GB2419071A

GB2419071A - Method of sound signal restoration

Info

Publication number: GB2419071A
Application number: GB0422146A
Authority: GB
Inventors: Vladimir Ivanovich Eliseev
Original assignee: Individual
Current assignee: Individual
Priority date: 2004-10-06
Filing date: 2004-10-06
Publication date: 2006-04-12
Also published as: GB0422146D0

Abstract

The method of restoring a sound signal distorted at recording, reproducing, transferring and processing by various devices of a sound path is intended to make a sound signal adequate to the original signal. It is possible to restore a sound signal by subtracting a balance error out of a sound signal. The error of balance is characterized by the change in time in the process of the sound signal change. The error of balance is accumulated as the sum of amplitudes of a bipolar sound signal at every moment of time. At every moment of time only a part of the current error of balance is subtracted from a sound signal. It is possible to calculated a part (A) of the current balance error (S) 4 as A = S/T, where T is the interval of time determined by the minimal frequency of a sound signal. The offered method of restoring of a sound signal can be realized as separate applied programs as well as a program code used in programmed devices (Digital Sound Processor). The sound signal in these cases should be submitted as a digital code.

Description

l 2419071 METHOD OF SOUND SIGNAL RESTORATIC)N The invention relates to

sound technology area and may be used while recording, reproducting, transferring and processing sound signals.

The majority of used devices in the area of sound technologies brings various distortions into the sound signal.

"Method of improving the intelligibility of a sound signal" (International number publication WO 00/75920 Al) is well-known. This method improves a sound signal only in a narrow range of frequencies, that's why application of the method is limited. Besides, the method does not restore a sound signal, though the signal becomes more appreciable at a level of noises.

"Method and circuit for distortion compensation and output control" (International number publication WO 02/27921 Al) based on using negative feedback is well-known. "Method of sound channel upgrading" (Patent GB2369278), based on extension of an infrasonic of a range of devices of sound channel to OHz and use of nonlinear negative feedback on a constant current, is well-known. These methods allow to decrease some distortions of those devices of a sound channel, to which these methods can be applied. But these methods can not restore a signal, which was deformed by other devices of a sound path.

The offered method of restoration of a sound signal while recording, reproducing, transferring and processing sound signals has no disadvantages typical of other methods.

Technical result of the declared invention is the restoration of a sound signal while recording, reproducing, transferring and processing sound signals, that allows to make this signal adequate to the original signal.

The r esult is reached by subtraction of the balance error from a sound signal.

The error of balance is characterized by its change in time in the process of a sound signal change. The error of balance is accumulated as the sum of amplitudes of a bipolar sound signal at each moment of time. At every moment of time only a part of the current balance error is subtracted from a sound signal. It is possible to calculate a part (A) of the current balance error (S) as A = S/T, where T is the interval of time determined by the minimal frequency of a sound signal.

The method of restoration of a sound signal implies the following way.

As shown in Fig I the original sound signal has the complex asymmetric form and it is possible to single out a spectrum of separate frequency components from this signal. Each frequency component is a fading oscillatory process as illustrated in Fig 2. The fading oscillatory process is characterized by the fact that it is balanced. The total area of positive half-waves of a signal is equal to the total area of negative half-waves of a signal on the interval of time aiming at infinity as shown in Fig 3. This condition is common for both each separate frequency component of a sound signal and a sound signal in the whole. In the process of a sound signal passing through the elements of a sound path the form of a signal is distorted, that inevitably results in infringement of total balance of positive and negative half-waves of a signal. It is possible to restore a sound signal by subtracting the error of balance out of a sound signal. The error of balance is characterized by the change in time in the process of the sound signal change. The error of balance is accumulated as the sum of amplitudes of a bipolar sound signal at every moment of time. At every moment of time only a part of the current error of balance is subtracted from a sound signal. It is possible to calculated part (A) of the current balance error (S) as A = S/T, where T is the interval of time determined by the minimal frequency of a sound signal.

The offered method of restoration of a sound signal can be realized as separate applied programs, and also as a program code used in programmed devices (Digital Sound Processor). The sound signal in these cases should be submitted as a digital code.

For a sound signal in a digital code PCM (pulse code modulation) the method can be implemented by the following actions: - the accumulator of a balance error, is initialized as S=O; - interval TO as TO=Fd/FO is calculated, i.e. Ed is the frequency of digitization of a code PCM, FO is the minimal frequency of a sound signal; - the cycle of reading of a code of a signal begins; - the next code of amplitude (Am) of a signal is read; - the part of an error of balance, such as Am = Am - (S + Am) / TO is subtracted from amplitude (Am); - the corrected code of amplitude (Am) of a signal is recorded; - the error of balance, such as S = S + Am is accumulated; - the cycle is finished upon termination of a code of a signal.

The minimal frequency for different sound signals is not constant and changes in a range, approximately from I to 50 Hz. It is necessary to choose approximately from 7 to 8Hz for effective work of a method for a sound signal frequency FO. it

Claims

I The method of restoring a sound signal distorted at recording, reproducing, transferring or processing a sound signal is realized by restoring a sound signal balance.
2 Balance restoration of a sound signal in claim 1 is realized by subtracting a balance error out of a sound signal.
3 The balance error in claim 2 is accumulated in the process of changing a signal as the sum of amplitudes of a bipolar sound signal at every moment of time.
4 In claim 3 at every moment of time only a part of the current error of balance is subtracted from a sound signal.

It is possible to calculate the part (A) of the current balance error (S) in claim 4 as A = S/T, where T is the interval of time determined by the minimal frequency of a sound signal.