JP2018022956A - Reaction estimation device, reaction estimation method, and program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To enable the estimation of reaction of a user due to quality deterioration in voice communication.SOLUTION: A reaction estimation device includes: a calculation unit for calculating a sound interruption rate, i.e. a ratio of a period which voice is interrupted in a communication period of voice via a network; and an estimation unit for estimating reaction of a user by applying the sound interruption rate calculated by the calculation unit to a relational expression indicating correspondence relation between the voice interruption rate and reaction of users due to interruption of voice.SELECTED DRAWING: Figure 2

Description

  The present invention relates to a reaction estimation device, a reaction estimation method, and a program.

  In order to maintain and improve the voice call service, it is important for the communication carrier to grasp the quality of the service and examine how the service quality affects the user. In particular, the situation in which the user perceives a specific quality degradation event (for example, whether it is difficult to hear a conversation; hereinafter referred to as “perception situation”), and the willingness to act as a result of perceiving the quality degradation event. (For example, whether you want to hang up the call and try again) (hereinafter referred to as “behavior intention”) is useful in improving the service.

Okuyama Takafumi, Kurashima Atsuko, Masuda Yuki, "A Study on Estimation Method of Listening Quality in VoLTE," IEICE Societies, B-11-21, pp.438, March 2015

  Although it is possible to directly ask the user about their perception status and behavioral intention when using the service in order to grasp the user's perception status and behavioral intention, there are problems such as reduced usability and cost and time required for installation of the aggregation system It becomes. For this reason, it is desirable to estimate the perceived state and action intention by acquiring data within the network or terminal without directly asking the user.

  Non-Patent Document 1 has been proposed as a method for estimating a subjective evaluation result of quality degradation using a packet. In Non-Patent Document 1, a listening MOS (Mean Opinion Score) is estimated in consideration of packet loss characteristics. However, Non-Patent Document 1 only estimates the listening MOS, and has a disadvantage that it cannot be applied to estimation of reactions such as perception status and action intention.

  The present invention has been made in view of the above points, and it is an object of the present invention to enable estimation of user reaction due to quality degradation of voice communication.

  Therefore, in order to solve the above-described problem, the reaction estimation apparatus includes a calculation unit that calculates a sound interruption rate that is a ratio of a period in which sound is interrupted in a communication period of sound through the network, the sound interruption rate, and the sound is interrupted. An estimation unit that estimates a user's reaction by applying a sound interruption rate calculated by the calculation unit to a relational expression indicating a correspondence relationship with a user's reaction.

  It is possible to estimate the user reaction due to the deterioration of voice communication quality.

It is a figure which shows the hardware structural example of the reaction estimation apparatus in 1st Embodiment. It is a figure which shows the function structural example of the reaction estimation apparatus in 1st Embodiment. It is a flowchart for demonstrating an example of the process sequence which the reaction estimation apparatus in 1st Embodiment performs. It is a figure which shows the function structural example of the reaction estimation apparatus in 2nd Embodiment. It is a flowchart for demonstrating an example of the process sequence which the reaction estimation apparatus in 2nd Embodiment performs.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings. FIG. 1 is a diagram illustrating a hardware configuration example of the reaction estimation apparatus according to the first embodiment. The reaction estimation device 10 in FIG. 1 includes a drive device 100, an auxiliary storage device 102, a memory device 103, a CPU 104, an interface device 105, and the like that are mutually connected by a bus B.

  A program for realizing the processing in the reaction estimation device 10 is provided by a recording medium 101 such as a CD-ROM. When the recording medium 101 storing the program is set in the drive device 100, the program is installed from the recording medium 101 to the auxiliary storage device 102 via the drive device 100. However, the program need not be installed from the recording medium 101 and may be downloaded from another computer via a network. The auxiliary storage device 102 stores the installed program and also stores necessary files and data.

  The memory device 103 reads the program from the auxiliary storage device 102 and stores it when there is an instruction to start the program. The CPU 104 executes a function related to the reaction estimation device 10 in accordance with a program stored in the memory device 103. The interface device 105 is used as an interface for connecting to a network.

  FIG. 2 is a diagram illustrating a functional configuration example of the reaction estimation apparatus according to the first embodiment. In FIG. 2, the reaction estimation apparatus 10 includes a sound interruption rate calculation unit 11 and a perceptual rate estimation unit 12. Each of these units is realized by processing that one or more programs installed in the reaction estimation apparatus 10 cause the CPU 104 to execute.

  In the first embodiment, captured voice packet capture data is input to the reaction estimation apparatus 10. The protocol of the packet flow uses RTP (Real-time Transport Protocol) that can identify the loss location and calculate the length of the loss interval from the sequence number and transmission time stamp. Note that it is not essential to capture packets in real time. It suffices if the packet can be acquired by some method. Further, the network is, for example, a mobile network, but the present embodiment is not limited to the mobile network, and can be applied to perceptual behavior estimation regarding voice calls in various networks.

  The sound interruption rate calculation unit 11 calculates a sound interruption rate that is a ratio of a period in which sound is interrupted in one call period (communication period). Specifically, the sound interruption rate calculation unit 11 calculates the sum (total sound interruption time) of the time length (sound interruption time) of the sound interruption period (period) specified by the sequence number of the packet header or the transmission time stamp. calculate. The sound interruption rate calculation unit 11 calculates the sound based on the ratio between the total time length of one call (for example, 1 minute, 60 minutes, etc.) and the total sound interruption time (for example, 10 seconds, 50 seconds, etc.). A break rate (for example, 1%, 3%, etc.) is calculated.

  The perception rate estimation unit 12 holds a correspondence relation expression (mapping function) indicating a correspondence function between the sound interruption rate and the perception rate. The perceptual ratio refers to the ratio of users who perceive deterioration in quality of voice communication (for example, the ratio of how many users out of 100 perceive quality deterioration). The perception ratio estimation unit 12 derives an estimated value of the perception ratio (hereinafter referred to as “perception ratio estimation value”) by substituting the sound interruption rate calculated by the sound interruption rate calculation unit 11 into the corresponding relational expression. A detailed example of the relational expression will be described later in the processing flow, and in the first embodiment, the perceived state of deterioration of voice communication quality is an example of user reaction.

  Hereinafter, the process procedure which the reaction estimation apparatus 10 performs is demonstrated. FIG. 3 is a flowchart for explaining an example of a processing procedure executed by the reaction estimation apparatus according to the first embodiment.

  For example, in a voice call service such as VoLTE (Voice over Long Term Evolution) connected via a network, a call is performed between terminals. The transmitted voice of a call is encoded, stored in a packet, reaches the receiving terminal via the network, is decoded by the receiving terminal, and the voice is output and received. Call packet capture data (hereinafter simply referred to as “packets”) acquired within the network or at the receiving terminal is input to the sound interruption rate calculation unit 11 of the reaction estimation apparatus 10 (S101).

  When a packet is input, the sound interruption rate calculation unit 11 specifies a loss point of the packet from the sequence number included in the packet related to one call (S102). Specifically, a part where the sequence numbers are not continuous is specified as a loss part.

  Subsequently, the sound interruption rate calculation unit 11 acquires the total time length of one call including the packet loss point from the transmission time stamp of the packet header (S103). Specifically, the total time length can be acquired by subtracting the time stamp of the first packet from the time stamp of the last packet of one call. Strictly speaking, the call time for the last packet may be added to the value obtained by subtracting the time stamp of the first packet from the time stamp of the last packet of the call. This is because the time stamp of each packet indicates the start time of the call time of the packet.

  Subsequently, for each packet loss point specified in step S102, the sound interruption rate calculation unit 11 calculates the sound interruption time of the loss point from the transmission time stamps included in the packets before and after the loss point (S104). ). Specifically, by subtracting the call time of the previous packet from the time obtained by subtracting the time stamp of the packet immediately before the loss location from the time stamp of the packet immediately after the loss location, the loss location The sound interruption time can be calculated.

  Subsequently, the sound interruption rate calculation unit 11 calculates the total sound interruption time by summing the sound interruption times for each loss point (S105). Subsequently, the sound interruption rate calculation unit 11 calculates the sound interruption rate from the ratio of the total time length and the total sound interruption time length (S106). That is, the sound interruption rate is calculated based on “total sound interruption time ÷ total time length of call”.

  Subsequently, the perceptual rate estimation unit 12 calculates the sound discontinuity rate calculated by the sound discontinuity rate calculation unit 11 with respect to the correspondence expression that indicates the correspondence relationship between the sound discontinuity rate and the perceived rate of quality degradation. By substituting, a perceptual ratio estimated value is calculated (S107). Subsequently, the perception ratio estimation unit 12 outputs the calculated perception ratio estimation value (S108).

  An example of a correspondence expression (perception estimation model) indicating the correspondence between the sound interruption rate and the perception ratio of quality degradation is shown below.

Perception ratio = A × sound break rate 2−B × sound break rate + C
The corresponding relational expression is a function having the sound interruption rate as an input, which is fitted to the relationship between the sound interruption rate and the perception rate obtained in advance by a subjective evaluation experiment. Based on the relationship between the sound interruption rate and the perception rate obtained through experiments, for example, the above function is defined as a mapping function for estimation of the perception rate. The coefficients A, B, and C are derived so as to obtain a function shape with high accuracy.

  The above mapping function for perceptual ratio estimation is an example. For example, an exponential function having the sound interruption rate as an input may be used as the mapping function for estimating the perception ratio.

  As described above, according to the first embodiment, it is possible to estimate a user reaction (perceived state of quality degradation) due to voice communication quality degradation using packet capture data. Further, according to the present embodiment, it is possible to estimate a user's reaction due to deterioration in voice communication quality in each call without requiring subjective evaluation for each call.

  Next, a second embodiment will be described. In the second embodiment, differences from the first embodiment will be described. Points that are not particularly mentioned in the second embodiment may be the same as those in the first embodiment.

  FIG. 4 is a diagram illustrating a functional configuration example of the reaction estimation apparatus according to the second embodiment. In FIG. 4, parts corresponding to those in FIG. In FIG. 4, the reaction estimation apparatus 10 a includes a sound interruption rate calculation unit 11 and an action rate estimation unit 13. Each of these units is realized by processing that one or more programs installed in the reaction estimation apparatus 10 cause the CPU 104 to execute.

  In the second embodiment, a call voice signal is input to the reaction estimation apparatus 10a. Note that it is not essential to obtain an audio signal in real time. It suffices if the audio signal can be acquired by some method. Further, the network is, for example, a mobile network, but the present embodiment is not limited to the mobile network, and can be applied to perceptual behavior estimation regarding voice calls in various networks.

  The sound interruption rate calculation unit 11 calculates the sum (total sound interruption time) of the time length (sound interruption time) of the sound interruption section specified by the signal level. The sound interruption rate calculation unit 11 calculates a sound interruption rate (for example, 10 seconds, 50 seconds, etc.) and a total sound interruption time (for example, 10 seconds, 50 seconds, etc.) For example, 1%, 3%, etc.) are calculated.

  The action ratio estimation unit 13 holds a correspondence expression (mapping function) indicating a correspondence relation between the sound interruption rate and the action ratio with respect to quality degradation. The action ratio refers to a ratio of users who show a specific or arbitrary action intention with respect to quality degradation (for example, a ratio of how many users out of 100 show a specific action intention). The action ratio estimation unit 13 derives an estimated value of action ratio (behavior ratio estimation value) by substituting the sound interruption rate calculated by the sound interruption rate calculation unit 11 into the corresponding relational expression. In the second embodiment, the intention to act for voice communication quality degradation is an example of a user reaction.

  FIG. 5 is a flowchart for explaining an example of a processing procedure executed by the reaction estimation apparatus according to the second embodiment.

  For example, in a voice call service such as VoLTE connected via a network, a call is performed between terminals. The transmitted voice of a call is encoded, stored in a packet, reaches the receiving terminal via the network, is decoded by the receiving terminal, and the voice is output and received. The audio signal decoded from the call packet acquired in the network or at the receiving terminal, or the audio signal directly acquired at the receiving terminal is input to the sound interruption rate calculating unit 11 of the reaction estimation apparatus 10a (S201).

  When the sound signal is input, the sound break rate calculation unit 11 specifies the sound break point from the level of the sound signal (S202). Silent sections and sound interruptions in conversation are distinguished from each other by the difference in level. Alternatively, a voice signal that is always sounded may be input from the transmission side, and the sound interruption point may be distinguished depending on the level.

  Subsequently, the sound interruption rate calculation unit 11 acquires the total time length of the call including the sound interruption point from the voice signal length (S203). Subsequently, the sound interruption rate calculation unit 11 calculates, for each sound interruption point specified in step S202, a sound interruption time at the sound interruption point from a comparison of levels before and after the sound interruption point (S204). Subsequently, the sound interruption rate calculation unit 11 calculates the total sound interruption time by summing the sound interruption times for each sound interruption point (S205). Subsequently, the sound interruption rate calculation unit 11 calculates the sound interruption rate from the ratio of the total time length and the total sound interruption time length (S206). That is, the sound interruption rate is calculated based on “total sound interruption time ÷ total time length of call”.

  Subsequently, the action rate estimation unit 13 calculates the sound interruption rate with respect to a correspondence relation expression that indicates a correspondence relationship between the sound interruption rate and the proportion of users who show an intention to act with respect to quality degradation (action rate). A behavioral rate estimated value is calculated by substituting the sound interruption rate calculated by the calculation unit 11 (S207). Subsequently, the action ratio estimation unit 13 outputs the calculated action ratio estimation value (S208).

  An example of a correspondence expression (behavior ratio estimation model) indicating a correspondence relation between the sound interruption rate and the action ratio with respect to quality degradation is shown below.

Action ratio = D × sound break rate 2−E × sound break rate + F
The corresponding relational expression is a function having the sound interruption rate as an input, which is fitted to the relationship between the sound interruption rate and the action ratio obtained in advance by a subjective evaluation experiment. Based on the relationship between the sound interruption rate and the action rate obtained by experiment, for example, a function as described above is defined as a mapping function for estimating the action rate. The coefficients D, E, and F are derived so as to obtain a function shape with high accuracy.

  The above mapping function for estimating the behavior ratio is an example. For example, an exponential function that uses a sound interruption rate as an input may be used as a mapping function for estimating a behavior ratio.

  As described above, according to the second embodiment, it is possible to estimate a user's action intention due to voice communication quality degradation using a voice signal. According to the present embodiment, it is possible to estimate a user's reaction due to quality degradation of voice communication in each call without requiring subjective evaluation for each call.

  Note that the method for calculating the sound interruption rate in the second embodiment may be applied to the first embodiment, and the method for calculating the sound interruption rate in the first embodiment is the second method. It may be applied to the form.

  In the present embodiment, the sound interruption rate calculation unit 11 is an example of a calculation unit. The perception rate estimation unit 12 or the behavior rate estimation unit 13 is an example of an estimation unit.

  As mentioned above, although the Example of this invention was explained in full detail, this invention is not limited to such specific embodiment, In the range of the summary of this invention described in the claim, various deformation | transformation・ Change is possible.

10, 10a Reaction estimation device 11 Sound interruption rate calculation unit 12 Perception rate estimation unit 13 Action rate estimation unit 100 Drive device 101 Recording medium 102 Auxiliary storage device 103 Memory device 104 CPU
105 Interface device B bus

Claims (7)

A calculation unit that calculates a sound interruption rate that is a ratio of a period in which the sound is interrupted in a communication period of the sound via the network;
An estimation unit that estimates a user's reaction by applying the sound interruption rate calculated by the calculation unit to a relational expression indicating a correspondence relationship between the sound interruption rate and a user's reaction due to the interruption of sound;
A reaction estimation apparatus comprising: The estimation unit applies the sound interruption rate calculated by the calculation unit to a relational expression indicating a correspondence relationship between the sound interruption rate and a ratio of users perceiving the sound interruption, and the user who perceives the sound interruption Calculate an estimate of the percentage,
The reaction estimation apparatus according to claim 1. The estimation unit applies the sound interruption rate calculated by the calculation unit to a relational expression indicating a correspondence relationship between the sound interruption rate and a ratio of users indicating intention to perform an action when the sound is interrupted. Calculate an estimate of the percentage of users who indicate their intention to
The reaction estimation apparatus according to claim 1. The calculation unit calculates the sound interruption rate using the total time length of the voice communication and the total time length of the sound interruption acquired from the capture data of the voice communication packet,
The reaction estimation apparatus according to any one of claims 1 to 3, wherein The calculation unit calculates the sound interruption rate using the total time length of the voice communication and the total time length of the sound interruption acquired from the voice signal of the voice communication,
The reaction estimation apparatus according to any one of claims 1 to 3, wherein A calculation procedure for calculating a sound interruption rate, which is a ratio of a period in which the sound is interrupted in a communication period of the voice via the network;
An estimation procedure for estimating a user's reaction by applying the sound interruption rate calculated in the calculation procedure to a relational expression indicating a correspondence relationship between the sound interruption rate and a user's reaction due to interruption of sound;
A reaction estimation method characterized in that a computer executes.   The program for functioning a computer as each part in the reaction estimation apparatus as described in any one of Claims 1 thru | or 5.

Images