Audio AI GUIDE
CREPE Pitch Estimation
CREPE is a convolutional neural network for estimating fundamental frequency in monophonic audio.
Listen to this guideAI-narrated · ~2 min
On this page2 min read
Overview
It operates directly on waveform frames and was evaluated against established pitch trackers, including tests of noise robustness. Its pitch estimates are model outputs and can still be wrong, especially when the input falls outside the task or conditions studied.
Deep Dive
CREPE (Convolutional Representation for Pitch Estimation), introduced by Kim, Salamon, Li, and Bello in 2018, estimates the fundamental frequency of monophonic sound such as a solo voice or instrument. The model applies a deep convolutional network directly to time-domain waveform frames instead of first computing a spectrogram. It frames pitch estimation as classification over 360 bins spanning roughly six octaves, spaced 20 cents apart; a local weighted average around the peak can refine the frequency estimate. The authors compared it with pYIN and evaluated generalization under noise, reporting performance that was equal or better in their tested settings. Those findings do not imply that CREPE is more accurate for every recording or task.
Real-World Implementation
Tracking a singer's pitch for real-time tuning feedback in vocal-training apps
Driving auto-tune and pitch-correction tools with accurate fundamental-frequency curves
Transcribing solo instrument melodies into MIDI or sheet music
Analyzing intonation and vibrato in music education and performance research
The Future of CREPE Pitch Estimation
Pitch tracking research explores polyphonic inputs, model latency, and compact inference on devices. CREPE’s waveform-first design remains a useful comparison point, while newer systems may combine learned representations with explicit signal-processing methods. Downstream tools should treat pitch and confidence outputs as estimates that may need smoothing or review.
Putting it into practice
Why it matters
- Access and reach. It improves accessibility through transcription, narration, and voice interfaces.
- Cost and budget. Media teams can ship polished audio faster with smaller budgets.
- Speed and scale. Customer-facing systems can process spoken interactions at larger scale.
Watch out for
- Voice misuse and impersonation risks increase when consent is missing.
- Accuracy can drop across accents, dialects, or noisy environments.
- Synthetic audio can be mistaken for authentic speech without clear labeling.
Next steps
- Obtain explicit consent for voice capture, cloning, and reuse.
- Test quality across diverse speakers and background conditions.
- Define when a human must review or approve outputs.
- Label synthetic audio and keep provenance records for accountability.
Keep Exploring
Free newsletter
Get the daily AI briefing
Three verified AI stories every weekday morning, written in plain English. Free forever, no ads.
One email each weekday. Unsubscribe in one click. We never sell or share your address.
Test yourself
Take the CREPE Pitch Estimation quiz
Instant feedback on every answer, and a shareable certificate with a verifiable ID once you pass a course.
Support free AI education. AI Understanding is a 501(c)(3) nonprofit — no ads, no paywall, ever. Make a donation
Frequently asked questions
What is CREPE Pitch Estimation?
CREPE is a convolutional neural network for estimating fundamental frequency in monophonic audio. It operates directly on waveform frames and was evaluated against established pitch trackers, including tests of noise robustness. Its pitch estimates are model outputs and can still be wrong, especially when the input falls outside the task or conditions studied.
What does CREPE estimate from an audio signal?
CREPE predicts the fundamental frequency, i.e. the perceived pitch, of monophonic audio.
CREPE is designed primarily for what kind of audio?
CREPE estimates pitch for monophonic signals such as a single voice or solo instrument.
How does CREPE frame the pitch-estimation task internally?
CREPE outputs a probability distribution over 360 pitch bins, treating estimation as classification.
What does CREPE take as its direct input?
Unlike spectrogram-based methods, CREPE operates on raw waveform frames through convolutional layers.
Approximately how far apart are CREPE's pitch bins?
CREPE uses 360 bins spaced 20 cents apart, giving sub-semitone resolution across about six octaves.
Keep learning
Related guides
More guides picked for this topic