Processing audio signals

1. A data processing system for processing received audio data comprising:

a classification unit configured to receive data derived from an audio signal and, based on the received data, to determine a classification state of an acoustic environment;

wherein access to a subsequent processing unit is controlled based on the classification state of the acoustic environment;

wherein the data processing system further comprises a control unit to determine a control state of the data processing system from a plurality of states including a first control state and a second control state, wherein the first control state is a default state; and

wherein, in the first control state, the control unit is configured to generate a positive streaming indication in order to allow data access to the subsequent processing unit, if an indication that data representing a presence of a trigger word is present in a received audio signal is received from a trigger word model.

2. A data processing system as claimed in claim 1 , wherein:

the classification unit comprises:

a model configured to receive the data and to derive a representation of the acoustic environment;

a comparator operable to compare the representation with a threshold and to output the classification state of the acoustic environment.

3. A data processing system as claimed in claim 1 , further comprising a trigger word model for detecting a presence of data corresponding to the trigger word in the audio signal.

4. A data processing system as claimed in claim 3 , further comprising a control unit configured to receive i) an indication from the trigger word model and ii) the classification state of the classification unit.

5. A data processing system as claimed in claim 1 , wherein in the second control state, the control unit is configured to generate a positive streaming indication in order to allow data access to the subsequent processing unit without requiring a presence of data representing the trigger word.

6. A data processing system as claimed in claim 1 , wherein the control unit is configured to transition from the first control state to the second control state if:

i) an indication that data representing the presence of a trigger word is present in the received audio signal is received from the trigger word model; and

ii) a predetermined classification state is received from the classification unit.

7. A data processing system as claimed in claim 6 , wherein the predetermined classification state is a close-talk state.

8. A data processing system as claimed in claim 1 , wherein the control unit is configured to transition from the second control state to the first control state if:

i) a classification state other than a predetermined classification state is received from the classification unit; or

ii) if a timer is below a predetermined value.

9. A data processing system as claimed in claim 3 , further comprising a timer, wherein the timer is initiated if the presence of data corresponding to the trigger word in the audio signal is determined.

10. A device comprising a data processing system according to claim 1 , wherein the device comprises a mobile telephone, headphone, acoustic noise cancelling headphones, a smart watch, an audio player, a video player, a mobile computing platform, a games device, a remote controller device, a toy, a machine, or a home automation controller, a domestic appliance or other portable device.

11. A method of processing audio data comprising:

determining a classification state of an acoustic environment;

controlling streaming of the audio data to a subsequent processing unit based on the classification state; and

determining a control state from a plurality of control states including a first control state and a second control state, wherein the first control state is a default state;

wherein, in the first control state, a control unit is configured to generate a positive streaming indication in order to allow data access to the subsequent processing unit, if an indication that data representing a presence of a trigger word is present in a received audio signal is received from the trigger word model.

12. A method as claimed in claim 11 , wherein the determination of the classification state comprises:

deriving a raw representation of the acoustic environment based on a model; and

comparing the raw representation with a predetermined value in order to determine the classification state.

13. A method as claimed in claim 12 , further comprising:

processing the received audio signal, or features extracted therefrom, to detect a presence of data corresponding to a trigger word.

14. A data processing system as claimed in claim 1 , wherein the determination of the classification state is based on a representation of a ratio of a direct to reverberant energy of the audio signal.

15. A method as claimed in claim 11 , wherein the determination of the classification state is based on a representation of a ratio of a direct to reverberant energy of the received audio signal.