4 ms·
The author makes a distinction between two different modeling approaches: 1. Representing music as a series of notes (with additional information about dynamic
by DiogenesKynikos 3y ago
The author makes a distinction between two different modeling approaches:
1. Representing music as a series of notes (with additional information about dynamics, etc.), and then building a model that transforms this musical score into sound.
2. Modeling the waveform itself directly, without reference to a score. This waveform could presumably be in either the frequency or the time domain, but the author chooses to use the time domain.
The author's terminology is a bit confusing, but I think that they mean option 2 when they say "the waveform domain."
- earthnail 3y agoThe frequency domain is a misleading name. I assume you’re referring to STFT, or spectrogram, which is a series of windowed time segments transformed into the frequency domain. But both you and OP are right in that waveform domain is usually called time domain.