4:07 at 0:34:28
20 second sample from a quarter of the way in. The delivered file is the full take at 24 bit.
01_0-34-29_take28_score88.wav · 62 MB · re-check: clean
Band practice, 8.14.26
Three hours and forty five minutes of tape, cut down to the takes that hold up. Every keeper runs longer than two minutes. False starts, drills, and dead air stayed behind.
19songs kept
148minmusic delivered
83takes found
78mincut away
24bit44.1 kHz wav
The session at a glance
Each bar below is one take, placed where it falls in the tape. Green bars became files.
1h2h3h
fragment (54)song (19)workshop (7)false start (3)
The 19 keepers
Ordered by where they land in the session. Each card carries the score breakdown that decided it, so any verdict here can be checked against the audio.
88
2:27 at 0:38:38
20 second sample from a quarter of the way in. The delivered file is the full take at 24 bit.
02_0-38-39_take29_score67.wav · 37 MB · re-check: FLAG
3:42 at 0:43:16
20 second sample from a quarter of the way in. The delivered file is the full take at 24 bit.
03_0-43-17_take31_score90.wav · 56 MB · re-check: clean
3:39 at 0:47:02
20 second sample from a quarter of the way in. The delivered file is the full take at 24 bit.
04_0-47-03_take32_score90.wav · 55 MB · re-check: clean
5:03 at 0:55:27
20 second sample from a quarter of the way in. The delivered file is the full take at 24 bit.
05_0-55-28_take36_score82.wav · 77 MB · re-check: clean
10:54 at 1:05:46
20 second sample from a quarter of the way in. The delivered file is the full take at 24 bit.
1.1s of dead air cut from the head
06_1-05-46_take41_score80.wav · 165 MB · re-check: FLAG
4:20 at 1:21:10
20 second sample from a quarter of the way in. The delivered file is the full take at 24 bit.
07_1-21-11_take47_score88.wav · 66 MB · re-check: clean
4:35 at 1:26:46
20 second sample from a quarter of the way in. The delivered file is the full take at 24 bit.
08_1-26-47_take49_score84.wav · 69 MB · re-check: FLAG
5:01 at 1:31:34
20 second sample from a quarter of the way in. The delivered file is the full take at 24 bit.
09_1-31-35_take50_score84.wav · 76 MB · re-check: clean
4:58 at 1:36:37
20 second sample from a quarter of the way in. The delivered file is the full take at 24 bit.
10_1-36-38_take51_score92.wav · 75 MB · re-check: FLAG
9:38 at 1:43:27
20 second sample from a quarter of the way in. The delivered file is the full take at 24 bit.
11_1-43-28_take54_score83.wav · 146 MB · re-check: clean
9:16 at 1:55:00
20 second sample from a quarter of the way in. The delivered file is the full take at 24 bit.
12_1-55-01_take57_score80.wav · 140 MB · re-check: clean
2:27 at 2:11:22
20 second sample from a quarter of the way in. The delivered file is the full take at 24 bit.
13_2-11-23_take66_score79.wav · 37 MB · re-check: FLAG
3:12 at 2:13:58
20 second sample from a quarter of the way in. The delivered file is the full take at 24 bit.
1.6s of dead air cut from the head
14_2-13-58_take67_score81.wav · 49 MB · re-check: FLAG
8:19 at 2:19:27
20 second sample from a quarter of the way in. The delivered file is the full take at 24 bit.
15_2-19-28_take72_score76.wav · 126 MB · re-check: clean
35:27 at 2:30:19
20 second sample from a quarter of the way in. The delivered file is the full take at 24 bit.
3.2s of dead air cut from the head / one continuous stretch, give it a listen ahead of any split
16_2-30-17_take74_score79.wav · 537 MB · re-check: FLAG
24:42 at 3:06:46
20 second sample from a quarter of the way in. The delivered file is the full take at 24 bit.
one continuous stretch, give it a listen ahead of any split
17_3-06-47_take77_score80.wav · 374 MB · re-check: clean
2:12 at 3:33:27
20 second sample from a quarter of the way in. The delivered file is the full take at 24 bit.
18_3-33-28_take80_score71.wav · 33 MB · re-check: clean
3:50 at 3:35:48
20 second sample from a quarter of the way in. The delivered file is the full take at 24 bit.
19_3-35-49_take81_score72.wav · 58 MB · re-check: clean
Two extra long takes
Give these two your ear first.
Files 16 and 17 run 35 and 25 minutes. The detector did not fail here. Measured against the
room floor, file 17 holds zero breaks longer than two seconds across its whole length, and
file 16 holds one three second break at the eight minute mark. The band played straight through
both. If either one covers several ideas you want apart, tell me where to place the cut.
Both are now split. Boundary analysis found 13 song changes across the two takes, so 60 minutes came out as 15 tracks. Every cut has 40 seconds of audio across it on the split page, with a vote button per cut.
Good takes under two minutes
Your two minute floor excluded these twelve. Each one scored like a real take. Take 25 stands out, since it is the successful retry of the strongest false start on the tape.
| Take | Starts | Length | Score | Labelled |
|---|---|---|---|---|
| 25 | 0:29:53 | 74s | 76 | fragment |
| 48 | 1:25:46 | 56s | 73 | fragment |
| 55 | 1:53:13 | 81s | 72 | fragment |
| 53 | 1:42:20 | 63s | 69 | fragment |
| 18 | 0:20:21 | 86s | 68 | workshop |
| 78 | 3:31:31 | 76s | 66 | fragment |
| 73 | 2:28:41 | 75s | 63 | fragment |
| 42 | 1:17:26 | 84s | 61 | fragment |
| 27 | 0:33:40 | 45s | 61 | fragment |
| 59 | 2:05:36 | 91s | 60 | fragment |
| 11 | 0:10:24 | 47s | 60 | fragment |
| 12 | 0:15:58 | 51s | 60 | fragment |
Say the word and any of these ship as wav files too.
False starts that got caught
Three takes were proven aborted by a later take replaying their opening. The number is the sequence match between the two heads.
| Take | Starts | Length | Evidence |
|---|---|---|---|
| 3 | 0:05:51 | 25s | take 5 replays this head, sequence match 0.67, so this attempt was aborted |
| 15 | 0:17:59 | 29s | take 18 replays this head, sequence match 0.64, so this attempt was aborted |
| 24 | 0:28:13 | 32s | take 25 replays this head, sequence match 0.94, so this attempt was aborted |
How the cut was made
Source ran 44.1 kHz and 24 bit stereo across 3.58 GB. The detector made one decode pass, tracked pulse, split the tape into 83 takes, then scored each one on length, section changes, pulse strength, continuity, ending shape, and tempo stability.
A take was proven aborted when a later take replayed its opening in the same pitch order. Sequence order carries the evidence, since an average across pitches would match any two songs sharing a key.
Every delivered file was then re-decoded from scratch and measured against the room noise floor of -40.2 dB, so no file counts as verified on the strength of the same numbers that selected it. Three files had leading dead air removed by hand after that check.