mirror of
https://github.com/snapotter-hq/SnapOtter.git
synced 2026-08-03 07:46:42 +02:00
SnapOtterandGitHub
1548d475ca
ffmpeg's loudnorm filter runs internally at 192 kHz and emits 192 kHz unless the chain resamples back, so normalize-audio produced 192 kHz output (~4.3x larger files) regardless of the input rate. Append aresample to restore the input's sample rate; runMediaTool now exposes the input audio sample rate to its args callback.
39 lines
1.3 KiB
TypeScript
39 lines
1.3 KiB
TypeScript
import { extname } from "node:path";
|
|
import type { FastifyInstance } from "fastify";
|
|
import { z } from "zod";
|
|
import { audioOutputFor, runMediaTool } from "../../lib/media-tool.js";
|
|
import { createToolRoute } from "../tool-factory.js";
|
|
|
|
const settingsSchema = z.object({});
|
|
|
|
export function registerNormalizeAudio(app: FastifyInstance) {
|
|
createToolRoute(app, {
|
|
toolId: "normalize-audio",
|
|
settingsSchema,
|
|
process: async () => {
|
|
throw new Error("normalize-audio is v2-only");
|
|
},
|
|
processV2: async (ctx) => {
|
|
const origExt = extname(ctx.inputs[0].filename) || ".mp3";
|
|
const out = audioOutputFor(origExt);
|
|
const base = ctx.inputs[0].filename.replace(/\.[^.]+$/, "");
|
|
const outName = `${base}_normalized${out.ext}`;
|
|
|
|
const { outPath } = await runMediaTool(ctx, outName, (inPath, outPath, info) => {
|
|
// loudnorm runs internally at 192 kHz and emits at 192 kHz unless we
|
|
// resample back, so restore the source rate to avoid inflating the file.
|
|
const sr = info.audioSampleRate ?? 44100;
|
|
return [
|
|
"-i",
|
|
inPath,
|
|
"-af",
|
|
`loudnorm=I=-16:TP=-1.5:LRA=11,aresample=${sr}`,
|
|
...out.encodeArgs,
|
|
outPath,
|
|
];
|
|
});
|
|
return { scratchPath: outPath, filename: outName, contentType: out.contentType };
|
|
},
|
|
});
|
|
}
|