2b-studio / studio /audio_analysis.py
2Btainment's picture
2B-Studio Phase 6: privater Space-Test (Shell + Worker-Routen)
45d92cc verified
Raw History Blame Contribute Delete
4.52 kB
from __future__ import annotations
import json
import math
import subprocess
import sys
import wave
from array import array
from dataclasses import asdict, dataclass
from pathlib import Path
from typing import Any
@dataclass(frozen=True, slots=True)
class AudioAnalysis:
duration_s: float
sample_rate: int
channels: int
peak_dbfs: float
rms_dbfs: float
crest_db: float
zero_crossing_rate: float
character: str
def as_dict(self) -> dict[str, Any]:
return asdict(self)
def _dbfs(amplitude: float) -> float:
if amplitude <= 0:
return -96.0
return max(-96.0, 20 * math.log10(amplitude / 32767.0))
def _describe_character(rms_dbfs: float, crest_db: float, zero_crossing_rate: float) -> str:
density = "dicht" if rms_dbfs > -12 else "ausgewogen" if rms_dbfs > -20 else "dynamisch"
tone = (
"brillant/perkussiv"
if zero_crossing_rate > 0.14
else "warm/dunkel"
if zero_crossing_rate < 0.055
else "tonal balanciert"
)
transients = "kompakt" if crest_db < 7 else "transientenreich" if crest_db > 13 else "offen"
return f"{density}, {tone}, {transients}"
def _read_audio_as_pcm16(audio_path: Path) -> tuple[int, int, int, bytes]:
try:
with wave.open(str(audio_path), "rb") as wav_file:
channels = wav_file.getnchannels()
sample_width = wav_file.getsampwidth()
sample_rate = wav_file.getframerate()
frame_count = wav_file.getnframes()
frames = wav_file.readframes(frame_count)
if sample_width == 2:
return channels, sample_rate, frame_count, frames
except wave.Error:
pass
probe = subprocess.run(
[
"ffprobe",
"-v",
"error",
"-select_streams",
"a:0",
"-show_entries",
"stream=sample_rate,channels",
"-of",
"json",
str(audio_path),
],
check=False,
capture_output=True,
text=True,
timeout=30,
)
if probe.returncode != 0:
raise ValueError(f"FFprobe konnte die WAV-Datei nicht lesen: {probe.stderr[-500:]}")
payload = json.loads(probe.stdout)
streams = payload.get("streams", [])
if not streams:
raise ValueError("Die WAV-Datei enthält keinen Audiostream.")
channels = int(streams[0]["channels"])
sample_rate = int(streams[0]["sample_rate"])
decoded = subprocess.run(
[
"ffmpeg",
"-v",
"error",
"-nostdin",
"-i",
str(audio_path),
"-f",
"s16le",
"-acodec",
"pcm_s16le",
"-",
],
check=False,
capture_output=True,
timeout=120,
)
if decoded.returncode != 0:
detail = decoded.stderr.decode("utf-8", errors="replace")[-500:]
raise ValueError(f"FFmpeg konnte die WAV-Datei nicht dekodieren: {detail}")
frame_count = len(decoded.stdout) // (2 * channels)
return channels, sample_rate, frame_count, decoded.stdout
def analyze_wav(path: str | Path) -> AudioAnalysis:
audio_path = Path(path)
channels, sample_rate, frame_count, frames = _read_audio_as_pcm16(audio_path)
if channels < 1 or sample_rate < 1 or frame_count < 1:
raise ValueError("Die WAV-Datei enthält keine analysierbaren Audiodaten.")
samples = array("h")
samples.frombytes(frames)
if sys.byteorder != "little":
samples.byteswap()
peak = max(abs(sample) for sample in samples)
square_sum = sum(sample * sample for sample in samples)
rms = math.sqrt(square_sum / len(samples))
mono_samples = samples[::channels]
crossings = sum(
1
for previous, current in zip(mono_samples, mono_samples[1:], strict=False)
if (previous < 0 <= current) or (previous >= 0 > current)
)
zero_crossing_rate = crossings / max(1, len(mono_samples) - 1)
peak_dbfs = _dbfs(float(peak))
rms_dbfs = _dbfs(rms)
crest_db = max(0.0, peak_dbfs - rms_dbfs)
return AudioAnalysis(
duration_s=frame_count / sample_rate,
sample_rate=sample_rate,
channels=channels,
peak_dbfs=round(peak_dbfs, 2),
rms_dbfs=round(rms_dbfs, 2),
crest_db=round(crest_db, 2),
zero_crossing_rate=round(zero_crossing_rate, 4),
character=_describe_character(rms_dbfs, crest_db, zero_crossing_rate),
)