DE102004041199B4 - Process and facility calculates damping factor for the damping of acoustic signals, multiplies first acoustic signal with damping value to generate output signal - Google Patents

Abstract

The invention concerns a process and the facility for the damping of acoustic signals. The process contains the following steps: - Obtaining a first acoustic signal s(t) and a second acoustic signal s(t-dt), which is delayed with a delay time dt compared to the first acoustic signal. - Generating a first signal power P1 from the first acoustic signal and a second signal power P2 from the second acoustic signal. - Determining the increase between the first and the second signal power. - Damping of the first acoustic signal s(t) with a damping factor B if the increase exceeds a threshold K, where the damping includes the following steps: - Determining the damping value A from the ratio of the first and the second signal power. - Generating the damping factor B through successive approximation of the damping value B from a start value to the damping value A. - Multiplying the first acoustic signal with the damping value to generate the output signal. - Independent claims are also included for; - (1) a device; - (2) a computer program product.

Description

The The present invention relates to a method and an apparatus for steaming of sound signals.

The The invention is in the field of signal processing for elimination of acoustic shock signals to the hearing of at least one user before suddenly onset of loud signals, the so-called shock signals suitable steaming to protect these signals.

The Elimination of acoustic shock signals in particular undergoes a growing Importance by the increase of often mobile devices, such as z. As mobile phones or other electro-acoustic devices, at which acoustic signals for Users are issued to protect the user from shock damage, to protect. Especially with the output of the acoustic signals via headphones, ear studs or the like can the shock signals to damage of hearing lead the user or just be uncomfortable.

to Avoidance of shock signals are known in the art Compressors known by means of which an audio signal according to a given characteristic curve all the more attenuated the louder it is. It has, however, as a disadvantage proved that the compressor works according to its characteristic only depending on the volume, and no further criteria of a potential shock signal are considered so that shock signals are not reliably eliminated.

Other Known devices use to reduce shock signals simple, analog circuits. For example, EP-A-1 383 298 a mobile telecommunication terminal, in which by means of a simple analog filter circuit and a second loudspeaker a sound signal is generated, the shock signals of a first speaker in an ear stud for to eliminate the user. The disadvantage is in particular that suddenly occurring shock signals by such devices not and also not completely can be eliminated and it about it out comes to artifacts in the resulting output sound signal and thus total shock signals can not be reliably eliminated.

A method and system for protection against acoustic shock signals is in the post-published EP 1 471 767 A2 described. Here, a pattern analysis based approach is applied to an input signal to perform feature extraction. A parameter space corresponding to the signal space of the input signal is identified. To detect an acoustic shock event, a rule-based decision-making approach is applied to the parameter space. Using such a technique, intelligent sound limiters can quickly and accurately detect and eliminate acoustic shock signals.

It is therefore an object of the invention, a method and an apparatus for steaming of sound signals indicating the sudden onset of shock signals safe and reliable eliminate without this a large Expenditure on memory requirements and computing power is required.

According to the invention this Task by a method according to claim 1 and a device according to Claim 10 and a computer program according to claim 11. advantageous Further developments of the invention are defined in the subclaims.

The inventive method for steaming of sound signals react very quickly to sudden onset of loud signals, whose power increase exceeds a predetermined threshold K. These, exceeding the threshold K Signals are recognized by the method as shock signals. Occurs according to rapidly rising shock signal, it gets so strong steamed, that its volume corresponds to the signal level before the onset of the shock signal. On This way, the known as shock signals popping sounds, such as z. B. sudden loud cracking or other sudden effectively eliminates short-term shock signals.

The method is expediently implemented as digital signal processing. The method makes low demands on memory requirements and computing power, so that the method can already be implemented on inexpensive microcontrollers. In such an implementation of the invention Method on a common microprocessor, the audio signal s (t) is a digital, discrete-time signal, which was obtained by analog-to-digital conversion of an analog time signal. The term sound signal is therefore understood, unless otherwise stated, according to the present invention, both for the analog time signal and the corresponding digital discrete-time signal.

According to one Aspect of the invention is a first sound signal s (t) and a relation to the first sound signal offset by a delay time .DELTA.t second sound signal s (t - .DELTA.t) detected. From the first and second sound signals are using window functions in each case a first and a second average signal powers P1 and P2 determined. The window functions define the type the time averaging in determining the average signal power. A possible Window function is z. B. a rectangular window with a length L and a height 1 / L, with which a signal power in accordance with the arithmetic mean can be determined is. Especially useful as a window function a so-called exponential window, which is in an analog invention Device as a simple low pass or in a digital invention Device can be realized as a recursive means.

The first and second mean signal power P1 and P2 differ by the time offset Δt between the first and second Sound signal from which the signal powers are generated. One Signal is detected as a shock signal when the slope of the middle Signal power (P1 - P2) / Δt an adjustable threshold K exceeds. A sound signal s (t) recognized as a shock signal according to the invention is a damping factor B steamed. To determine the damping factor B becomes a preliminary Attenuation value A expediently as the square root of the quotient of the mean signal powers P1 and P2 calculated. The attenuation value A represents the value by which the shock signal is to be damped, to achieve a signal level as before the onset of the shock signal. However, as a sudden strong damping the sound signal to artifacts in the output signal and thus unpleasant Can cause auditory noise, becomes an abrupt onset of the attenuation value A to the sound signal avoided. Rather, according to the invention on the sound signal applied damping factor B with a preferably adjustable step size the attenuation value A approximated. The so determined damping factor B is used to generate the output signal with the sound signal s (t) multiplied.

According to one preferred embodiment The invention experienced shock signals compared to blasts over a long time stop, such as B. signal horns or horns generated shock signals, a signal treatment that is can be divided into two variants. In the first variant, the shock signal is attenuated over its entire duration and thus practically inaudible made. However, since in particular with warning signals, such. B. signals of signal horns or horns, this whole Elimination, however, under certain circumstances undesirable and for the user dangerous because he is supposed to be able to perceive these signals is for them Cases the second variant provided according to the damping the shock signal is reduced over time. The reduction the damping is preferably achieved by a simple parameter change, so that the Signal over time for the user becomes perceptible, but the harmful, rapid signal rise eliminated or at least mitigated. According to the second variant remains so the shock signal audible, it only reduces the speed of the volume increase.

Further embodiments The invention will be explained with reference to the accompanying drawings.

1a - 1b show a schematic flow diagram of a method according to another embodiment of the invention.

2a - 2 B show a basic flowchart of a method according to an embodiment of the invention.

3 shows by way of example a window function F (t) by means of which a mean signal power can be generated according to the method according to the invention.

One The basic idea of the invention is to provide shock signals based on a sudden onset strong performance increase. This is received from a Input sound signal a first sound signal s (t) and one opposite to the first sound signal offset by a delay time .DELTA.t second sound signal s (t - .DELTA.t) detected and of these sound signals, the respective signal power P1 and P2 generated.

. Die ersten und zweiten mittleren Signalleistungen werden dann anhand der Formeln

ermittelt. Die jeweils um die Verzögerungszeit Δt versetzten Fensterfunktionen F1(t) und F2(t) definieren dabei die Art der zeitlichen Mittelung bei der Bestimmung der mittleren Signalleistungen. Zur Erzeugung der Signalleistungen P1 und P2 können daher auch andere geeignete Fensterfunktionen verwendet werden.According to one embodiment of the invention, the signal powers P1 and P2 are the first and second average signal power using a respective window function F1 (t) and F2 (t) determined. The window functions are as in 3 shown as an exponential window and fulfill the constraint

, The first and second average signal powers are then calculated using the formulas

determined. The window functions F1 (t) and F2 (t) offset by the delay time Δt in each case define the type of time averaging in the determination of the average signal powers. Therefore, other suitable window functions can be used to generate the signal powers P1 and P2.

Regarding 1a and 1b Now, a method for attenuating sound signals will be explained with reference to a flowchart 100 described in more detail. First, a sound signal s at a time t as the first sound signal s (t) and offset by the delay time .DELTA.t detected as a second sound signal s (t - .DELTA.t) takes place an initialization of the sizes attenuation value A, damping factor B and hold value P3 and the parameter threshold K. , Delay time .DELTA.t, lower power threshold M, and the first, second and third step size S1, S2, S3 as in step 10 specified. In step 20 is then the first signal power P1 from the first sound signal (step 21 ) and the second average signal power P2 from the second sound signal (step 22 ) are respectively generated by means of the window functions F1 (t) and F2 (t). In step 30 it is checked whether the hold value is 0. If this is the case, no valid hold value is currently stored and the method continues in step 50 by determining the slope between the first and second signal powers by determining the term (P1-P2) / Δt and checking whether the slope exceeds the adjustable threshold K. If this is not the case, there is no shock signal according to the method, and the attenuation value A and the hold value P3 are set to their initialization values. However, if the slope is greater than the threshold K, the method detects a shock signal and branches to step 60 in which the value of the second signal power P2 is temporarily stored as a new hold value P3 and in the next step 70 the attenuation value A is calculated as the square root of the quotient of the first signal power and the holding value according to the following formula:

In The following steps will now be applied to the sound signal later damping factor B with an adjustable increment the damping value A approximated that on the one hand prevents or at least reduces artifacts and on the other hand one as possible fast damping the shock signal is reached. With falling damping factor B is therefore a larger increment S2 for approach to the attenuation value A chosen, across from a smaller increment S1 to approach with increasing damping factor. This ensures that the damping with fast rising shock signals sufficiently fast effective becomes.

Furthermore, it is ensured that with increasing damping factor (step size S1), the damping factor B does not exceed the damping value A and that when the damping factor (step S2) decreases, the damping factor B does not fall below the damping value A. The thus determined damping factor B is in step 90 multiplied by the sound signal s (t) so as to obtain a damped output signal.

According to the invention then begins the passage of the process from the front. In step 30 This is after the generation of the average signal power in step 20 checks whether the hold value P3 has a value other than zero, that is, whether a hold value in step 60 which indicates the existence of a shock signal. If so, the performance increase criterion (step 50 ) is not checked to detect a shock signal, instead the method branches to step 40 in which the current first signal power P1 is compared with the hold value P3 and a lower power threshold M. If the test proves that P1 is less P3 or P1 less than M, then according to the method, the shock signal is over and the hold value P3 is set to zero. Otherwise, the method recognizes that the shock signal is still ongoing, and the attenuation value A becomes the new value of the first signal power P1 in step 70 updated. Subsequently, as described above, the damping factor B to be applied to the sound signal is applied to the new one Attenuation value A approximated.

The lower power threshold M is used in the step 40 to avoid a "dead lock", which could otherwise occur if the hold P3 is extremely low, so a shock signal from almost complete silence occurs. Then falls the noise level after the end of the shock signal is no longer below the level of the holding value P3, z. B. if it is not as quiet due to ambient noise as before the shock signal, otherwise would never be detected on "shock signal to end". In order to avoid this "dead lock", it is additionally checked whether the first signal power has at least fallen below the value of M again. If this additional "M criterion" is met, even with a small holding value P3, the state that a shock signal is present is terminated, so that there can be no corresponding "dead lock". In the lower power threshold M is an adjustable value of preferably fixed z. B. is stored in the microcontroller used and is set based on the expected signals and the system properties.

If it is detected by the method that the shock signal is over because the hold value P3 has been set to zero, damping of the sound signal is no longer necessary, so that the damping value A is set to 1. In further runs will be in step 50 again checked whether the criterion for the beginning of a shock signal is present in order, if necessary, to attenuate the sound signal again.

In the 2a and 2 B becomes a flowchart 200 an alternative embodiment of the method according to the invention. However, the alternative embodiment is largely the same as with reference to FIG 1a and 1b described embodiment, wherein identical reference numerals denote identical method steps. In the alternative embodiment according to 2 If the shock signal persists ("No" branch after step 40 ) the attenuation value is not recalculated in step 70 but it gets straight to step 80 branched, in which the attenuation value is instead increased stepwise with a third step S3 to the value 1, so that a longer-lasting shock signal is initially attenuated, however, as time progresses assumes its original volume and thus long-lasting shock signals such. As a siren, for the user are audible, however, the initial attenuation, the sudden onset of a loud signal is prevented.

The inventive method allows a reliable one and easy steaming all types of shock signals, in particular via signal processing equipment with speakers, headphones, studs or similar be forwarded to the user. By the procedure will be doing unpleasant and possibly leading to short-term or long-term hearing damage, sudden onset of loud signals steamed, being through gradual approximation of the damping factor acting on the signal the attenuation value a gentle damping achieved and unwanted Artifacts are avoided.

application areas The invention therefore extends from the shock signal elimination in electronic dictation machines or in voice transmission in telecommunications equipment to use in public address systems all kinds.

A Implementation of the inventive method takes place expedient manner in suitable signal processing devices. Such signal processing devices include appropriate data processing systems and in particular microprocessor and controller based systems or DSP (Digital Signal Processor) -Systems on which the procedure as program code in a memory is implemented. Alternative are also pure hardware solutions, z. In an FPGA or ASIC for implementing the method of the invention.

Farther are all of the invention Combinations of the claims and the description and drawings disclosed features and embodiments as belonging to the invention to watch.

P1

first signal power

P2

second signal power

P3

hold

F (t)

window function

F1 (t)

first window function

F2 (t)

second window function

s (t)

first sound signal

s (t - Δt)

second sound signal

.delta.t

Delay Time or time offset

L

Length of one Rectangular window as window functions

K

threshold

B

damping factor

A

 attenuation value

S1

first increment

S2

second increment

S3

third increment

M

adjustable lower power threshold

t

time basis the signal sample

Claims (11)

A method of attenuating sound signals comprising the steps: - To capture a first sound signal s (t) and one with respect to the first sound signal offset by a delay time Δt second sound signal s (t-Δt); - Produce a first signal power P1 from the first sound signal and a second signal power P2 from the second sound signal; - Determine the slope between the first and the second signal power; - steaming the first sound signal s (t) by a damping factor B when the slope exceeds a threshold K; in which the step of stewing further comprising the steps of: - Determine an attenuation value A out of proportion from first to second signal power; - Generate the damping factor B by gradual approach the damping factor B from a starting value to the damping value A; and - Multiply of the first sound signal with the damping factor for generation an output signal. The method of claim 1, wherein the first and second signal powers are first and second mean signal power and each have the form:

F1 (t) and F2 (t) a first and a second window function, each with the constraint

and t is the time base. Method according to one of claims 1 or 2, wherein the slope form: (P1 - P2) / Δt, and for the case that the slope exceeds the threshold K, the signal power P2 as valid Hold value P3 is stored. The method of claim 3, wherein the damping is of the form:

Method according to one of claims 3 or 4, wherein the steps to go through again and continue the process before steaming comprising: - Check if a valid Hold P3 is stored after the signal powers P1, P2 were generated; - To compare the first signal power P1 with the hold value P3 and a lower one Power threshold M, if a valid Hold value is stored; - In case P1 <P3 or P1 <M: Set the Hold value P3 to an invalid Value and the damping value A to "1"; - In the case, that neither P1 <P3 P1 <M: determine the current attenuation value A with the current value of the first signal power P1. The method of claim 5, further comprising: - Put the attenuation value A to "1" and the hold value P3 to "0", If P1 <P3 or P1 <M; or - if (P1 - P2) / Δt are the same the threshold K is. Method according to one of claims 1 to 6, wherein the generating the damping factor the steps include: - To compare the damping factor B with the attenuation value A; - in the case, that B <= A and at the same time B + S1 <A, where S1 is a first step size: increasing B by S1; - in the case, that B> A and at the same time B - S2> A, where S2 is a second one Increment is: decreasing B by S2; and - for the remaining ones Cases, that B <= A and at the same time B + S1> = A and B> A and at the same time B - S2 <= A: setting of B to the value of A. The method of claim 7, wherein the first step size S1 is smaller than the second step S2. Method according to one of claims 5 to 8, wherein in the case that a valid Hold P3 stored and neither P1 <P3 nor P1 <M, the current attenuation value A is not determined from the signal powers, but as follows: - Approaching the attenuation value A by a third step S3 to the value "1". Apparatus for carrying out the method according to one the claims 1 to 9. Computer program containing program code included in execution on a data processing system, a method according to a the claims 1 to 9 performs.