Adaptive loudness levelling for digital audio signals
View Patent ↗A time-domain method of adaptively levelling the loudness of a digital audio signal is proposed. It selects a proper frequency weighting curve to relate the volume level to the human auditory system. The audio signal is segmented into frames of a suitable duration for content analysis. Each frame is classified to one of several predefined states and events of perceptual interest is detected. Four quantities are updated each frame according to the classified state and detected event to keep track of the signal. One quantity measures the long-term loudness and is the main criterion for state classification of a frame. The second quantity is the short-term loudness that is mainly used for deriving the target gain. The third quantity measures the low-level loudness when the signal is deemed to not contain important content, giving a reasonable estimate of noise floor. A fourth quantity measures the peak loudness level that is used to simulate the temporal masking effect. The target gain to maintain the audio signal to the desired loudness level is calculated by a volume leveller, regulated by a gain controller that simulates the temporal masking effect to get rid of unnecessary gain fluctuations, ensuring a pleasant sound.
1. A method for adaptive loudness levelling, comprising:
receiving a plurality of audio input signals;
weighting one or more frequencies associated with each audio input signal according to human audio perception;
characterizing a state for each audio input signal;
detecting whether each audio input signal is experiencing an event;
producing a target levelling gain for each audio input signal based on the characterized state and/or detected event;
temporally masking the target levelling gain; and
combining the temporally masked target levelling gain with a delayed plurality of audio input signals.
2. The method of claim 1 wherein characterizing includes classifying each frame of the plurality of audio input signals according to a plurality of signal attributes.
3. The method of claim 2 wherein one of the plurality of signal attributes is an estimated long term loudness measure and wherein the estimated long term loudness measure is updated at a predetermined update rate.
4. The method of claim 3 wherein responsive to detecting that the plurality of audio input signals is experiencing an event, updating the plurality of signal attributes at an increased update rate greater than the predetermined update rate.
5. The method of claim 1 wherein temporally masking includes smoothing a peak signal level for each frame of the plurality of audio input signals.
6. The method of claim 1 wherein temporally masking includes smoothing an average loudness over a predetermined number of frames of the plurality of audio input signals.
7. The method of claim 1 wherein temporally masking includes smoothing each frame of the plurality of audio input signals based on content characterization.
8. The method of claim 1 further comprising smoothing the target levelling gain for each frame of the plurality of audio input signals proportionally to a difference between a current signal level and a prior peak signal level.
9. A method for modifying digital audio signal amplitude to maintain consistent loudness, comprising:
relating a volume level of a plurality of audio input signals to human perception using a weighting curve;
computing a volume levelling target gain by tracking a plurality of audio signal attributes for each of the audio input signals wherein tracking includes classifying a state of each of the audio input signals and detecting whether each of the audio input signals is associated with an event;
applying the volume levelling target gain to a delayed audio input signal producing a consistent loudness level; and
temporally masking the target levelling gain and combining the temporally masked target levelling gain with a delayed plurality of audio input signals.
10. The method of claim 9 further comprising segmenting each of the audio input signals into frames.
11. The method of claim 9 wherein the plurality of audio signal attributes includes a long-term loudness of the signal, a short-term loudness, a low-level loudness, and a peak loudness.
12. The method of claim 9 wherein a higher priority level is associated with a higher loudness estimate.
13. The method of claim 9 wherein the state of each of the audio input signals is primarily based on a long-term loudness attribute of the signal.
14. The method of claim 9 wherein detecting includes analyzing short-term loudness of each of the audio input signals.
15. The method of claim 9 wherein the volume levelling target gain is based on a short-term loudness level.
16. The method of claim 9 further comprising simulating temporal masking.
17. The method of claim 9 wherein the weighting curve is derived according to equal-loudness contour curves, the weighting curve being selected based on expected sound pressure levels.
18. The method of claim 9 wherein a short-term loudness level is updated based on the state and detected event and is used for computing the volume levelling target gain.
19. The method of claim 9 wherein the plurality of audio signal attributes are discriminatively updated based on event detection.
20. The method of claim 9 further comprising simulating a temporal masking effect of human hearing by tracking a peak loudness, an average loudness and the state of each of the plurality of audio input signals.