Adaptive noise reduction

A computer program product, encoded on a non-transitory computer-readable medium, operable to cause a data processing apparatus to perform operations comprising:

• receiving digital audio data;
• receiving input specifying a noise threshold, the noise threshold identifying a level at which one or more segments of audio data are considered to be noise, the noise threshold specifying a noisiness factor derived from a plurality of parameters of the audio data including an amplitude value of the audio data and a corresponding duration of the audio data, the noise threshold being applied to each of a plurality of frequency bands of the audio data;
• analyzing a first segment of the digital audio data at a first frequency band to identify noise, such that when audio data in the first segment exceeds the noise threshold, the first segment is identified as including a first noise and the audio data is compressed including:
• determining a first amplitude of the audio data corresponding to the first noise, and
• attenuating audio data of the first frequency band of the first segment according to the first amplitude of the first noise;
• analyzing a second segment of the digital audio data at the first frequency band to identify noise, such that when audio data in the second segment exceeds the noise threshold, the second segment is identified as including a second noise and the audio data is compressed including:
• determining a second amplitude of the audio data corresponding to the second noise, and
• attenuating audio data of the first frequency band of the second segment according to second amplitude of the second noise, where the second amplitude is distinct from the first amplitude such that the compression is adapted to compress the second noise at the second amplitude.

A system comprising:

• a user interface device;
• one or more computers operable to interact with the user interface device to:
• receive digital audio data;
• receive input specifying a noise threshold, the noise threshold identifying a level at which one or more segments of audio data are considered to be noise, the noise threshold specifying a noisiness factor derived from a plurality of parameters of the audio data including an amplitude value of the audio data and a corresponding duration of the audio data, the noise threshold being applied to each of a plurality of frequency bands of the audio data;
• analyze a first segment of the digital audio data at a first frequency band to identify noise, such that when audio data in the first segment exceeds the noise threshold, the first segment is identified as including a first noise and the audio data is compressed including:
• determining a first amplitude of the audio data corresponding to the first noise, and
• attenuating audio data of the first frequency band of the first segment according to the first amplitude of the first noise;
• analyze a second segment of the digital audio data at the first frequency band to identify noise, such that when audio data in the second segment exceeds the noise threshold, the second segment is identified as including a second noise and the audio data is compressed including:
• determining a second amplitude of the audio data corresponding to the second noise, and
• attenuating audio data of the first frequency band of the second segment according to second amplitude of the second noise, where the second amplitude is distinct from the first amplitude such that the compression is adapted to compress the second noise at the second amplitude.