Free self-hosted STT for VRChat. https://git.yummers.dev/TaSTT.git/atom?h=v0.1 2022-12-31T21:17:36+00:00 2022-12-31T21:17:36+00:00 yum yum.food.vr@gmail.com 2022-12-31T21:16:47+00:00 urn:sha1:841a6a87b0ba00576c11fc64833b61ad9ade956e Re-paging anything on screen N causes screens N+1...infinity to completely re-page. This fixes cases where we go back and draw something at the bottom of the board, and it never gets overwritten. 2022-12-30T10:33:07+00:00 yum yum.food.vr@gmail.com 2022-12-30T10:25:50+00:00 urn:sha1:d9a0224466189d0fe1d46f21d586fa8a1a58c687 Boards whose size is an even multiple of CHARS_PER_SYNC would lose the entire last region. * Attempt to fix runaway memory usage of GUI text frames, but this needs more work 2022-12-30T09:35:11+00:00 yum yum.food.vr@gmail.com 2022-12-30T09:35:11+00:00 urn:sha1:9a7190a04bbe6feba2abe4a1590eb8114c04f683 Users can pick longer transcription durations for accuracy-critical tasks, or shorter durations for latency-critical tasks. 2022-12-30T09:10:32+00:00 yum yum.food.vr@gmail.com 2022-12-30T09:10:32+00:00 urn:sha1:d1024fef1b216af5d3d991228c6b83311a71bb42 VRChat won't update the FX layer associated with an avatar unless its GUID changes. Delete the GUID file when overwriting our generated FX layer to work around this. * Change paging behavior: when a region is updated, we re-page everything that comes after it. This fixes the issue where we go back to update something, then jump back to the current screen, leaving some random chunk of text somewhere on the board. * Reduce transcription time from 28s to 10s. I'm going to expose this to the user since there's a fundamental latency/stability tradeoff here. 2022-12-30T08:01:28+00:00 yum yum.food.vr@gmail.com 2022-12-30T08:01:28+00:00 urn:sha1:abdaa7ce215086bf1070d6093731cd35df866cbb Bump up recording window to 28 seconds. This helps a lot with long-form transcription tasks, s.a. transcribing an audiobook. We should expose this as a parameter, since at 10s the transcription delay is typically 300ms, while at 28s it's typically 1.1-1.2s. 2022-12-30T07:01:09+00:00 yum yum.food.vr@gmail.com 2022-12-30T07:01:09+00:00 urn:sha1:12bcd1d40df21cb9bad6ae353ea4f6697e1275a6 Users can now control how many letters wide and tall the board is. Tested at 4x48, 5x60, 10x120, and 20x240. At 20x240, Unity freezes and does not make forward progress. Perhaps creating 4800 float parameters isn't a truly scalable interface. 2022-12-30T02:49:24+00:00 yum yum.food.vr@gmail.com 2022-12-30T02:47:57+00:00 urn:sha1:7214653f9354ef885ef4131e455e0a63f12a17f2 Now it's possible to generate shaders with a custom number of rows, columns, and bytes per character. All edits to the shader should go through TaSTT_template.shader. To generate a new shader from the template: $ ./Scripts/generate_shader.py \ --bytes_per_char 2 \ --rows 1 \ --cols 12 --shader_template $(pwd)/Shaders/TaSTT_template.shader \ --shader_path $(pwd)/Shaders/TaSTT.shader 2022-12-30T01:20:39+00:00 yum yum.food.vr@gmail.com 2022-12-30T01:17:49+00:00 urn:sha1:06aece2ebade3d25c16fe1c143fba5cd472d004c Users can now see the number of avatar parameter bits they'll use prior to committing. 2022-12-29T23:28:32+00:00 yum yum.food.vr@gmail.com 2022-12-29T23:28:32+00:00 urn:sha1:618487e712f15d92ac9b6655c892cd94ff59d656 An off-by-one issue in numRegions() would result in one extra layer trying to drive a letter in the last region, which would wrap back around to the 0th character slot (cell). * GUI explicitly logs when it's done generating avatar stuff * OSC layer no longer tries to update cells which don't exist 2022-12-29T22:47:49+00:00 yum yum.food.vr@gmail.com 2022-12-29T22:47:49+00:00 urn:sha1:0b5c3841b5c2ed99b718c880bf0f161ac4db0788 The transcription engine beeps when you start/stop transcribing so you know that it's listening. Users can now disable this. * add help text to all input fields in GUI * make TaSTT generated file textctrls readonly, since I haven't tested them being reassigned * document idea to configure unity & transcription apps with config files * controller input thread no longer crashes if steamvr isn't running, it just slowly spins and waits * when you stop transcribing, the transcription engine re-transcribes a few times. I think this should improve end-of-transcription tail latencies * transcribe.py now prints out its args