Technology

Processing a standardized data collection with deep learning and music information retrieval

In three years of research, we have established a unique data collection method at the interface of music, emotion and situation. This allows us not only to guarantee high quality of results but also to adapt our technology to the needs of each client.

Technology

Processing a standardized data collection with deep learning and music information retrieval

In three years of research, we have established a unique data collection method at the interface of music, emotion and situation. This allows us not only to guarantee high quality of results but also to adapt our technology to the needs of each client.

Quantify the emotional effect of music for safe decision-making

Quantify the emotional effect of music for safe decision-making

Save time, money & risk

Cyanite processes music 65x faster than real-time and makes your music selection process more efficient. Avoid wrong decisions and long discussion loops with quantified analysis results. Our servers are located in Germany and GDPR compliant.

Thrive through clarity

Objective AI evaluation compares songs on a neutral level and creates clear categories as well as practical tags for your music titles. Thanks to an intuitive interface, you don’t need to be an AI wizard for this.

Benefit from a unique dataset

The core of our AI are real, emotional and situational music experiences of people. This unique dataset comes from our own app for Music Moments and grows continuously.

Save time, money & risk

Cyanite processes music 65x faster than real-time and makes your music selection process more efficient. Avoid wrong decisions and long discussion loops with quantified analysis results. Our servers are located in Germany and GDPR compliant.

Thrive through clarity

Objective AI evaluation compares songs on a neutral level and creates clear categories as well as practical tags for your music titles. Thanks to an intuitive interface, you don’t need to be an AI wizard for this.

Benefit from a unique dataset

The core of our AI are real, emotional and situational music experiences of people. This unique dataset comes from our own app for Music Moments and grows continuously.

How the technology works

How the technology works

Learning characteristic sound features

In a continuous training process, our AI recognizes and remembers certain sound features within pieces of music. They are patterns and structures that characterize the sound of a musical piece. Using the training data, it links these characteristics to attributes that stand for emotions or situations.

Learning characteristic sound features

In a continuous training process, our AI recognizes and remembers certain sound features within pieces of music. They are patterns and structures that characterize the sound of a musical piece. Using the training data, it links these characteristics to attributes that stand for emotions or situations.

Sound characteristics within audio source

The AI examines to what extent it finds the patterns and structures learned from the training data in your music file.

Sound characteristics within audio source

The AI examines to what extent it finds the patterns and structures learned from the training data in your music file.

Mood & Genre Scores

If the AI recognizes a significant number of patterns and structures in your song which are related to a certain attribute, it will make an accurate assessment of how likely it is that your song has this attribute. This attribute can be the emotion it evokes or a genre it classifies for.

Mood & Genre Scores

If the AI recognizes a significant number of patterns and structures in your song which are related to a certain attribute, it will make an accurate assessment of how likely it is that your song has this attribute. This attribute can be the emotion it evokes or a genre it classifies for.

Visualization & recommendation

The analysis results, accurate to eight decimals, are presented in your project in an intuitive and easy-to-interpret form. For music recommendations, songs with similar scores are suggested.

Visualization & recommendation

The analysis results, accurate to eight decimals, are presented in your project in an intuitive and easy-to-interpret form. For music recommendations, songs with similar scores are suggested.