Reply to thread

Message: [QUOTE="Shamiso, post: 92484, member: 160"] An octave error happens when a pitch detector chooses half or double the intended fundamental frequency, shifting the resulting MIDI by 12 semitones. Bass and 808 material make this failure especially obvious because a wrong note can still sound related to the one you played. You might convert a clean bass phrase and get seven correct notes plus one random leap an octave higher. Move the bad note down 12 semitones, and it suddenly fits perfectly. Pretty good clue. This is a different problem from an 808 sample being mapped to the wrong root note inside a sampler. [B][URL='https://goldmidi.com/community/threads/never-found-introduced-mg-mini-for-midi-capture.77885/']MG-Mini's Single Line bass capture[/URL][/B] is dealing with detected pitch from audio, so the first job is figuring out whether the MIDI note itself is wrong or the instrument playing it back is misleading you. [HEADING=2]Low bass can fool pitch detection by exactly one octave[/HEADING] A bass note is not one lonely frequency. Real bass sounds contain a fundamental plus upper harmonics, and some of those harmonics can carry more energy than the fundamental after recording, filtering, saturation, distortion, or ordinary instrument resonance. The second harmonic sits at twice the fundamental frequency. If a detector latches onto it instead of the fundamental, the musical result lands exactly one octave higher. The note name stays the same, which is why an E can turn into another E and still look almost reasonable in the piano roll. Low notes make the decision harder because their waveform cycles take longer. A detector examining a short slice of audio gets fewer complete cycles from deep bass than from a midrange note, while the attack may be full of transient energy that does not represent the settled pitch very well. Hard distortion can complicate things further. Saturation adds harmonic content, which may help a bass speak on smaller playback systems but also changes the evidence a pitch estimator receives. A heavily processed 808 can therefore be easier to hear in a mix while being less tidy as transcription input. Octave mistakes can also go downward. Period-based pitch methods can confuse a repeating pattern with a period twice as long as the true one, producing half the intended frequency. The classic [B][URL='https://doi.org/10.1121/1.1458024']YIN fundamental-frequency paper[/URL][/B] was built around reducing pitch-estimation errors caused by exactly these kinds of periodic ambiguities. [HEADING=2]An 808 does not always hold one pitch[/HEADING] A long 808 is not necessarily a stationary sine wave with one pitch from start to finish. The attack can begin above the eventual bass note and fall rapidly, while processing or a longer envelope can make more of that movement audible. A 2026 analysis of pitch in contemporary popular music describes the TR-808 as a reference point for continuous frequency trajectories. In the examples discussed there, the fundamental can fall before settling, and some longer bass gestures move through much larger intervals. Feed a moving pitch into a converter and somebody has to decide what counts as the note. A detector that gives too much weight to the attack may report a higher pitch than one that favors the stable tail, even though both are looking at the same sample. This is why chopping a long 808 into tiny pieces before conversion can backfire. Short slices may contain mostly the pitch drop or transient instead of enough settled tail for a useful note estimate. Longer clean sections give the system more context, provided adjacent notes are not overlapping. Slides create another trap. If an 808 deliberately glides between pitches, a converter may output several stepped MIDI notes rather than one note plus continuous pitch movement. Those extra notes are not necessarily octave errors, even when one briefly lands 12 semitones away during a large slide. [HEADING=2]Fix the octave error before touching the groove[/HEADING] Start by soloing the source and comparing each suspicious MIDI note with the audio at the same moment. If the converted note is exactly 12 semitones away and shifting it by one octave restores the phrase, you are probably looking at a detection error rather than a harmonic decision in the bassline. Next, inspect the source before rewriting the MIDI. A high-pass filter, aggressive saturation, clipped recording, layered kick, or noisy attack can weaken the fundamental or make another component dominate the first part of the note. Convert to a cleaner version when one is available. For an 808 with a clear pitch drop, listen to the sustained portion separately from the transient. The tail often tells you more about the note you intend to use musically, while the attack may be doing its own short pitch movement for impact. Do not automatically transpose the entire conversion because three notes jumped. A global 12-semitone shift only makes sense when the whole source was interpreted in the wrong register. Random isolated jumps need local correction. Finally, load a plain synth patch and play the corrected MIDI before blaming the transcription again. If the MIDI notes now follow the source but your favorite 808 still sounds an octave wrong, check its root-note mapping, tuning, and sampler setup instead. Different bug, different fix. [/QUOTE]

Name