Build a mix from a tracklisting
Given a tracklisting, one track per line, the dermixen command builds the mix in that order, resolving each line to a file in the library, which is everything Dermixen has learned about your tracks, kept in one file in your data folder. An agent driving the app does the same, which is the second workflow DESIGN.md names.
Prerequisites
- A library that contains the tracks, from Scan your music.
- The tracklisting as a text file, one track per line, in the order the mix plays. Leading track numbers and timestamps are fine.
Resolve each line
The top match's score measures how well the line matches the track. A score of at least 0.8 is a resolved line. A score below 0.5 is a guess: report the line rather than build on it. Between the two, look at the path the match names before you trust it.
Build the mix in order
dermixen mix new set.dmx
dermixen mix add set.dmx "/Users/dermixenuser/audio/goa/comp/VA - Analog Dreams [DATCD005]/2 Space Tribe - The Great Spirit (Original Mix) - mastered.mp3"
dermixen mix add set.dmx "/Users/dermixenuser/audio/goa/comp/VA - Sun Trip [LRR52231CD]/01 Koxbox - Point of No Return.mp3"
Each mix add appends the track, joins it to the one before it with the default transition, the blend, and prints the timeline. mix add refuses an add it can't place, with a message that says why, and leaves the document as it was, so the timeline it prints is always a mix that renders.
Do it in one script
The script below reads a tracklisting, resolves every line, and builds the mix only when every line resolved. It stops at the first line it can't resolve and writes no mix, so a set with a hole in it is never built by mistake.
import json
import subprocess
import sys
listing, mix = sys.argv[1], sys.argv[2]
lines = [line.strip() for line in open(listing) if line.strip()]
paths = []
for line in lines:
out = subprocess.run(
["dermixen", "library", "find", line, "--limit", "1", "--json"],
capture_output=True, text=True, check=True,
).stdout
matches = json.loads(out)
if not matches or matches[0]["score"] < 0.8:
print(f"unresolved: {line}", file=sys.stderr)
sys.exit(1)
paths.append(matches[0]["record"]["path"])
subprocess.run(["dermixen", "mix", "new", mix], check=True)
for path in paths:
subprocess.run(["dermixen", "mix", "add", mix, path], check=True)
The lines that don't resolve are the ones to fix by hand: a track the scan hasn't seen, a file with no tags whose name the parser read wrongly, or a title spelled differently in the listing than in the file.
Check the result
mix show prints the timeline with every track's tempo, gain, anchors, and start and end times. Check a transition by ear covers rendering or playing one transition at a time, and Change a transition covers presets, lengths, and anchors when the default isn't right for a pair of tracks.