My voice check analyzes your voice based on acoustic principles. Here's a detailed step-by-step explanation:
1. Digital Conversion of Sound
Your voice is originally an analog signal (continuous air vibrations). It needs to be converted into a digital signal for computer processing.
- Sampling: Continuous sound is measured tens of thousands of times per second and recorded as numbers. Typically 44,100Hz (44,100 measurements per second) or 48,000Hz is used.
- Quantization: Measured values are rounded to fixed units and stored as digital numbers, expressed in 16-bit or 24-bit units.
Like drawing a smooth curve by plotting very dense points and connecting them. More points mean closer to the original sound.
2. Frequency Analysis (FFT)
The digitally converted sound is analyzed using Fast Fourier Transform (FFT).
- Time Domain → Frequency Domain: Converts time-varying sound into "which frequencies are included and how much."
- Fundamental Frequency (F0) Extraction: Finds the lowest frequency that determines voice pitch. Typically 85-180Hz for males, 165-255Hz for females.
- Harmonics Analysis: Analyzes harmonics appearing at integer multiples of the fundamental frequency to identify timbre characteristics.
Like listening to a chord and decomposing "which notes are mixed."
3. Voice Feature Extraction
Various characteristics of the voice are quantified from frequency analysis results:
- Pitch: Measures voice pitch based on fundamental frequency.
- Spectral Centroid: Center of mass of frequency distribution, indicating whether the voice is bright or dark.
- Energy: Measures sound volume and intensity.
- Zero Crossing Rate: Frequency of signal crossing zero, identifying roughness of sound.
- MFCC (Mel-Frequency Cepstral Coefficients): Timbre feature values considering human auditory characteristics.
4. Result Classification
Extracted features are synthesized to classify into one of 12 types:
- Multidimensional Analysis: Considers multiple features simultaneously to identify overall voice characteristics.
- Weighting: Considers importance of each feature to determine final type.
- Result Presentation: Presents best matching type along with main voice characteristics.
5. Analysis Limitations and Accuracy
Digital voice analysis is a very sophisticated process, but it is not 100% accurate:
- Impact of Recording Quality: If there are silences or gaps during recording, calculation errors may occur and results may vary. For example, when analyzing tone brightness (spectral centroid), including silent sections can make the values inaccurate.
- Environmental Factors: Background noise, microphone distance, speaking speed, and more all affect analysis results.
- Individual Variation: Even for the same person, voice changes depending on time of day, condition, and emotional state, so results may vary.
- Entertainment Purpose: This service is designed for fun and interest rather than scientific rigor.
For the most accurate results, speak continuously for at least 1 second in a consistent tone in a quiet environment.
All processing is done in real-time in your browser, and recorded files or analysis data are not sent to the server.
My voice check is a service for fun and entertainment, not a professional voice analysis tool.
Analysis Characteristics
- Classification is based on phonetic features, but not absolute standards
- Results may vary depending on recording environment and condition, even for the same person
- Microphone quality and ambient noise can affect results
Use the results for reference only, and enjoy sharing them with friends!
Analysis may fail or be inaccurate in the following cases:
Main Failure Causes
- Recording duration is too short (less than 1 second)
- Background noise is too loud
- Distance from microphone is too far or volume is too low
- Microphone volume is too high causing sound distortion
- Microphone permission not granted
Please speak naturally in a quiet environment while maintaining an appropriate distance from the microphone.
No, no data is stored whatsoever.
Privacy Protection
- All voice analysis is processed only in your browser
- Recording files or analysis data are not sent to the server
- All data is deleted when you refresh the page
- We do not collect or store personal information
Use the service with confidence. It operates 100% client-side only.
Not at all! The analysis results do not judge whether your voice is good or bad.
Meaning of Results
- Each voice type simply represents different characteristics
- Every voice has its own unique charm and characteristics
- Analysis measures only objective acoustic characteristics
- Results are for fun and interest, not value judgment
Use it to understand and enjoy your voice characteristics!
Even for the same person, results may vary slightly with each recording.
Result Variation Factors
- It's difficult to speak with exactly the same tone and pitch every time
- Voice changes depending on condition, time of day, and emotional state
- Distance and angle from microphone may differ slightly each time
- Background noise or environment can have an effect
We recommend measuring multiple times and referring to characteristics that appear consistently.
You don't need a special microphone. The built-in microphone of a typical device is sufficient.
Recommendations
- Built-in microphones of laptops and smartphones can be used
- Using headset or earphone microphones is more stable
- Maintain a distance of about 15-30cm from the microphone
- Sound may be distorted if too close
Speaking naturally in a quiet environment is most important.
It can be used in most modern browsers.
Supported Browsers
- Chrome, Edge (Chromium-based) - Full support
- Safari (including iOS) - Supported
- Firefox - Supported
- Mobile browsers (Chrome, Safari) - Supported
May not work on some older browsers. We recommend using the latest version.
Yes, you can easily share your analysis results.
Sharing Methods
- You can copy and share the URL of the result page
- You can share via KakaoTalk
- You can also capture the result screen and share it as an image
- Have fun comparing results with your friends!
Shared URLs contain only analysis data and do not include actual recording files.
Have More Questions?
If you didn't find your answer above, try our service to experience the fun of voice analysis yourself.
Start Now