Choosing a Dynamic Range
AI material is born flat, and how much contrast to keep is an aesthetic call rather than a setting.
Related
The problem
AI music often stays equally full from start to finish. Verse and chorus differ in instrumentation but not in force. Compress that into a master and it sounds professional without ever moving. Push levels the other way to manufacture contrast and you get pumping and fake swells. How much range to keep is a decision, not a default.
How it works
Dynamics work on two levels. Macro dynamics are the level differences between sections; micro dynamics are the crest factor between notes and transients. AI material is usually flat on both, because generation outputs a fixed density and post-processing limits it. Compression can reduce dynamics but cannot create them. Real movement comes from arrangement density, playing force and volume automation; a compressor only gathers what already exists.
Practice
Set a relative level for each section first. A verse 2 to 4 LU below the chorus, a bridge perhaps lower still, written with volume automation rather than squeezed out of a compressor. Shape transitions over two to four bars instead of stepping, since the ear notices a jump more than a gap. Leave the drum bus 2 to 4 dB of crest factor instead of flattening every hit to one height. Use bus compression only where you need to gather, with no more than 2 dB of gain reduction. Check the loudness range, where 5 to 8 LU is normal for pop, wider for classical and film, narrower for background music. Finally, turn the monitor level down; if the movement survives, it was not carried by loudness.
What to listen for
In Mahler's Wenn mein Schatz Hochzeit macht, hear the voice move from murmur to outcry, which is macro dynamics. In Mel Bonis's La Cathédrale blessée, hear the structure build step by step and whether the sections leave a stair. In the slow movement of Beethoven's String Quartet No. 12, hear the tiny differences of force between long notes, which is micro dynamics. Bring all three to the same loudness first.
Listen in the library
All tracks from midicn-lib — inline badge is the license tier (C1 commercial / C2 non-commercial / C3 study only).
Sources
- 宏观与微观动态的区分,依混音与母带通识核对,未抄文本
- Macro and micro dynamics, based on general mixing and mastering practice, no text copied
Used for fact-checking only; all prose is original.