3d01ff4db1
PatchScript.GestureDescriptor had no case for n, so it fell to the alveolar approximant (no velum). IpaGestureExperiment.BuildWordScript runs gestures back to back on the same scaffold as BuildCandidateScript (refactored onto one Scaffold). tools/WordRender renders a word offline through native libfaust, driving surface sliders from ControlSplineTimeline because the Faust exporter's summed spline offsets do not sequence gestures. Two tests. Nothing was heard: the render is a near-pure sine with no vowel quality. Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
87 lines
4.5 KiB
C#
87 lines
4.5 KiB
C#
using System.Runtime.CompilerServices;
|
|
using System.Runtime.InteropServices;
|
|
using System.Security.Cryptography;
|
|
using System.Text.Json;
|
|
using AquaSynth.Dsl;
|
|
using AquaSynth.Faust;
|
|
|
|
// Offline probe: one word as a run of IPA gestures on the IpaGestureExperiment scaffold, rendered through the native
|
|
// libfaust path to a mono 16-bit WAV plus script.aqua, source.dsp and manifest.json. It measures peak, rms and NaN;
|
|
// it cannot say how the word sounds.
|
|
// usage: WordRender <outDir> <label> <freq-hz> <ipa>:<ms>:<intensity> ... e.g. WordRender out nibu 140 n:95:1 i:180:1.2 b:95:1 u:180:.9
|
|
const int Rate = 48000, Frames = 512;
|
|
if (args.Length < 4) { Console.Error.WriteLine("usage: WordRender <outDir> <label> <freq-hz> <ipa>:<ms>:<intensity> ..."); return 2; }
|
|
var outDir = Path.GetFullPath(args[0]);
|
|
var label = args[1];
|
|
var freq = float.Parse(args[2], System.Globalization.CultureInfo.InvariantCulture);
|
|
var phones = args.Skip(3).Select((spec, i) =>
|
|
{
|
|
var p = spec.Split(':');
|
|
var ipa = p[0];
|
|
var ms = float.Parse(p[1], System.Globalization.CultureInfo.InvariantCulture);
|
|
var intensity = float.Parse(p[2], System.Globalization.CultureInfo.InvariantCulture);
|
|
return (new IpaGestureExperimentTarget($"{label}_{i}_{ipa}", ipa, "", ms / 1000f), intensity);
|
|
}).ToArray();
|
|
Directory.CreateDirectory(outDir);
|
|
|
|
var script = IpaGestureExperiment.BuildWordScript(phones, freq);
|
|
File.WriteAllText(Path.Combine(outDir, "script.aqua"), script + "\n");
|
|
var seconds = phones.Sum(p => p.Item1.DurationSeconds) + 0.12f;
|
|
var home = AquaSynthFaustToolchain.Find();
|
|
Console.WriteLine($"faust home: {home ?? "(not found)"}");
|
|
using var compiler = new AquaSynthPatchCompiler(new AquaSynthNativeOptions(SampleRate: Rate));
|
|
// The exporter bakes splines in as offsets summed per surface; the engine's ControlSplineTimeline merges them per surface
|
|
// into one absolute curve. Back-to-back gestures only sequence under the timeline, so the DSP is emitted without splines
|
|
// (surface sliders only) and the timeline drives the sliders block by block, as NativeFaustRuntime.ControlValuesAt does.
|
|
var scripted = PatchScript.Parse(script);
|
|
var timeline = ControlSurfaceCatalog.FromPatch(scripted).CreateTimeline();
|
|
var source = FaustEmitter.Emit(scripted with { ControlSplines = [] }, new FaustExportOptions(label)).Source;
|
|
File.WriteAllText(Path.Combine(outDir, "source.dsp"), source);
|
|
if (!compiler.TryCompileSource(new AquaSynthCompileIdentity(label, label, ""), source, seconds, out var patch0, out var cerr))
|
|
{
|
|
Console.Error.WriteLine("COMPILE FAILED: " + cerr);
|
|
return 1;
|
|
}
|
|
|
|
using var patch = patch0!;
|
|
using var stream = patch.CreateStreamingPatch();
|
|
var buf = GC.AllocateArray<float>(Frames, pinned: true);
|
|
var ptrs = GC.AllocateArray<IntPtr>(1, pinned: true);
|
|
unsafe { ptrs[0] = (IntPtr)Unsafe.AsPointer(ref buf[0]); }
|
|
var outP = Marshal.UnsafeAddrOfPinnedArrayElement(ptrs, 0);
|
|
var blocks = (int)Math.Ceiling(seconds * Rate / Frames);
|
|
var pcm = new short[blocks * Frames];
|
|
double sum = 0; float peak = 0; var nan = false;
|
|
for (var blk = 0; blk < blocks; blk++)
|
|
{
|
|
var ctl = patch.ControlValuesAt(timeline, blk * Frames / (float)Rate);
|
|
stream.ProcessBlock(Frames, IntPtr.Zero, outP, ctl);
|
|
for (var i = 0; i < Frames; i++)
|
|
{
|
|
var v = buf[i];
|
|
nan |= float.IsNaN(v) || float.IsInfinity(v);
|
|
peak = Math.Max(peak, Math.Abs(v));
|
|
sum += (double)v * v;
|
|
pcm[blk * Frames + i] = (short)(Math.Clamp(v, -1f, 1f) * 32767);
|
|
}
|
|
}
|
|
|
|
var wav = Path.Combine(outDir, $"{label}.wav");
|
|
using (var w = new BinaryWriter(File.Create(wav)))
|
|
{
|
|
var db = pcm.Length * 2;
|
|
w.Write("RIFF"u8); w.Write(36 + db); w.Write("WAVEfmt "u8); w.Write(16); w.Write((short)1); w.Write((short)1);
|
|
w.Write(Rate); w.Write(Rate * 2); w.Write((short)2); w.Write((short)16); w.Write("data"u8); w.Write(db);
|
|
foreach (var s in pcm) w.Write(s);
|
|
}
|
|
|
|
var rms = Math.Sqrt(sum / pcm.Length);
|
|
File.WriteAllText(Path.Combine(outDir, "manifest.json"), JsonSerializer.Serialize(new
|
|
{
|
|
label, freq, phones = args.Skip(3), seconds, sampleRate = Rate, frames = pcm.Length, peak, rms, nonFinite = nan,
|
|
wavSha256 = Convert.ToHexString(SHA256.HashData(File.ReadAllBytes(wav))).ToLowerInvariant(),
|
|
scriptSha256 = Convert.ToHexString(SHA256.HashData(File.ReadAllBytes(Path.Combine(outDir, "script.aqua")))).ToLowerInvariant(),
|
|
os = RuntimeInformation.OSDescription, faustHome = home
|
|
}, new JsonSerializerOptions { WriteIndented = true }));
|
|
Console.WriteLine($"{wav} {seconds:F2}s peak={peak:F3} rms={rms:F4} nonFinite={nan}");
|
|
return nan ? 1 : 0;
|