Precisione di Whisper: come migliorare la qualità della trascrizione

• Sviluppatori che ottimizzano la precisione di trascrizione con Whisper
• Creator che trascrivono podcast e video
• Ricercatori che lavorano con audio
• Chi cerca suggerimenti sulla precisione di Whisper

Cosa influenza la precisione di Whisper

• Qualità audio (la più importante)
• Dimensione del modello (scelta)
• Rilevamento della lingua
• Pre-elaborazione audio
• Parametri di configurazione
• Durata audio e segmentazione

Suggerimento 1: Scegliere la dimensione del modello giusta

import whisper

# Model sizes from fastest to most accurate:
# tiny, base, small, medium, large

# For maximum accuracy, use medium or large
model = whisper.load_model("medium")  # Best balance
# or
model = whisper.load_model("large")  # Maximum accuracy

import whisper

def transcribe_with_optimal_model(audio_path, prioritize_accuracy=True):
    """
    Select model based on accuracy vs speed priority.
    
    Args:
        audio_path: Path to audio file
        prioritize_accuracy: True for accuracy, False for speed
    """
    if prioritize_accuracy:
        model_size = "medium"  # or "large" for best accuracy
    else:
        model_size = "base"  # or "small" for balanced
    
    model = whisper.load_model(model_size)
    result = model.transcribe(audio_path)
    
    return result

# For critical transcriptions
result = transcribe_with_optimal_model("important_meeting.mp3", prioritize_accuracy=True)

Suggerimento 2: Specificare la lingua quando è nota

import whisper

model = whisper.load_model("base")

# Auto-detect (less accurate)
result_auto = model.transcribe("audio.mp3")

# Specify language (more accurate)
result_en = model.transcribe("audio.mp3", language="en")
result_zh = model.transcribe("audio.mp3", language="zh")
result_es = model.transcribe("audio.mp3", language="es")

• Meno errori di rilevamento lingua
• Risultati migliori per parlanti multilingue
• Elaborazione più veloce (salta il rilevamento)
• Migliore gestione di accenti e dialetti

import whisper
import langdetect

def transcribe_with_language_detection(audio_path, model_size="base"):
    """
    Detect language first, then transcribe with explicit language.
    """
    model = whisper.load_model(model_size)
    
    # Quick language detection
    result_quick = model.transcribe(audio_path, language=None)
    detected_lang = result_quick["language"]
    
    # Re-transcribe with detected language for better accuracy
    result = model.transcribe(audio_path, language=detected_lang)
    
    return result

result = transcribe_with_language_detection("audio.mp3")

Suggerimento 3: Pre-elaborare l’audio prima della trascrizione

import whisper
import numpy as np
from scipy.io import wavfile
from scipy import signal

def preprocess_audio(audio_path, output_path):
    """
    Preprocess audio to improve transcription accuracy.
    """
    # Read audio file
    sample_rate, audio = wavfile.read(audio_path)
    
    # Normalize audio (scale to [-1, 1])
    if audio.dtype == np.int16:
        audio = audio.astype(np.float32) / 32768.0
    elif audio.dtype == np.int32:
        audio = audio.astype(np.float32) / 2147483648.0
    
    # Remove DC offset
    audio = audio - np.mean(audio)
    
    # Normalize volume
    max_val = np.max(np.abs(audio))
    if max_val > 0:
        audio = audio / max_val * 0.95  # Leave headroom
    
    # Resample to 16kHz (Whisper's optimal sample rate)
    if sample_rate != 16000:
        num_samples = int(len(audio) * 16000 / sample_rate)
        audio = signal.resample(audio, num_samples)
        sample_rate = 16000
    
    # Save preprocessed audio
    wavfile.write(output_path, sample_rate, (audio * 32767).astype(np.int16))
    
    return output_path

# Usage
preprocessed = preprocess_audio("raw_audio.wav", "preprocessed.wav")
model = whisper.load_model("base")
result = model.transcribe(preprocessed)

1. Normalizzare i livelli – volume uniforme
2. Rimuovere l’offset DC – eliminare bias costante
3. Ricampionare a 16 kHz – frequenza ottimale per Whisper
4. Rimuovere il silenzio – concentrarsi sulla voce
5. Ridurre il rumore – pulire lo sfondo

Suggerimento 4: Impostazioni di temperatura per risultati migliori

import whisper

model = whisper.load_model("base")

# Default temperature (0.0)
result_default = model.transcribe("audio.mp3")

# Lower temperature for more deterministic results
result_low_temp = model.transcribe(
    "audio.mp3",
    temperature=0.0,  # Most deterministic
    best_of=5,  # Try multiple decodings, pick best
    beam_size=5  # Beam search size
)

• temperature=0.0: più deterministico, migliore precisione
• temperature=0.2: poca casualità, buon equilibrio
• temperature=0.6: predefinito, bilanciato
• Valori alti: più “creativo”, meno preciso

def transcribe_with_optimal_settings(audio_path, model_size="base"):
    """
    Use optimal settings for maximum accuracy.
    """
    model = whisper.load_model(model_size)
    
    result = model.transcribe(
        audio_path,
        temperature=0.0,  # Most deterministic
        best_of=5,  # Try 5 decodings, pick best
        beam_size=5,  # Beam search
        patience=1.0,  # Patience for beam search
        condition_on_previous_text=True,  # Use context
        initial_prompt="This is a conversation about technology."  # Context hint
    )
    
    return result

Suggerimento 5: Prompt iniziale per il contesto

import whisper

model = whisper.load_model("base")

# Without context
result_basic = model.transcribe("meeting.mp3")

# With context (much better accuracy)
result_context = model.transcribe(
    "meeting.mp3",
    initial_prompt="This is a business meeting discussing project timelines and deliverables."
)

# For technical content
result_tech = model.transcribe(
    "lecture.mp3",
    initial_prompt="This is a computer science lecture about machine learning and neural networks."
)

• Contenuto tecnico: includere termini di dominio
• Nomi e luoghi: menzionare nomi propri importanti
• Accenti: descrivere accento o dialetto
• Contesto: descrivere ambiente o argomento

def transcribe_with_context(audio_path, context_description):
    """
    Transcribe with context for better accuracy.
    """
    model = whisper.load_model("medium")
    
    result = model.transcribe(
        audio_path,
        initial_prompt=context_description,
        language="en"
    )
    
    return result

# Example usage
result = transcribe_with_context(
    "interview.mp3",
    "This is an interview with Dr. Sarah Johnson about medical research. "
    "The conversation includes technical medical terminology."
)

Suggerimento 6: Gestire correttamente file audio lunghi

import whisper
from pydub import AudioSegment
import os

def transcribe_long_audio(audio_path, model_size="base", chunk_length_minutes=30):
    """
    Transcribe long audio by splitting into optimal chunks.
    """
    model = whisper.load_model(model_size)
    
    # Load audio
    audio = AudioSegment.from_file(audio_path)
    chunk_length_ms = chunk_length_minutes * 60 * 1000
    
    # Split into chunks
    chunks = []
    for i in range(0, len(audio), chunk_length_ms):
        chunks.append(audio[i:i + chunk_length_ms])
    
    # Transcribe each chunk
    full_text = []
    all_segments = []
    
    for i, chunk in enumerate(chunks):
        chunk_path = f"temp_chunk_{i}.wav"
        chunk.export(chunk_path, format="wav")
        
        print(f"Transcribing chunk {i+1}/{len(chunks)}")
        result = model.transcribe(chunk_path)
        
        # Adjust timestamps for chunk offset
        offset = i * chunk_length_ms / 1000.0
        for segment in result["segments"]:
            segment["start"] += offset
            segment["end"] += offset
            all_segments.append(segment)
        
        full_text.append(result["text"])
        
        # Clean up
        os.remove(chunk_path)
    
    # Combine results
    combined_result = {
        "text": " ".join(full_text),
        "segments": all_segments,
        "language": result["language"]
    }
    
    return combined_result

# Usage
result = transcribe_long_audio("long_podcast.mp3", model_size="medium", chunk_length_minutes=30)

• Dividere in segmenti da 20–30 minuti
• Usare la stessa dimensione di modello per tutti i segmenti
• Mantenere contesto tra i segmenti
• Unire i segmenti con timestamp corretti

Suggerimento 7: Ottimizzare per audio rumoroso

import whisper
import noisereduce as nr
import soundfile as sf
import numpy as np

def transcribe_noisy_audio(audio_path, model_size="medium"):
    """
    Reduce noise before transcription for better accuracy.
    """
    # Load audio
    audio, sample_rate = sf.read(audio_path)
    
    # Reduce noise
    reduced_noise = nr.reduce_noise(
        y=audio,
        sr=sample_rate,
        stationary=False,  # For non-stationary noise
        prop_decrease=0.8  # Reduce noise by 80%
    )
    
    # Save cleaned audio
    cleaned_path = "cleaned_audio.wav"
    sf.write(cleaned_path, reduced_noise, sample_rate)
    
    # Transcribe with larger model (better for noisy audio)
    model = whisper.load_model(model_size)
    result = model.transcribe(cleaned_path)
    
    # Clean up
    os.remove(cleaned_path)
    
    return result

# Usage
result = transcribe_noisy_audio("noisy_recording.mp3", model_size="medium")

• Usare modelli medium o large
• Pre-elaborare con riduzione rumore
• Aumentare best_of
• Fornire contesto sulle condizioni di rumore

Suggerimento 8: Timestamp a livello di parola per più controllo

import whisper

model = whisper.load_model("base")

# Get word timestamps
result = model.transcribe(
    "audio.mp3",
    word_timestamps=True  # Enable word-level timestamps
)

# Access word timestamps
for segment in result["segments"]:
    print(f"Segment: {segment['text']}")
    print(f"Start: {segment['start']:.2f}s, End: {segment['end']:.2f}s")
    
    if "words" in segment:
        for word in segment["words"]:
            print(f"  Word: {word['word']} ({word['start']:.2f}s - {word['end']:.2f}s)")

• Sottotitoli: sincronizzazione precisa parola per parola
• Correzione errori: individuare parole problematiche
• Ricerca: trovare parole nella trascrizione
• Analisi del parlato: pattern di voce

Suggerimento 9: Combinare più decodifiche

import whisper

model = whisper.load_model("base")

# Single decoding (default)
result_single = model.transcribe("audio.mp3")

# Multiple decodings, pick best (more accurate)
result_best = model.transcribe(
    "audio.mp3",
    best_of=5,  # Try 5 decodings
    temperature=(0.0, 0.2, 0.4, 0.6, 0.8)  # Different temperatures
)

• Precisione: più alta con più decodifiche
• Velocità: più lenta (5× con best_of=5)
• Quando: precisione critica, velocità secondaria

Suggerimento 10: Post-elaborare le trascrizioni

import re
import whisper

def post_process_transcript(text):
    """
    Fix common transcription errors.
    """
    # Fix common contractions
    text = re.sub(r"\b(\w+) '(\w+)\b", r"\1'\2", text)  # Fix spacing in contractions
    
    # Fix common homophones (add your own)
    replacements = {
        "there": "their",  # Context-dependent
        "its": "it's",  # Context-dependent
        # Add more based on your domain
    }
    
    # Capitalize sentences
    sentences = re.split(r'([.!?]\s+)', text)
    capitalized = []
    for i, sentence in enumerate(sentences):
        if sentence.strip():
            capitalized.append(sentence[0].upper() + sentence[1:] if len(sentence) > 1 else sentence.upper())
        else:
            capitalized.append(sentence)
    
    return "".join(capitalized)

# Usage
model = whisper.load_model("base")
result = model.transcribe("audio.mp3")
processed_text = post_process_transcript(result["text"])

Esempio completo: ottimizzazione della precisione pronta per la produzione

import whisper
import os
from pathlib import Path

def transcribe_with_maximum_accuracy(
    audio_path,
    model_size="medium",
    language=None,
    context_prompt=None,
    output_format="txt"
):
    """
    Transcribe audio with maximum accuracy using best practices.
    
    Args:
        audio_path: Path to audio file
        model_size: Whisper model size (medium or large recommended)
        language: Language code (None for auto-detect)
        context_prompt: Initial prompt for context
        output_format: Output format (txt, json, srt)
    """
    # Load model (medium or large for best accuracy)
    print(f"Loading Whisper model: {model_size}")
    model = whisper.load_model(model_size)
    
    # Prepare transcription parameters
    transcribe_kwargs = {
        "temperature": 0.0,  # Most deterministic
        "best_of": 5,  # Try multiple decodings
        "beam_size": 5,  # Beam search
        "patience": 1.0,
        "condition_on_previous_text": True,
        "word_timestamps": True,  # Get word-level timestamps
    }
    
    # Add language if specified
    if language:
        transcribe_kwargs["language"] = language
    
    # Add context prompt if provided
    if context_prompt:
        transcribe_kwargs["initial_prompt"] = context_prompt
    
    # Transcribe
    print(f"Transcribing: {audio_path}")
    result = model.transcribe(audio_path, **transcribe_kwargs)
    
    # Post-process
    result["text"] = post_process_transcript(result["text"])
    
    # Save result
    base_name = Path(audio_path).stem
    output_path = f"{base_name}_transcript.{output_format}"
    
    if output_format == "txt":
        with open(output_path, "w", encoding="utf-8") as f:
            f.write(result["text"])
    elif output_format == "json":
        import json
        with open(output_path, "w", encoding="utf-8") as f:
            json.dump(result, f, indent=2, ensure_ascii=False)
    
    print(f"✓ Transcription saved: {output_path}")
    print(f"  Language: {result['language']}")
    print(f"  Duration: {result['segments'][-1]['end']:.2f}s")
    
    return result

# Example usage
result = transcribe_with_maximum_accuracy(
    audio_path="important_meeting.mp3",
    model_size="medium",
    language="en",
    context_prompt="This is a business meeting discussing quarterly results and project updates.",
    output_format="txt"
)

Confronto di precisione: prima e dopo l’ottimizzazione

Riepilogo delle best practice

Per la massima precisione:

1. ✅ Modello medium o large
2. ✅ Lingua esplicita
3. ✅ Contesto con initial_prompt
4. ✅ temperature=0.0 per risultati deterministici
5. ✅ word_timestamps per output dettagliato
6. ✅ Pre-elaborare audio rumoroso
7. ✅ Dividere file lunghi
8. ✅ best_of=5 per contenuti critici

Per equilibrio velocità/precisione:

1. ✅ Modello small o base
2. ✅ Rilevamento automatico della lingua
3. ✅ Temperatura predefinita
4. ✅ Senza best_of
5. ✅ File elaborati così come sono (pre-elaborazione minima)

Errori comuni da evitare

❌ Modello tiny per contenuti importanti

❌ Non specificare la lingua

❌ Ignorare il contesto

❌ Impostazioni predefinite con molto rumore

❌ File molto lunghi in un unico blocco

Risoluzione dei problemi di precisione

Problema: Bassa precisione su termini tecnici

result = model.transcribe(
    "technical_audio.mp3",
    initial_prompt="This audio contains technical terminology related to machine learning, neural networks, and deep learning."
)

Problema: Scarsa precisione con accenti

# Use larger model
model = whisper.load_model("medium")

# Provide accent context
result = model.transcribe(
    "accented_audio.mp3",
    initial_prompt="This speaker has a British accent.",
    language="en"
)

Problema: Errori su nomi propri

# Include names in initial prompt
result = model.transcribe(
    "interview.mp3",
    initial_prompt="This interview features Dr. Sarah Johnson and Professor Michael Chen discussing research."
)

Conclusione

• Modello: medium o large per contenuti critici
• Configurazione: temperatura e decodifica ottimali
• Contesto: informazioni di dominio
• Pre-elaborazione: audio pulito prima della trascrizione
• Post-elaborazione: correzione automatica degli errori comuni

1. La dimensione del modello ha il maggior impatto
2. Specificare la lingua migliora molto i risultati
3. I prompt di contesto aiutano su contenuti specializzati
4. Più decodifiche (best_of) aumentano la precisione ma rallentano
5. La qualità audio resta il fattore principale