This page is BELIEVEYOUME, mode-aware key verdicts, V3. It is one of the projects Brandon built and put online. The grown-up details are below, tap Grown-up mode to read them the normal way.
Confidence is the score gap from the top mode candidate to the runner-up, expressed as a percent of the top score.
Source-MIDI tag. Whether the third is reliably observable depends on whether the vocal track is present. The third is often carried by the singer in Shiner's writing, so an instrument-only MIDI can systematically pull the verdict toward Aeolian or Dorian.
Shiner-TheAlligator-FROMCDN.mid full-vocal, 138 notes| Window | Beat range | Verdict | Score | Gap % | Runner-up | Top 5 pitch classes |
|---|---|---|---|---|---|---|
| 1 | 0.0 to 69.1 | G Mixolydian | 0.1020 | 4.2 low | G Ionian | G, B, D, A, Bb |
| 2 | 69.1 to 138.2 | G Mixolydian | 0.1063 | 5.2 medium | G Aeolian | G, C, D, B, A |
| 3 | 138.2 to 207.3 | G Aeolian | 0.0800 | 21.1 high | G Mixolydian | G, B, C, D, Bb |
| 4 | 207.3 to 276.4 | G Aeolian | 0.0886 | 11.4 medium | C Dorian | G, C, D, Eb, B |
| 5 | 276.4 to 345.5 | G Aeolian | 0.0818 | 1.6 low | C Dorian | G, Eb, F, C, E |
What the table actually says. G is the tonal center in every window. The third is unstable. Early in the song the singer holds the major third up against a rootless riff, so the window reads Mixolydian. Later in the song the minor third leaks into the rhythm side, and the windows tip Aeolian. Window 3 is the only high-confidence segment, and it is on Aeolian, which means the verse-versus-chorus split is a real structural feature rather than detector noise.
The song-wide gap from Mixolydian to Aeolian was only 1.8 percent. Calling the song Mixolydian outright pretends the algorithm settled an argument it did not settle. The V3 headline states what the algorithm actually found: a G-centered song whose mode tips between Mixolydian and Aeolian by section.
G 0.308 D 0.123 B 0.118 C 0.103 Eb 0.085 A 0.084 Bb 0.069 F 0.061
B (natural third) at 0.118 versus Bb (flat third) at 0.069 is the headline number. The major third wins, which is why Mixolydian beats Aeolian by a hair song-wide. F (flat seven) at 0.061 is the structural fingerprint of Mixolydian over Ionian. The coexistence of both thirds is doing most of the work, which is why the song-wide verdict is low confidence.
| Rank | Root | Mode | Score | Note |
|---|---|---|---|---|
| 1 | G | Mixolydian | 0.0996 | cadence-aligned |
| 2 | G | Aeolian | 0.0978 | cadence-aligned, the chorus mode |
| 3 | G | Ionian (the old label) | 0.0767 | cadence-aligned, ruled out by Bb and F |
| 4 | G | Dorian | 0.0720 | cadence-aligned |
| 5 | C | Dorian | 0.0584 | parent-IV interpretation |
The harmonic-spine table on the chapter page is the ground truth for this song. It lists the bass and chord events directly: G5add9, Bb6/9, Gmaj9, with the bass descent at the end of each bIII bar. Mode labels summarize that table. When the mode confidence is low, rely on the spine table.
sofarso-06-03-2026.mid full-vocal, 236 notes| Window | Beat range | Verdict | Score | Gap % | Runner-up | Top 5 pitch classes |
|---|---|---|---|---|---|---|
| 1 | head | D Mixolydian | 0.1874 | 12.1 medium | D Ionian | D, F#, C, A, G |
| 2 | verse | D Mixolydian | 0.1382 | 7.5 medium | Gb Locrian | D, F#, C, Bb, A |
| 3 | chorus | D Aeolian | 0.0871 | 1.4 low | Bb Lydian | D, Bb, C, A, F |
| 4 | bridge | A Dorian | 0.0935 | 4.8 low | D Mixolydian | D, E, C, F#, A |
| 5 | outro | Bb Lydian | 0.1414 | 5.3 medium | Bb Ionian | Bb, C, D, F, A |
What the table actually says. D Mixolydian at the head means F# (the major third) is present and heavy, the song opens in the major-leaning end of the spectrum. By the chorus the minor third has taken over and the window reads Aeolian. The outro pivots toward Bb as a tonal center, which is the cousin of the I-bVI-bVII rock vamp the chapter card was already describing. The chapter card's bare "D minor" label captured the chorus and missed the verse, the outro, and the bridge. The honest sentence is what the headline now says.
D Aeolian wins the song-wide vote because Bb and C are heavy in the chorus and outro and the algorithm is averaging the whole song. The fact that the song-wide gap is 11.5 percent (medium) while window 3 alone is 1.4 percent (low) tells you the algorithm is more confident in the average than in any single section. That is the expected behavior of averaging. The section table is the truer picture.
D 0.284 Bb 0.165 C 0.157 F# 0.128 A 0.075 F 0.062 E 0.055 Db 0.053
F# at 0.128 against F-natural at 0.062 is the diagnostic. Bare D Aeolian is impossible because Aeolian forbids F#. The template still scores Aeolian top because Bb and C are very heavy, and Aeolian natively contains both.
| Rank | Root | Mode | Score | Note |
|---|---|---|---|---|
| 1 | D | Aeolian (template fit, F# disproves a literal reading) | 0.0608 | |
| 2 | Bb | Lydian | 0.0538 | matches outro |
| 3 | C | Mixolydian | 0.0526 | cadence-aligned, rejected because C is bVII |
| 4 | G | Dorian | 0.0440 | |
| 5 | F | Ionian | 0.0428 |
The harmonic-spine table on the So Far So chapter card lists the bass and chord events directly, including the chorus pedal on Bb that the song-wide pitch histogram is reflecting. Read the spine first, then read the mode line as a summary that holds in the verse and drifts in the chorus and outro.
Shiner-BrokenSatellites-FROMCDN.mid full-vocal, 258 notes| Window | Beat range | Verdict | Score | Gap % | Runner-up | Top 5 pitch classes |
|---|---|---|---|---|---|---|
| 1 | head | E♭ Lydian | 0.1314 | 5.9 medium | C Dorian | E♭, B♭, C, F, D |
| 2 | verse | B♭ Ionian | 0.1264 | 0.2 low | E♭ Lydian | E♭, F, D, B♭, C |
| 3 | chorus | E♭ Lydian | 0.1118 | 1.1 low | C Dorian | E♭, F, C, B♭, G |
| 4 | bridge | C Aeolian | 0.1065 | 3.1 low | E♭ Ionian | C, E♭, G, B♭, A♭ |
| 5 | outro | E♭ Lydian | 0.1754 | 15.8 high | C Dorian | E♭, D, G, C, B♭ |
What the section table says. Windows 1, 3, 5 are unambiguously E♭ Lydian, with the outro the strongest. Window 2 reads B♭ Ionian by 0.2 percent, which is the parent key of E♭ Lydian heard from its 5th (so this is the same parent key, just centered on the V). Window 4 (bridge) is where the song darkens: A♭ appears, A-natural drops out, and C Aeolian takes the window. C Aeolian = C E♭ F G A♭ B♭, which is the parallel-minor borrow over the E♭ root, the classic IV-of-bIII-or-darken-the-bridge move.
E♭ 0.291 C 0.133 D 0.129 B♭ 0.125 G 0.108 F 0.095 A 0.079 A♭ 0.031
E♭ dominates as the tonic. G (major third) is structural. A-natural (♯4) over A♭ is the Lydian diagnostic. D (major 7) is heavy at 0.129, ruling out Mixolydian or Aeolian.
| Rank | Root | Mode | Score | Note |
|---|---|---|---|---|
| 1 | E♭ | Lydian | 0.1613 | cadence-aligned, high confidence |
| 2 | E♭ | Ionian (the old label, the natural-4 disproves it) | 0.1357 | |
| 3 | C | Dorian | 0.1273 | relative-minor reading |
| 4 | A | Locrian | 0.1225 | |
| 5 | G | Aeolian | 0.1214 |
The harmonic-spine table on the Broken Satellites chapter card already lists the E♭ tonic prolongation, the B♭ (V) parking, and the Cm (vi) pedal in the loud middle. The mode line now adds that the V parking is window 2 in the section table and the vi pedal underpins the C-Aeolian bridge in window 4. Read the spine first, then the section table as its modal coloring.
Shiner-TheMutiny-FROMCDN.mid full-vocal, 196 notes| Window | Beat range | Verdict | Score | Gap % | Runner-up | Top 5 pitch classes |
|---|---|---|---|---|---|---|
| 1 | head | B♭ Phrygian | 0.1241 | 0.8 low | B♭ Aeolian | B♭, G♭, F, G, D♭ |
| 2 | verse | G♭ Ionian | 0.1833 | 0.0 low, tied with G♭ Lydian | G♭ Lydian | G♭, D♭, B♭, F, D |
| 3 | chorus | F Aeolian | 0.1382 | 0.5 low | F Dorian | F, A♭, C, D♭, D |
| 4 | bridge | F Phrygian | 0.0884 | 7.4 medium | B♭ Aeolian | F, B♭, G♭, A♭, C |
| 5 | outro | G♭ Lydian | 0.1293 | 3.3 low | B♭ Aeolian | B♭, G♭, D♭, F, A |
What the section table actually says. No window calls B♭ Aeolian directly. The song-wide verdict is the integral of these windows, not a duplicate of any one of them. Window 2 pivots to G♭ (the ♭VI of B♭) as the apparent center. Window 3 and 4 pivot to F (the 5 of B♭) with Phrygian color. Window 5 returns to G♭. The verdict-by-section here is less about modal drift than about tonal-center drift: the song keeps pivoting between B♭, G♭, and F as the chord roots take the foreground.
Drop-C rhythm guitar dyads on this song play rootless flat-VI and flat-VII voicings that ring loud. Any whole-mix histogram weights those notes heavily and pulls the verdict to minor. The vocal track strips that bias out and exposes the third the singer is actually using. The 1.7-to-1 D-over-D♭ ratio in the vocal-only run reproduces the chapter card's 16.9 versus 9.1 percent finding within rounding error.
G♭ 0.202 B♭ 0.200 F 0.191 A♭ 0.117 D♭ 0.082 C 0.073 G 0.038 E♭ 0.034
G♭ (the ♭6 / ♭VI), B♭ (the root), and F (the 5) are essentially tied at the top. That is the chapter card's "tonic ringed by ♭VI and ♭VII" claim showing up directly in the histogram. The 4 (E♭) is almost absent, which is why some windows tip Phrygian: rootless pentatonic Aeolian with no 4 is one step from Phrygian-flavored Aeolian.
| Rank | Root | Mode | Score |
|---|---|---|---|
| 1 | D | Phrygian (cadence-aligned) | 0.0356 |
| 2 | B♭ | Ionian (the chapter card's call) | 0.0330 |
| 3 | B♭ | Aeolian | 0.0279 |
| 4 | D | Locrian | 0.0209 |
| 5 | B♭ | Mixolydian | 0.0205 |
The Mutiny chapter card already names both layers: the loud Drop-C dyads ringing ♭VI and ♭VII, and the singer landing the major third. The V3 add is just to say that both readings produce a defensible mode label, B♭ Aeolian for the band layer and B♭ major for the voice layer, and the song lives in the gap between them rather than in one mode or the other.
My Mirror Hates Me, Lazarus, Endless Summer, and Not Too Much. Each pass uses the sectionalized detector and the full-vocal source MIDI where one exists.