A | B | C | D | E | F | G | H | I | J | K | L | M | N | O | P | Q | R | |
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
1 | Date | Description | Dataset | Data | Epochs | Train | Converged? | Score | Silent | Wind | Tones | Voice | Rhythmic | Gnarly | Garbled | Hellish | Noise | Comments |
2 | 2018 02 19 | cee-lo "fuck you" acc | a-fuck | 556 | 10 | 2.3 | ok | 3 | X | very strained ceelo-like vocals | ||||||||
3 | 2018 02 16 | (h)ywfd mixed acapellas | ywfdacc | 476 | 15 | 3.0 | ok | 3 | X | very windy - many voices mixed in background. occasional vowel overlap | ||||||||
4 | 2018 02 26 | school chant 1 | school1 | 521 | 8 | 1.4 | ok | 3 | X | X | hoarse vocal noises | |||||||
5 | 2018 02 26 | school chant 2 | school2 | 511 | 8 | 1.4 | ok | 3 | X | X | early monster vocalizations, then breathy/fry voices | |||||||
6 | 2018 01 17 | stacked "ahs" form major chord | bigahchord | 629 | 15 | 3.7 | ok | 3 | X | X | renders had areas with clear voices, but sometimes were overtaken by LOUD static | |||||||
7 | 2018 02 28 | school chant tunafish | tunafish | 500 | 12 | 1.3 | bad | 1 | X | sounds underwater and demonic, turns into wordlike tones | ||||||||
8 | 2018 03 03 | ae "vi scose poise" | scosepoise | 509 | 64 | 0.9 | great! | 6 | X | X | highly developed scose poise... | |||||||
9 | 2018 02 26 | jlin drums | jlin4 | 655 | 80 | 0.5 | great! | 6 | X | X | slight improvements but tbh this one is done | |||||||
10 | 2018 02 26 | sascha kuschel excerpt | kuschel | 466 | 27 | 1.9 | great! | 6 | X | X | making normal sounds again. didn't do anything different. fluke? | |||||||
11 | 2018 01 19 | tap glass with spoon, layered | glassmix | 1309 | 17 | 1.1 | great! | 6 | X | X | REALLY good. lots of interesting variations. rendering at higher resolution helped - double the data! | |||||||
12 | 2018 02 08 | glassloop_left pruned | glass_space1 | 556 | 22 | 2.0 | great! | 6 | X | noise decreases the further it trains.. but there is more silence and the breaking is further apart! | ||||||||
13 | 2018 02 16 | track 'fm9' | fm9 | 540 | 14 | 1.8 | great | 5 | X | X | X | mostly tone runs from the beginning of the track. funny bass/vocal tone in the background | ||||||
14 | 2018 02 01 | flufdril_2 just drums @172bpm | flufdril_2a_drums | 745 | 33 | 1.4 | great | 5 | X | X | more of the same, still pretty spaced out | |||||||
15 | 2018 02 01 | plastic cut drums @172bpm | plasticcut_4 | 600 | 30 | 1.6 | great | 5 | X | X | likes weird bass notes | |||||||
16 | 2018 02 26 | vsnares "sundae" | sundae | 411 | 70 | 2.4 | great | 5 | X | X | no real change, producing weird rhythms | |||||||
17 | 2018 02 16 | hito "this is all i can say" | thisisallicansay | 495 | 71 | 1.7 | great | 5 | X | X | gets better at generating the phrase regularly, but still breaks up a bit | |||||||
18 | 2018 03 03 | lovelove loop | lovelove | 508 | 17 | 1.3 | great | 5 | X | X | produces "love love love" chants with odd spacing | |||||||
19 | 2018 03 03 | lyra bladerunner vocals | lyraloop | 503 | 8 | 0.7 | great | 5 | X | X | really good singing! doesn't fixate so much. | |||||||
20 | 2018 02 21 | autechre known | known | 508 | 14 | 1.5 | great | 5 | X | more of the same tones.. semi-evolving but not super varied | ||||||||
21 | 2018 02 21 | autechre seeonsee | seeonsee | 518 | 14 | 2.0 | great | 5 | X | very consistent.. generates legato changes. only a couple constant sticky tones. | ||||||||
22 | 2018 02 21 | henry cow "teenbeat" | teenbeat | 503 | 24 | 2.2 | great | 5 | X | tense trumpet tones.. cymbals and rhythms evolve.. | ||||||||
23 | 2018 02 19 | avril "what the hell" acc | a-hell | 509 | 14 | 2.1 | good | 4 | X | consistent vocals, piercing noise blobs, could be cleaned up | ||||||||
24 | 2018 01 22 | ensemble compressed | ensemble2 | 387 | 8 | 2.9 | good | 4 | X | much better than previous ensemble training. compression helped, but didn't saturate. | ||||||||
25 | 2018 02 28 | psalm compressed | psalm_comp | 505 | 12 | 2.1 | good | 4 | X | gets the guy's voice well, avoids sustained tones, but kind of random | ||||||||
26 | 2018 02 18 | "we found love" song | wefoundlove-song | 506 | 11 | 3.8 | good | 4 | X | random rhythms, a vague build. the organ stab is prominent toward the end | ||||||||
27 | 2018 02 28 | f.a.t.y. drums | jlin-faty | 519 | 20 | 1.3 | good | 4 | X | X | sounds promising | |||||||
28 | 2018 02 28 | f.a.t.y. voice | holly-faty | 509 | 10 | 1.9 | good | 4 | X | X | X | diverse loops and stuff, fwiw | ||||||
29 | 2018 02 26 | "can't sleep" vocal | cantsleep_vocal | 503 | 28 | 1.0 | good | 4 | X | X | X | fixates, no change | ||||||
30 | 2018 02 01 | uneasy melody | uneasy | 497 | 28 | 1.0 | good | 4 | X | X | surprisingly generated sections of 3-4 notes | |||||||
31 | 2018 02 08 | re-rendered just quiet glass | glass_space2 | 544 | 14 | 1.6 | good | 4 | X | very few crashes - mostly little pops and things | ||||||||
32 | 2018 03 03 | holly bladerunner chorus | hhbr | 510 | 8 | 1.1 | good | 4 | X | disordered vocal shapes, fixates but has cool vibrato / crescendo | ||||||||
33 | 2018 01 17 | all 3 singing "ah" | ah_mix_long | 635 | 16 | 2.4 | good | 4 | X | 1st epoch already sounded good. 16th epoch SUPER clear, could hear everyone's voices. | ||||||||
34 | 2018 02 26 | "can't sleep" source mix | cantsleep_mix | 512 | 32 | 2.1 | good | 4 | X | fixates, no change | ||||||||
35 | 2018 02 26 | "can't sleep" synth | cantsleep_synth | 511 | 24 | 2.2 | good | 4 | X | fixates, no change | ||||||||
36 | 2018 02 12 | rashad becker "dances v" | dancesv | 454 | 61 | 1.0 | good | 4 | X | no real development or new structure | ||||||||
37 | 2018 02 25 | susumu "hagoromo" edit | susumu4 | 437 | 14 | 0.7 | good | 4 | X | fixates, harp/violin tones | ||||||||
38 | 2018 03 03 | lucierization of the office | lucier | 511 | 15 | 0.7 | good | 4 | X | whistly, ghostly... interesting upper harmonics develop. mostly a big air cloud. | ||||||||
39 | 2018 02 21 | brigitte fontaine "l'ete l'ete" | brigitte | 504 | 24 | 1.4 | good | 4 | X | generated a lot of vocal loops and horn hits | ||||||||
40 | 2018 02 05 | glass foley (stretch) | glassloop_stretch | 460 | 14 | 3.2 | good | 4 | X | less whistling by the end - probably the best? | ||||||||
41 | 2018 01 20 | ensemble singing chords | omnir39 | 260 | 12 | 2.1 | ok | 3 | X | X | very noisy, learned some phrasing. maybe the most promising once we get the dataset together | |||||||
42 | 2018 02 21 | nsync "bye bye bye" | byebyebye | 508 | 16 | 3.0 | ok | 3 | X | some swipey nsync sounds, mostly caught in tones | ||||||||
43 | 2018 02 19 | cata talking | catamachine | 548 | 22 | 1.2 | ok | 3 | X | X | not much development | |||||||
44 | 2018 02 16 | cata's 'clamstep' | clamstep | 568 | 8 | 3.6 | ok | 3 | X | X | midrange whistle - like a howl - some shapes in the background resembling original sound | |||||||
45 | 2018 01 23 | clouds35b | clouds2 | 649 | 4 | 0.7 | good | 3 | X | X | X | produced bubbily notes like the original. later training epochs popped less | ||||||
46 | 2018 02 19 | jules + cata simul talking | freebsd | 523 | 24 | 1.4 | ok | 3 | X | X | strained cycles | |||||||
47 | 2018 02 13 | susumu album | susumu2 | 504 | 19 | 1.0 | ok | 3 | X | sustained harp stuff. a little schizo given it trained on bits of 12 songs. some good tones at the end | ||||||||
48 | 2018 01 23 | jlin master mix | jlin3 | 679 | 4 | 1.4 | good | 3 | X | X | X | drum hit + vocal, more coherent sequences, bass sound | ||||||
49 | 2018 02 01 | shlucht "stolen pics" | stolenpics | 418 | 26 | 1.7 | good | 3 | X | cheeky tone is still prominent. harmonizes with itself once?? | ||||||||
50 | 2018 02 01 | rashad becker "dances" | dances | 590 | 60 | 1.7 | ok | 3 | X | found a couple tones it likes, on the whole seems too complex to train on.. | ||||||||
51 | 2018 02 07 | rec1 compressed | glass_rec1_comp | 446 | 8 | 1.6 | ok | 3 | X | best so far of the ones i recorded. a bit whistly, noisy | ||||||||
52 | 2018 01 20 | holly granular stuff | hollydream | 753 | 12 | 2.0 | ok | 3 | X | given this was a lot of granular/fft freeze stuff it produced similar material with no real transitions | ||||||||
53 | 2018 01 23 | jlin holly vocal overlapped | jlin2 | 464 | 4 | 2.3 | good | 3 | X | wandering noisy vocals | ||||||||
54 | 2018 02 26 | school chant 3 | school3 | 483 | 8 | 1.2 | ok | 3 | X | boy vocalizations | ||||||||
55 | 2018 02 15 | susumu "azukiiro no kaori" | susumu3 | 529 | 10 | 0.8 | ok | 3 | X | nice drones, got vocal, but also loud bursts of noise randomly | ||||||||
56 | 2018 02 01 | vibra tres @172bpm | vibratres | 492 | 26 | 0.6 | ok | 3 | X | starts getting more active. still pretty indecisive, never got the structure | ||||||||
57 | 2018 02 05 | glass foley (L/R/mix) | glassloop | 460 | 14 | 2.9 | ok | 3 | X | produced glass sounds, all on top of each other, some whistling | ||||||||
58 | 2018 02 05 | glass foley (L) | glassloop_left | 460 | 14 | 3.0 | ok | 3 | X | whistly and vague | ||||||||
59 | 2018 01 23 | jlin mix, drums, vocal in sequence | jlin1 | 500 | 4 | 1.4 | ok | 3 | X | drum hit + vocal drone | ||||||||
60 | 2018 01 26 | ensemble hocket | chapter3 | 731 | 8 | 1.9 | ok | 3 | X | more training didn't help! just made it more airy | ||||||||
61 | 2018 01 27 | ensemble hocket | chapter3 | 731 | 5 | 2.3 | ok | 3 | X | re-training but stopping sooner. got some maniacal flute stuff toward the end | ||||||||
62 | 2018 02 07 | space3 reversed | glass_space_rev | 526 | 8 | 1.3 | ok | 3 | X | some good renders, no whistling, a lot of silence | ||||||||
63 | 2018 02 13 | lonne loop | lonneloop | 559 | 19 | 1.1 | ok | 3 | X | very very bassy. ignored most of the song. | ||||||||
64 | 2018 02 21 | waka "o lets do it" | oletsdoit | 498 | 22 | 2.0 | ok | 3 | X | mostly quiet static... almost sounds like data-loading / hard disk noise | ||||||||
65 | 2018 03 04 | mix of bladerunner vox | brmix | 497 | 8 | 1.2 | ok | 3 | X | X | fixates, lyra voice dominates | |||||||
66 | 2018 03 04 | psalm, 55 ms grains | psalm055 | 509 | 8 | 1.5 | ok | 3 | X | X | skips around, likes producing a couple voices only | |||||||
67 | 2018 03 04 | psalm, 115 ms grains | psalm115 | 509 | 8 | 1.5 | ok | 3 | X | X | fixates a bit more | |||||||
68 | 2018 03 04 | psalm, 234 ms grains | psalm234 | 509 | 8 | 1.5 | ok | 3 | X | X | fixates more | |||||||
69 | 2018 03 04 | psalm, 468 ms grains | psalm468 | 509 | 8 | 1.5 | ok | 3 | X | X | fixates on a couple quiet notes | |||||||
70 | 2018 02 28 | psalm cut up rearranged | psalm_cut | 508 | 12 | 1.7 | ok | 2 | X | certain epochs were very noisy. moves breathily between passages | ||||||||
71 | 2018 02 01 | cata's harsh noise | textile | 446 | 19 | 4.0 | ok | 2 | X | high pitched sound everywhere but the noise bursts more interesting | ||||||||
72 | 2018 01 22 | holly tts training sentences | hollytrainset | 733 | 8 | 2.4 | ok | 2 | X | X | got some voice, but the more it trained, the more it turned into noise. epochs 7-8 the best | |||||||
73 | 2018 02 26 | "main artery of the mother" | artery | 496 | 16 | 0.5 | ok | 2 | X | X | staticky, irregular heartbeat | |||||||
74 | 2018 03 03 | really short countdown | countdown | 500 | 7 | 0.8 | ok | 2 | X | X | struggles to say "one" | |||||||
75 | 2018 02 26 | harp of king david | harpdavid | 510 | 14 | 0.9 | ok | 2 | X | fixates, on tones | ||||||||
76 | 2018 02 06 | space3 louder, roll off highs | glass_space4 | 573 | 8 | 1.3 | ok | 2 | X | less defined high end, whistles, not very glassy | ||||||||
77 | 2018 02 05 | glass foley (L compressed) | glassloop_comp | 460 | 14 | 3.1 | ok | 2 | X | very whistly | ||||||||
78 | 2018 02 16 | "we found love" acc | wefoundlove | 476 | 14 | 2.8 | ok | 2 | X | high pitched tone throughout. eventually stops rendering voice | ||||||||
79 | 2018 02 18 | liam talking (18 seconds) | liam | 451 | 14 | 0.9 | ok | 2 | X | mostly silence. some voice. seems like it picked up on the split-second of very quiet music mostly! | ||||||||
80 | 2018 01 25 | jlin drums processed | hhprocessedjlindrums | 524 | 4 | 4.8 | bad | 1 | X | turned into quickly-fading noise. might need a long time to come into focus | ||||||||
81 | 2018 01 28 | dj shlucht pocket mix | 494 | 12 | 2.3 | bad | 1 | X | harsh noise with some structure | |||||||||
82 | 2018 02 13 | susumu track | susumu1 | 488 | 19 | 0.9 | bad | 1 | X | bits of harps, a lot of loud noise though. why did this train poorly? | ||||||||
83 | 2018 03 03 | inoj loveyoudown vox | loveyoudown | 516 | 8 | 2.1 | bad | 1 | X | occasional mumbled words... mostly noise | ||||||||
84 | 2018 01 22 | clouds35b | clouds | 472 | 8 | 0.8 | bad | 1 | X | maybe the samples were too compressed? very cloudy noises, weird bass shapes | ||||||||
85 | 2018 02 06 | space2 pruned | glass_space3 | 588 | 8 | 1.3 | bad | 1 | X | really jumbled and noisy | ||||||||
86 | 2018 02 19 | kesha "die young" acc | a-die | 524 | 10 | 2.7 | bad | 1 | X | turns into a high pitched tone | ||||||||
87 | 2018 02 08 | rec1 compress + slow | glass_rec1_slow | 484 | 8 | 0.9 | bad | 1 | X | sort of glassy but focused on repeating glass tones | ||||||||
88 | 2018 01 20 | holly glissando + ah concat | gliss_and_ah | 432 | 12 | 2.2 | bad | 1 | X | the constant tones made it focus on certain notes. "hallucinated" a deep "ah" | ||||||||
89 | 2018 01 17 | many glissandos mixed together | gliss_long_mix | 1164 | 15 | 2.9 | bad | 1 | X | loud, harsh, distinct squeak at ?? hz - vague "ah" in the background. further renders all the same. | ||||||||
90 | 2018 02 06 | space4 timestretch | glass_space5 | 573 | 8 | 1.3 | bad | 1 | X | whistly, wet mud sound | ||||||||
91 | 2018 01 14 | slowed down glass sounds | hito glass | 164 | 600 | 2.0 | bad | 1 | X | turned into pitched wind sounds and bursts of noise | ||||||||
92 | 2018 01 21 | ensemble going "ah ah ah" | ensemble | 1962 | 12 | 0.9 | bad | 1 | X | generated mostly silence | ||||||||
93 | 2018 01 26 | 400 kick drums | kiq | 428 | 1 | 0.1 | bad | 1 | X | 256 quantization steps. nothing rendered. too much silence! | ||||||||
94 | 2018 01 26 | 400 kick drums | kiq | 428 | 1 | 0.2 | bad | 1 | X | 512 quantization steps. ditto | ||||||||
95 | 2018 02 28 | madli reading in estonian | madli | 508 | 2 | 2.2 | bad | 1 | X | X | X | a few words toward the beginning - really scattered - trained 2x and it was noisier but not better | ||||||
96 | 2018 02 07 | broken glass recorded | glass_rec1 | 476 | 8 | 0.8 | bad | 0 | X | noise - glassy in aspect - doesn't improve | ||||||||
97 | 2018 02 07 | rec1 repitch and overlap | glass_rec1_mix | 471 | 8 | 2.4 | bad | 0 | X | noise - glassy in aspect - doesn't improve | ||||||||
98 | 2018 02 07 | rec1 reversed | glass_rec1_rev | 471 | 6 | 0.9 | bad | 0 | X | silence, occasional glassy noise which sounds poorly trained |