Classifications

• • G—PHYSICS
  • G10—MUSICAL INSTRUMENTS; ACOUSTICS
  • G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
  • G10L21/00—Speech or voice signal processing techniques to produce another audible or non-audible signal, e.g. visual or tactile, in order to modify its quality or its intelligibility
  • G10L21/02—Speech enhancement, e.g. noise reduction or echo cancellation
  • G10L21/0208—Noise filtering
• • G—PHYSICS
  • G10—MUSICAL INSTRUMENTS; ACOUSTICS
  • G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
  • G10L21/00—Speech or voice signal processing techniques to produce another audible or non-audible signal, e.g. visual or tactile, in order to modify its quality or its intelligibility
  • G10L21/02—Speech enhancement, e.g. noise reduction or echo cancellation
  • G10L21/0208—Noise filtering
  • G10L21/0216—Noise filtering characterised by the method used for estimating noise
  • G10L2021/02168—Noise filtering characterised by the method used for estimating noise the estimation exclusively taking place during speech pauses

Definitions

• the invention relates to a communication system for connecting several users in conference mode, that is to say where each user can talk and at the same time hear all other users.
• the invention relates more particularly to a device for improving the intelligibility of speech when users speak in a noisy environment, for example a sporting event in a stadium.
• a well-known solution for suppressing noise during the silent phases is that used by walkie-talkies, that is to say that the user, with a button, switches his terminal between a mode of operation. transmission only and reception mode only. Nevertheless, this solution is unsuitable when the number of users who can speak exceeds three, it is constraining because it monopolizes a hand, and it does not allow a user who speaks to hear an important message that could transmit another user.
• Patent application EP 1843326 describes such a system.
• FIG. 1 represents a block diagram of a terminal as described in patent application EP 1843326.
• a microphone 10 transmits the signal of The signal is then pre-filtered at 14 in order to remove the components out of the speech band and then converted into digital by an analog / digital converter 16. The signal thus converted powers a DSP digital signal processing circuit.
• the DSP is programmed to perform the signal processing envisaged, including the improvement of speech intelligibility.
• An example of treatment is described in the aforementioned patent application. It involves the detection of the speech signature and the calculation of parameters of a filter which makes it possible to subtract the ambient noise of the signal while preserving the speech signal.
• the signal output from the DSP is routed to an antenna 18 via an RF transmission module 20 providing the processing required to pass the digital signal provided by the DSP into a signal transmittable on the antenna according to the standard used by all terminals.
• the antenna 18 also receives the signals transmitted by other terminals, which the RF module 20 converts and transmits to the DSP. These received signals are processed by the DSP and sent to a loudspeaker 22 via a shaping circuit 24 which takes care of the digital / analog conversion, filtering, and amplification.
• a terminal of the type of FIG. 1 is effective in terms of speech intelligibility and noise suppression, provided that the gain of the amplifier 12 is always adapted to the type of microphone, and that the microphone is located at a microphone. precise location relative to the user's mouth. Any deviation can be detrimental to the effectiveness of the terminal.
• a lapel microphone with which the speech measurement conditions will vary over time depending on the movements of the user and the orientation of his head.
• a device for improving the intelligibility of a signal originating from a source subjected to a noisy environment, said source printing a signature specific to the signal comprising a processing circuit receiving the signal, and means for analyzing the signal and setting the processing circuit according to characteristics of the signature present in the signal.
• the device comprises a first low-distortion channel carrying the signal from the source to the analysis means, and a second channel capable of introducing a distortion, conveying the signal from the source to the processing circuit.
• FIG. 1, previously described, represents a block diagram of a conventional terminal that can be used in a conference-type multi-user communication
• FIG. 2 shows a block diagram of a terminal embodiment referred to in this patent application.
• FIG. 3 represents improvements that can be made to the embodiment of FIG. 2.
• variable gain input amplifier in a feedback loop that adjusts the gain accordingly.
• the signal envelope for example, the signal envelope.
• the amplifier gain control loop does not react fast enough to avoid the saturation of the chain. (By the way, the reaction time of the loop is deliberately slow to reduce distortion under nominal usage conditions.)
• variable gain amplifier solution is unsuitable in a terminal of the type of Figure 1 (in the conditions where it is desired to use it) to compensate for variations in the location of the microphone. Indeed, the saturation of the processing chain disturbs the signature detection of the speech so much that it causes too many false detections and consequently inefficient noise filtering.
• a dynamic compressor which is an amplifier with a non-linear gain curve, flattening asymptotically towards the saturation limit.
• Fig. 2 shows elements of a terminal incorporating a microphone location compensation system embodiment. It contains the same elements as in Figure 1, designated by the same references.
• the signal from the microphone 10 is transmitted to the DSP by a first channel incorporating the amplifier 12, filter 14, and converter 16 described in connection with Figure 1.
• the gain k of the amplifier 12 is chosen small enough that the saturation of the track is unlikely, or occurs occasionally but for short periods. This gain k must however be sufficient for a speech signal from a microphone placed far from the mouth is treatable by the DSP.
• the first channel has a low distortion on the entire dynamics of the input signals.
• the DSP will be able to detect the signature of the speech.
• This first channel is analyzed by a process 26 of the DSP which detects the speech signature and calculates filter parameters according to the characteristics of the signature.
• the signal from the microphone 10 is transmitted to a second process 28 of the DSP by a second channel comprising a gain amplifier K, a filter 32 attenuating the frequencies out of the band of speech, and an analog / digital converter 34.
• the gain K of the amplifier 30 is chosen to produce an audible speech signal under most conditions. It does not matter whether the channel saturates with peaks of ambient noise because this channel is not used for speech detection.
• the gain K is variable and controlled by the process 26 in order to better adapt the amplitude of the signal to the dynamics of the second channel.
• the gain is determined, for example, according to the envelope of the signal carried in the first channel.
• a dynamic compressor for example incorporated in the amplifier 30.
• a dynamics compressor will introduce more distortion into the signal. situations where the path would not be saturated, but it has the advantage of producing a more intelligible signal in saturation situations.
• the process 28 implemented by the DSP on the second channel performs a noise filtering using the parameters calculated by the process 26.
• This filtering can, as in the patent application EP 1843326, consist in subtracting the ambient noise of the signal , thereby preserving the speech signal.
• Figure 3 shows the device of Figure 2 embellished with several improvements that can be used together or separately. To improve the noise level situation during the silent phases, it is expected to inhibit the output of the second channel during the phases where the process 26 does not detect a speech signature. This functionality is symbolized by a gate 36 disposed in the second channel (30, 32, 34) upstream of the filtering process 28.
• the gain of the amplifier 12 will be adapted to the sensitivity of the microphone. This can of course be done by providing a manual gain setting, such as a switch, but it goes against this type of terminal that wants to be ready to operate under any circumstances.
• the terminal will be preferred to equip the terminal with an automatic detection means of the type of microphone used.
• the professional-quality microphones used with such terminals are not equipped with connectors, so that the terminal manufacturer can equip them with the connectivity of his choice.
• the microphones are not equipped with connectors, so that the terminal manufacturer can equip them with the connectivity of his choice.
• the microphones are not equipped with connectors, so that the terminal manufacturer can equip them with the connectivity of his choice.
• the microphone 10 is provided with a connector 38 incorporating, for example, a resistor 40 of specific value associated with the type of microphone. This resistor is connected between a ground terminal GND of the connector and an identification terminal 42 of the connector.
• the terminal 42 is connected to a supply voltage Vcc by a current source 44.
• the voltage drop across the resistor 40 is converted into digital by a converter 46 and analyzed by the process 26.
• the process 26 adjusts the gain k of the amplifier 12 and possibly other parameters, such as a bias current required for "electret" type microphones.
• the bias current is supplied, for example, by a current source 48 connected between the potential Vcc and a dedicated terminal of the connector 38.
• the analysis process 26 also receives the signal from the second channel. This makes it possible, if necessary, to implement in the analysis process 26 of the finer algorithms of signature detection and determination of filter parameters.
• burst noise is an audible frequency noise generated by the envelope of RF signals that alternate between transmission and reception.
• the system described here can be applied to the detection of other signatures.
• the process 26 may also be provided to detect the signature of a whistle.
• the signature detection is used to trigger a signal. The signal may be sent to a particular terminal which will make the desired use.

Landscapes

• Engineering & Computer Science (AREA)
• Human Computer Interaction (AREA)
• Quality & Reliability (AREA)
• Signal Processing (AREA)
• Health & Medical Sciences (AREA)
• Audiology, Speech & Language Pathology (AREA)
• Computational Linguistics (AREA)
• Physics & Mathematics (AREA)
• Acoustics & Sound (AREA)
• Multimedia (AREA)
• Circuit For Audible Band Transducer (AREA)
• Soundproofing, Sound Blocking, And Sound Damping (AREA)
• Telephonic Communication Services (AREA)

Applications Claiming Priority (2)

Publications (1)

Family

ID=41668720

Family Applications (1)

Country Status (6)

Families Citing this family (1)

Family Cites Families (7)

• 2009
  • 2009-06-23 FR FR0903038A patent/FR2947122B1/fr not_active Expired - Fee Related
• 2010
  • 2010-06-22 CA CA2766293A patent/CA2766293A1/en not_active Abandoned
  • 2010-06-22 EP EP10736749A patent/EP2446436A1/de not_active Withdrawn
  • 2010-06-22 US US13/379,451 patent/US8762140B2/en not_active Expired - Fee Related
  • 2010-06-22 CN CN2010800374550A patent/CN102483927B/zh not_active Expired - Fee Related
  • 2010-06-22 WO PCT/FR2010/000457 patent/WO2010149875A1/fr active Application Filing

Non-Patent Citations (1)

Also Published As

Similar Documents

Legal Events