match heavy guitar sounds in metal songs. I look for a similar sounding preset and work from there. For high percussive loops I choose a similar style of loop rather than trying to match it perfectly. For pad and lead sounds I listen to the relationship between them and the solid foundation I've already created. Essentially, the more songs I recreate, the better I'm able to hear small differences in sounds. Over time, when selecting sounds, I'm eliminating more of them before I note them down as candidates. Hopefully one day I'll reach a point where I can just go through and choose the best sound without making any notes. More experience with my sample libraries will help, as would an enhanced knowledge of synthesis.
To sum up, our music libraries are the best learning resource we have. If we take the time to study them and to experience their creation directly (to the best of our ability), we'll make tremendous progress. Each song is a new experience, which will stretch us and make us grow. Great masters of classical music transcribed music of the composers they admired in order to achieve this direct experience with the music they loved. I heartily encourage the practice.
I hope you've found this useful, keep making great music!
Fabian
No comments:
Post a Comment