jb/paseo
1
0
Fork 0
Code Issues 1 Pull Requests Actions Packages Projects Releases Wiki Activity

refactor(app): replace voice hooks with voice runtime and audio engine architecture

Browse Source
This commit is contained in:
Mohamed Boudra
2026-03-13 20:53:55 +07:00
parent 730f34753a
commit 80a2f8052c
51 changed files with 4308 additions and 2380 deletions
Download Patch File Download Diff File Expand all files Collapse all files
packages/app/src/hooks/use-dictation-audio-source.web.ts
+1 -78
View File
@@ -1,4 +1,5 @@
import { useCallback, useEffect, useMemo, useRef, useState } from "react";
import { parsePcm16Wav } from "@/utils/pcm16-wav";
import { getTauri } from "@/utils/tauri";
import type { DictationAudioSource, DictationAudioSourceConfig } from "./use-dictation-audio-source.types";
@@ -65,84 +66,6 @@ const int16ToFloat32 = (input: Int16Array): Float32Array => {
return out;
};
type Pcm16Wav = {
sampleRate: number;
samples: Int16Array;
};
const parsePcm16Wav = (buffer: ArrayBuffer): Pcm16Wav | null => {
if (buffer.byteLength < 44) {
return null;
}
const view = new DataView(buffer);
const readAscii = (offset: number, length: number): string => {
let out = "";
for (let i = 0; i < length; i += 1) {
out += String.fromCharCode(view.getUint8(offset + i));
}
return out;
};
if (readAscii(0, 4) !== "RIFF" || readAscii(8, 4) !== "WAVE") {
return null;
}
let offset = 12;
let channels = 0;
let sampleRate = 0;
let bitsPerSample = 0;
let dataOffset = 0;
let dataSize = 0;
while (offset + 8 <= buffer.byteLength) {
const chunkId = readAscii(offset, 4);
const chunkSize = view.getUint32(offset + 4, true);
const chunkDataOffset = offset + 8;
if (chunkDataOffset + chunkSize > buffer.byteLength) {
break;
}
if (chunkId === "fmt " && chunkSize >= 16) {
const audioFormat = view.getUint16(chunkDataOffset, true);
channels = view.getUint16(chunkDataOffset + 2, true);
sampleRate = view.getUint32(chunkDataOffset + 4, true);
bitsPerSample = view.getUint16(chunkDataOffset + 14, true);
if (audioFormat !== 1) {
return null;
}
} else if (chunkId === "data") {
dataOffset = chunkDataOffset;
dataSize = chunkSize;
break;
}
offset = chunkDataOffset + chunkSize + (chunkSize % 2);
}
if (!dataOffset || !dataSize || sampleRate <= 0 || bitsPerSample !== 16 || channels <= 0) {
return null;
}
const sampleCount = Math.floor(dataSize / 2);
const interleaved = new Int16Array(buffer, dataOffset, sampleCount);
if (channels === 1) {
return { sampleRate, samples: new Int16Array(interleaved) };
}
const frameCount = Math.floor(interleaved.length / channels);
const mono = new Int16Array(frameCount);
for (let frame = 0; frame < frameCount; frame += 1) {
let sum = 0;
for (let ch = 0; ch < channels; ch += 1) {
sum += interleaved[frame * channels + ch] ?? 0;
}
mono[frame] = Math.round(sum / channels);
}
return { sampleRate, samples: mono };
};
const int16ToBase64 = (pcm: Int16Array): string => {
const bytes = new Uint8Array(pcm.buffer, pcm.byteOffset, pcm.byteLength);
let binary = "";