|
SongTrellis |
|
||||||||||||
|
|
|||||||||||||
|
Kristian, I am certain that the chord arrangement posted doesn't match the string arrangement. I was simply trying to ascribe the correct chord name to the chords used in the arrangement so that one could use those changes to accompany someone who was singing Etta James' vocal line or so that they could improvise over those changes. The chord arrangement that resulted was just a side effect of completing my chord progression transcription. When I worked on this tune, it was one of the first times I had tried to do something practical with a feature I had recently added to my music editor. I had the new capability to synchronize the playback of an mp3 file as my editor played and animated my transcription attempt. I first imported the mp3 of Etta's performance into an empty score in my editor. I then figured out the elapsed time in the performance at which the first beat of the A section of the tune occurred after that gorgeous string intro. After that I figured out the tempos that were used across the performance. The tempo that seemed most accurate for the A sections was 58.150 (my app handles tempos down to the thousandth of a beat per minute). I verified that the tempo was correct by watching the beat counter in my animation and verified that the beat transitions that happened were synchronized with the drummer's time keeping. After that I typed in the chord progression for the A sections and the partial B section that my sax teacher had given me. My editor synthesizes new accompaniments from the changes entered every time I would play the score. By listening to this with the MP3 I was able to verify that what we had notated so far sounded correct. It was left to me to figure out the last 5 bars of B. In my score they were initially represented by 5 whole note rests in the chordal part of my score. I would select the next bar in the chord score that had not been filled and press play to listen to the corresponding segement of the mp3 performance. I would make a guess about the chords that were used, enter them into the score and then listen to them with the mp3 playing. Invariably my guess would sound wrong, so I would select the incorrect chord and bring up the app's Chord Grid dialog. The grid consists of a rectangular grid of chord specifications. The rows specifiy the forty or so chord types known by the editor and the columns specify the twelve possible chord roots upon which chords of those types could be built. By picking a square in the grid, I could temporarily substitute a different chord from what I had originally specified and audition the score with that substitution in place. If the substituted chord sounded like the best expression for what happened in the performance I would press the OK button in the dialog, which would replace the original chord in the score with a chord of the newly selected root and type. Using this method I was able to figure out the last part over about an hour. I'd guess that someone who is proficient at chord identification would have been able to do the same in three or four minutes. Anyway the chord voicings that are published are just the voicings that my application figured out in a fraction of a second the last time I pressed my editors Play Score button before I saved the score in MIDI format. As you can hear the voicings published are just alternative arrangement that would work OK as a substituted parts for the actual arrangement that was used. The string arrangements for Etta's performance were done by a master arranger so I'm sure they are worth transcribing exactly. I didn't need that level of specificity to satisfy my subscriber's request. I'm sure it would take me several days to get an exact transcription done correctly. Best of luck with your soundalike project. blog comments powered by Disqus
|
|||||||||||||
|
Last update: Sunday, February 3, 2002 at 10:55 AM. |
|||||||||||||