Patent Application
20070129022
The invention relates in general to mobile radio communications, and more particularly to the link quality of mobile communications between mobile communication radios.
In radio voice communication, link problems can arise between radios due to noise and other interference with the transmitted signal in the signal channel. Interference from other sources may disturb the channel. Multi-path effects of the signal being transmitted are also known to degrade link quality. Although efforts may be taken to reduce the effect of interference, often the interference combines with the signal in a way that results in corruption of the information being transmitted. The result is that the listener hears the effects of the interference, potentially corrupting the information to the point of making it, or parts of it, unintelligible.
However, interference in the channel between radios is not the only way in which a signal may become perceptively corrupted or distorted. Digital mobile voice communication relies on voice coding, or voice encoding, techniques to reduce the bandwidth needed to transmit voice information. Voice coding is performed by a vocoder, and takes advantage of the relatively slow time varying nature of speech, as well as other aspects of speech, to model speech with a set of parameters and coefficients. When speech in the acoustic audio signal is mixed with other acoustic sounds, generally referred to as background noise, the voice coding process becomes less effective, resulting in audio artifacts being mixed in with speech at the listener's equipment.
To the listener, it is difficult to determine, if possible at all, whether degraded speech received at their equipment from another party is the result of ambient acoustic conditions at the remote party, of if it is due to interference in the radio channel. Typically the listener will prompt the speaking party to “speak up,” and increase their speaking volume. The speaker, of course, has no way of knowing what their speech sounds like at the listener's equipment. Furthermore, communications equipment is constantly measuring radio link quality to determine if some communication process needs to be undertaken, such as a handover to a new serving cell, or to inform a user's equipment to increase transmission power. However, communication equipment does not evaluate the voicing quality of the information it is receiving for transport to another party, only the radio link quality. Therefore, there is a need by which users of communications systems can be informed as to the voicing quality of the speech signal they are transmitting so that preemptive action may be taken to address radio and voice processing issues.
While the specification concludes with claims defining the features of the invention that are regarded as novel, it is believed that the invention will be better understood from a consideration of the following description in conjunction with the drawing figures, in which like reference numerals are carried forward. The invention solves the problem of informing communication system users of the voicing quality of the speech signal they are transmitting by determining a voicing quality metric based on vocoder parameters produced while vocoding the audio signal at the speaker's equipment, and either making the voicing quality metric perceptible to the user, or using the voicing quality metric to compare with the radio signal reception quality to determine if there is a problem with the radio link.
Referring now to
According to an embodiment of the invention, as the user of the mobile communication device speaks into the microphone 108, an audio signal is digitized by the audio processor 106 and fed to the vocoder 104 for encoding. The audio signal contains the speech signal as well as any ambient noise that may have been present at the microphone while the user speaks. The vocoder encodes the speech and in so doing determines certain coefficients and parameters of the audio signal on a frame by frame basis. Typically included in the output of the vocoder is a voicing level parameter and a background noise parameter. The voicing level parameter indicates the degree to which the present frame appears to be voiced content, and may depend on certain characteristics such as pitch, pitch trajectory, periodicity, and so on. In frames that do not appear to be voiced, the vocoder may provide a noise estimation corresponding to the non-voiced content. During periods where the user is not speaking, the vocoder may output what are referred to as comfort noise frames which provide minimal acoustic content so that the receiving party still “hears” the user's call because a completely silent frame will often make a listener think the call has been disconnected or interrupted. The voicing level parameter may be directly output from the vocoder, or it may be determined by mathematical operation on a combination of parameters output by the vocoder. Comparing the voicing level parameter to the background noise parameter is one way of providing a voicing quality metric that indicates how well the speaker's voice overcomes the ambient noise. Various other vocoder parameters, depending on the vocoder, may be used to generate the voicing quality metric, so long as they relate to how well the speaker's voice overcomes ambient noise. The comparison may be made simply by a ratio of voicing level to background noise, but it is contemplated that these parameters may be scaled or weighted, or even adaptively scaled or weighted depending on the acoustic circumstances. If the voicing quality metric indicates the volume of the speaker's voice is not enough to sufficiently overcome the ambient noise, an indication may be given to the speaker to increase their speaking volume. The indication is provided in the form of perceptible feedback, such as, for example, visual, audible, tactile, feedback, or combinations thereof. For example, a light source such as an LED may be blinked, or flashed, or provided at a certain color to indicate a voicing quality problem. A vibration device may be employed equivalently to provide tactile feedback to the speaker, as another example. The precise selection of value in comparing the voicing quality metric with the background noise metric at which to prompt the user to adjust speaking manner is a matter of engineering choice, and will depend on the particular vocoder parameters used, the algorithm used by the vocoder, and the present acoustic conditions, for example. Furthermore, it is contemplated that the vocoder, upon detecting audio saturation, may reduce the gain of the microphone input. Audio saturation can occur in noisy environments where the user speaks much louder than normal, possibly over-compensating for ambient noise.
Referring now to
Referring now to
In another embodiment, infrastructure equipment such as the base radio and other infrastructure components conventionally pass the signal from one mobile communication device 302 to the remote mobile communication device 312. The remote mobile communication device 312 receives the signal originated by the calling mobile communication device 302 from a base radio 304. During a call between the mobile communication devices the mobile communication devices determine their respective voicing quality metrics, and insert the voicing quality metric into their outbound signal. As the mobile communication devices receive the respective signals, if the audio quality appears to be poor, the receiving mobile communication device can compare the reception quality metric and the received voicing quality metric to determine where the problem is occurring. If both metrics are within acceptable limits, then the assumption is there is a link problem between the transmitting mobile communication device and its respective base radio. In response, the receiving mobile communication device may send a message to the sending mobile communication device so that the sending mobile communication device may decide whether or not to search for a new base station 314, increase transmit power, prompt the user to raise the antenna, or other such corrective actions to improve transmit signal quality.
If the voicing quality metric is low then the assumption is ambient noise is interfering with the voice encoding process at the speaker's mobile communication device. In this case the receiving mobile communication device may reply with a message to the sending mobile communication device to prompt the user of the sending mobile communication device to speak differently (louder, slower, clearer, etc.) via any of the perceptible feedback modalities discussed previously. If the reception quality metric and voicing quality metric indicate a problem with reception from the present serving base station 304, the mobile communication device may search for a new base station 314 for a high quality signal.
Alternatively, the base station may send a control message to the mobile communication device to prompt the user to speak differently in an attempt to overcome channel interference, even when the voicing quality metric indicates the mobile communication device is producing high fidelity encoded voice.
In an embodiment of the invention, the mobile communication device determines an audio quality metric of the voice, such as a ratio of voicing level to background noise, a pitch trajectory, or another attribute representative of continuous voice. Pitch results from vibration of the vocal chords which are normally continuous and smooth over sufficiently short time periods. Speech varies relatively slowly because of the articulation requirements in producing voice. Hence, an audio quality metric or measure should be able to identify the attributes of good form voice, i.e. voice in good form with true voice characteristics. When the vocoder is poorly encoding, which can be due to background noise, the encoding performance becomes reduced and the parameters representing the encoded speech may not adequately comply with the slowly time varying model. For example, the vocoder parameters representing the pitch should have a smooth trajectory over vowelic portions of speech because vowelic speech is periodic. With significant background noise, the encoding may perturbate the pitch for certain frames which lead to a jumpy or jittery pitch track, or pitch rejection. The audio quality metric is not coupled with or related to the link quality estimates which use forward error correction (FEC) values or bit energy over noise figures (Eb/No) to determine signal strength. Improved link performance is typically attained with higher power or more forward error correction coding.
The mobile communication device may packetize the voicing quality metric within silence frames between speech content frames and send it to the base radio. The base radio is operably coupled to a transcoder, as is well known in the art. The transcoder decodes the data and assess the voice quality as well as the link quality. Note, if the link quality is bad, the audio quality will become corrupted. But if the link quality is good and the audio quality is poor, then the base radio assumes the mobile communication device is having difficulty encoding, likely due to environmental conditions like background noise. If so, the base radio may send a control packet to the mobile communication device to prompt the user to change speaking manner. The user may speak louder, slower, with more articulation, and so on. The base radio can determine if the voicing quality is poor by comparing an estimate it makes on the decoded speech with the voicing quality metric transmitted by the mobile communication device. The base radio may use the same evaluation criteria, and if the voicing quality metric produced by the base radio do not sufficiently match that provided by the mobile communication device, then it indicates a link problem. Link quality measured in this manner can be confirmed with the SQE and RSSI measurements.
If the vocoding at the mobile communication device is acceptable but the link is poor, then speaking slower may help retain the quality over the link because there are more packets given the same number of encoding errors. For example, if the link distorts every other frame, by slowing speech in half, then twice as many good frames are received, which, although doesn't improve quality so much as intelligibility. Alternatively, if the user articulates differently, which affects pronunciation, inflection, intonation, and pitch, it may condition the speech to be more robust to the communication channel errors. In practice the errors are randomly distributed, like burst, convolution, and so on. If the user speaks with more inflection than the pitch trajectory has more emphasis, it is more dynamic. If the user speaks with more intonation, the pitch track better exhibits characteristic behaviors, such as an upward frequency sweep near the end of words. Vocoders depend on pitch and are sensitive to reconstruction errors when the pitch estimates are off. Accordingly, link errors can affect pitch values which will deteriorate reconstructed audio quality. Accordingly, prompting the user to speak differently can alter the way the decoder reconstructs speech in poor link conditions. Also, in variable rate vocoders or differential vocoders, providing pronounced variation or limiting variation can affect the encoding process. Highly inflected voices will have wider dynamic range which can reduce sensitization to quantization noise. Increasing inflection thus may improve robustness when link errors corrupt the encoding parameters. Wider dynamic range requires the vocoder to allocate more bits to those fields. Requesting the user to speak differently introduces variance and redundancy into the encoding parameters which can increase the robustness of the parameters across the link.
In addition to the voicing quality, the mobile communication device, according to one embodiment of the invention, may provide a comprehensive signal quality indication, including the present voicing quality and the present radio link quality. The present radio link quality is based on the signal received from a base radio with which the mobile communication device is presently affiliated, and may be based on radio signal strength index (RSSI), signal quality estimation (SQE), the degree of forward error correction, bit error rate, and so on. The mobile communication device may indicate its radio link quality or comprehensive signal quality to the user via perceptible feedback. The comprehensive signal quality, which may include both the voicing quality metric and radio link quality metric, indicates the overall quality of the signal as received by the base radio, but may be passed on to the remote party. Having both indicators of voicing quality and radio link quality allows for decisions as to what sort of communication action may be taken to improve the quality of the signal heard by the remote party. If the voicing quality is high, but the radio link quality is low, remote radios may send control messages to the originating mobile communication device to improve signal conditions, such as by raising an antenna, increasing transmit power, handing over to a different serving cell, and so on. Likewise, if radio link quality is high, but the voicing quality is low, remote radios may send control messages to prompt the user of the originating mobile communication device to speak differently (louder, slower, etc). in an attempt to improve voicing quality.
Furthermore, the comprehensive signal quality received from the initiating party may be high, while the remote party's reception quality may also be high, but the decoded audio quality at the remote party may be low. Such conditions indicate a problem in the link from the initiating party to the communication system, also known as the inbound link. In this situation the remote party's device may respond with a control message to the initiating party to take corrective action, such as adjusting transmission parameters, or prompting the user to speak differently.
Referring now to
Referring now to
Thus, the invention provides a method of providing voicing quality feedback to a user of a mobile communication device, commenced by receiving an audio signal containing voice information at the mobile communication device, via, for example a microphone, and vocoding the audio signal into a vocoded signal. The output of the vocoder is used for determining a present voicing quality metric of the vocoded signal. The method then commences providing perceptible feedback to the user of the mobile communication device in correspondence to the present voicing quality metric. The perceptible feedback may include provide a visual, audible, or tactile feedback. In one embodiment the voicing quality metric is determined by the voicing level parameter and the background noise parameter, such as by a ratio.
The invention also provides a method of informing a remote party of a calling party's voicing quality, commenced by receiving an audio signal containing voice information at the mobile communication device, followed by vocoding the audio signal into a vocoded signal. The output data of the vocoder or vocoding process is used in determining a present voicing quality metric of the vocoded signal. The mobile communication device then commences transmitting the present voicing quality metric to the remote party, and may further transmit a comprehensive signal quality indication including the voicing quality metric and a radio ling quality metric. The voicing quality metric may be determined by the voicing level parameter, pitch parameter, pitch trajectory parameter, along with a parameter indicating back ground noise. The voicing quality metric may be inserted into silence period frames, or in fields in the frame structure with the vocoded voice data. The voicing quality metric may be used by the receiving mobile communication device to let the remote party know audio quality is suffering due to link conditions, not because of ambient noise at the speaker's equipment.
While the preferred embodiments of the invention have been illustrated and described, it will be clear that the invention is not so limited. Numerous modifications, changes, variations, substitutions and equivalents will occur to those skilled in the art without departing from the spirit and scope of the present invention as defined by the appended claims.
