KR840002491B1 - Improvements in signal processing systems and switching system therefor - Google Patents

Abstract

The signal compressor is controlled in response to frequency and amplitude level characteristics of the applied signal. A filter excludes the high and low frequency extreme regions from controlling the compressor and attenuates these regions in the compressed signal. The filter pref. comprises a band rejecting network having a steep drop-off, with a corner frequency located to attenuate the frequency extreme regions. the band rejecting network pref. has drop-offs of 12-18 dB/octave. The network may be located in the input signal path to the compressor.

Description

Signal compressor

1-4 are schematic block diagrams illustrating various arrangements of spectral distortion networks in accordance with the present invention.

5 is a response curve diagram generally illustrating characteristics of a high frequency spectral distortion network and a compensation non-distortion network for use in a cassette tape recorder and a regenerator.

6 is a characteristic curve for the noise increase and decrease of the standard CCIR.

7-10 illustrate representative response curves of typical cassette tape recorders and regenerators.

11-16 are representative curve diagrams useful for understanding the present invention.

BACKGROUND OF THE INVENTION 1. Field of the Invention [0001] The present invention generally relates to a circuit for changing the dynamic region of a signal, that is, a compressor for compressing the dynamic region of a signal, and a circuit for changing the climb region, that is, a compressor for compressing a dynamic region of a signal and an expander for extending the dynamic region. In particular, it relates to a signal compressor.

The invention is particularly useful for processing audible signals and can also be applied to other signals.

Compressors and stretchers are typically used together for noise reduction (ie, compensator systems), where the compressor compresses before a signal is transmitted or recorded and the expander stretches after receiving or reproducing from a transmission channel. However, when the compressed signal is suitable for the end purpose, the compressor can be used alone to reduce the dynamic range of the signal, e.

Compressors are also used alone in certain products, in particular audio products which only transmit or record compressed broadcast or pre-recorded signals. And stretchers are also used alone in certain products, especially audio products that only receive or play back already compressed broadcast or pre-recorded signals.

In some products, in particular audio recording and playback products, the single device is often arranged for mode operation which can be switched to a compressor for recording signals and to an expander for playing compressed broadcast or pre-recorded signals.

It is well known that a transmission channel has a frequency dependent characteristic. Therefore, when the frequency spectrum of the received or reproduced signal is changed and the compressed signal is applied to the transmission channel, the compensation extension of the signal is degraded by the transmission channel frequency characteristic.

In a compensator type noise reduction system, compensability requires that the expander not only have inherently reverse characteristics of the compressor, but also that the transmission channel between the compressor and the expander maintains relative signal amplitude at all frequencies within the bandwidth of the compressed signal. Do.

Since the transmitted signal is received at the expander, the relative level change caused by the transmission channel is indistinguishable from the signal processed by the compressor.

Therefore, errors in the transmission channel cause the compressor to have an extended signal that is different from the input signal. This difference can be significant and audible depending on the spectral content of the signal. Errors in the transport channel for high compression and stretching ratios become more important. Typically, the maximum audible effect is not a direct effect on the ultra high frequency (VHF) or ultra high frequency (VHF) signal itself, but rather modulation on the signal between the bandwidths caused by the attenuation of the high and low frequency signals reaching the stretcher. Effect. This effect is referred to as an intermediate frequency band modulation effect for explaining the purpose of the present invention.

In wideband comparators, the amplitude ether at the control frequency becomes somewhat apparent in itself in all other parts of the spectrum. This may or may not be received.

In a sliding band type comparator as described below, if the control frequency is across the bandwidth, the error at this frequency is increased at the actual intermediate frequency. Conversely, if the control frequency is at the intermediate frequency as usual, any error across the frequency is reduced. This is the advantage of the sliding band type press.

As used herein, the "transport channel" is defined to cover all parts of the system between the compressor and the expander.

In the case of noise reduction systems for recording compressed signals on relatively narrow bandwidth media, such as low speed magnetic tapes, i.e., small cassettes, it is particularly difficult to provide an accurate compensation extension of the compressed signals.

This is due in particular to the inaccuracy of the system which provides a flat signal amplitude response at low and high frequencies. However, this problem also exists in professional compressor and expander systems. The inaccuracy may occur after the compressor and may be caused by a recorder, tape, playback device or combination thereof including a bandwidth limiting filter.

Similarly in other systems, such as FM broadcast or satellite broadcast, an error occurs after compression by a transmitter, broadcast signal transmission medium, receiver, or a combination thereof.

In practical compressor and expander systems, it is necessary to attach a sharp cut filter in front of the compressor and expander. The filter should include at least a lowpass portion and preferably include a highpass portion.

Such band-limiting filters have node frequencies at or outside the useful bandpass of the system in order not to limit the bandwidth of the system. These filters have several functions. In other words,

1) Attenuation of 19KHz pilot tones used in subcarrier components and FM broadcasts to prevent bias "birdie" ringing (small noise) and mistracking of encoder and decode noise processor circuits.

2) Attenuation of the tape recorder bias, which may be a leak in the signal circuit, to prevent miss tracking of the encoder and decoder.

3) Attenuation of radio frequencies in the encoder input signal that may be caused by an audible horn modulated product, and also attenuation of the ultrasonic components in the encoder input signal which may be caused by a bias "burdy".

4) Attenuation of ultrasonic tape noise or other transmission channel noise at the decoder inputs to prevent mistracking of encoders and decoders.

5) Signal bandwidth limiting function to enhance the compensator for the response of encoder and decoder.

For professional applications, it is desirable to use high frequency bandwidth limited filters (eg at 20-25 KHz) and also low frequency bandwidth limited filters (eg at 20 Hz).

Strictly speaking, if an ideal channel is present between the encoder and the decoder, the input filter for the decoder should not be connected, resulting in slight non-compensation in some signal states when the input filter is connected. (The encoder signal must be dependent on one filter stage and the decoder signal must be dependent on two filter stages). However, the removal of the decoder input filter is not practical because the above items 1) to 5) should be taken into consideration. Therefore, even when there is an ideal channel between the encoder output and the decoder input, a very necessary decoder input filter (for protection purposes) is present in the original uncompensated system under some signal condition.

For this reason, it is necessary to consider the expander input filter which is an integral part of the transmission channel.

It is an object of the present invention to suppress adverse effects on the compensability of piezoelectric noise reduction systems caused by errors in the transmission channel amplitude response.

In other words, it is an object of the present invention to suppress the adverse effect on the compensability of a compander type noise reduction system caused by an error in amplitude response occurring between a compressor and an expander.

It is a particular object of the present invention to suppress non-compensatory effects in ultrashort (ultra) waves that produce an audible effect at intermediate frequencies in order to reduce the mind-band modulation effect.

It is a further object of the present invention to suppress this adverse effect in systems where the bandwidth of the compressed signal exceeds the relatively flat bandwidth of the transport channel amplitude response.

It is a still further object of the present invention to suppress adverse effects in an audio tape system comprising a video cassette and an audio portion of a video disc and using small cassettes or other bandwidth constrained media.

Another object of the present invention is also to reduce non-compensatory modulation of low level intermediate frequency signals (eg, 500 Hz to 1 KHz) when an ultrahigh frequency signal (eg 10 KHz or more) is present together in an audiotape system.

Another object is to reduce the non-compensatory effect accompanied by the use of an expander input filter.

It is a still further object of the present invention to reduce the aforementioned effects in a sliding band system of the type described below.

In bandwidth constrained systems, as the frequency bandwidth of the compressed signal reaches or exceeds the bandwidth available for the recording and playback transmission channels, such systems are susceptible to errors in the recording and playback frequency response, especially in the high and low frequency bands. So in the realm. There are two aspects to this problem. In other words,

오디오장치에서 비교적 평평한 대역폭을 초과할 수 있다.1) Compressor output bandwidth has its own recording and playback response

The audio device may exceed the relatively flat bandwidth.

2) Compressor output bandwidth of equipment used to prepare pre-recorded tapes or discs may exceed the bandwidth where audio device playback response is flat.

Although errors can theoretically occur anywhere in the spectrum, the present invention is intended to suppress the effects of errors across the bandwidth of the system and, in particular, to suppress the mid-modulation effects caused by such errors.

The present invention, which is an answer to the mid-modulation problem, is very surprising in its simplicity. More specifically, the signal processed by the compressor is subject to good node frequencies within the useful passband of the system and to rapid high frequency (low frequency) attenuation with somewhat higher or lower frequencies where the transmission channel or recording and playback responses have practically errors. do. It is also desirable that the signal processed by the expander be subject to a compensation boost in order to maintain the overall flat frequency response. If the lack of an overall flat frequency response is tolerated, the present invention can only be coupled to the compressor section in the compander system.

Therefore, according to the present invention, the portion (or portions) of the transmission spectrum or the signal spectrum where the recording and playback response has an error is attenuated at a level where the attenuated portion excludes actual compression and decompression control.

In accordance with the present invention, the spectral content of the signal processed by the compressor is altered or distorted (“spectral distortion”) such that the compressor action is significantly less affected by signals above abrupt roll-off frequencies.

One preferred embodiment of the present invention is to place a suitable filter (or network) in the signal path of the compressor input (preferably the compensating filter in the signal path of the expander output).

This is desirable because it requires minimal circuitry and operates at all signal levels.

However, the equivalent configuration is to use both filters. That is, one is located in the control circuit path of the compressor and the other is located in the signal path of the compressor output.

Similarly for the expander, one filter is placed in the control circuit path and the other in the signal path of the expander input. And in the case of double-path compressors or expanders (for example, as described in US Pat. No. 3,846,719 and US Pat. No. 28,426, 426), another approach is possible, ie the two signals are full-path. A suitable filter can be placed only at the processor input to pass through and also two filter configurations, which are side circuit control circuits or equivalent, can be used, where one filter is placed in the control circuit path for the side path and the other One is placed in the signal output path to the side path. It is well known in the comparator system as described below to supply a roll off at high frequency within the compressor side of the system to reduce tape saturation and to provide compensatory boost on the expander side of the system.

See, for example, Lund Hunkteken on Jan. 22, 1978. VS-PS 3,846,719 and VS-PS 4072914, H.2, pages 63-74, issued to Mitilungen. However, the roll off is slow and is not sufficient to maintain a high level signal at high frequencies (low frequencies) in the uncertain response area of the system under the influence of the compressor carrying out the invention. It is well known in the patent of this example to place the saturation filter in several locations of the signal path, i.e., before and after the compressor and the expander, and within the compressor and the expander simultaneously included in its control circuit. Therefore, there are many possible positions for the unsaturated filter in the compressor and expander signal paths provided for the purpose of reducing the high frequency signals applied to the recording tape. In contrast, the circuit according to the present invention should be located more accurately because it affects the compressor and expander by changing the frequency spectrum of the signal that the compressor processes.

Although the present invention does not allow to mitigate the saturation effect, it also helps to prevent tape saturation to some extent. However, tape saturation is described, for example, in the audio paper, pages 20-26, published in May 1980, since the nodal frequencies drop sharply above the frequency of the region where saturation begins to occur. Is processed by.

As in the known unsaturated configuration, the present invention results in a loss of noise reduction effect in the region where the signal drops sharply as a function of frequency. However, the nodal frequencies dropping sharply toward the upper frequency stages of the system will be at relatively high frequencies in areas where the human ear is much less sensitive to noise.

For the use of the present invention at the low frequencies of the system, the ears should be less sensitive and noise free, even at reduced noise reduction, as well as inherently audible at the frequencies of a well designed system.

In the present invention, spectral distortion occurs only at both the high frequency and the Jeju frequency, because in the actual system, this frequency region is the only region in which the transmission channel amplitude response has a significant error. Another reason is that the loss of noise reduction can be tolerated at both ends due to the response of the human ear.

In another known system, a bandpass filter of 12 decibels per octave with a node frequency of about 20 Hz and 10 KHz is used in the control circuit of a wideband audio tape compensator type noise reduction system (sold under the trademark "dbxII"). .

However, such a circuit cannot achieve the results of the present invention because the control circuit is a control circuit no matter what level of signal the high level high frequency (low frequency) signal of 10 KHz or more (or 20 Hz or less) is provided without any filter in the signal path. This is because it is amplified when it exists within the pass band of. Therefore, this circuit overdrives the transmission channel when there is no high amplitude signal within the 20Hz to 10KHz band, resulting in excessive amplification of the high level high frequency (low frequency) signal (except the 20Hz to 10KHz band). Such an overshoot can be driven by the present invention.

Known video noise reduction systems (VS-PS No. 3,846,719) focus on color television subcarrier frequencies to prevent chroma signals from suppressing compressors and extenders and removing noise reduction at frequencies below the color subcarrier frequency. There is provided a variable-Q notch filter that responds to chroma levels. Thus, the filter center is within the signal bandwidth of the system and within the expected response of the system transport channel, and moreover, is not intended to overcome errors in recording and playback response, but merely forms the basis of the present invention.

In addition, the known systems are provided with lowpass and highpass filters in double-path compressors and expanders (VS-PS No. 3,846,719; Audio Engineering Magazine, October 1976, Vol. 15, No. 4, page 383). To 388).

However, these filters are for a completely different purpose, i.e. to separate the bands of the noise reduction operation in separate side paths and do not affect the upper or lower frequencies of the compressor input signal.

Likewise, the aforementioned sharp cutoff limiting filters have node frequencies deliberately located outside the useful system bandwidth because they do not affect across the upper or lower frequency bandwidth.

Up to now, it has been considered undesirable to change the frequency spectrum within the system bandwidth that is actually available. For example, in some professional recordings it was considered unthinkable to locate the node frequency of any upper band limiting filter below 20KHz. Similarly, in FM broadcasting, the 15KHz upper band limit is precisely maintained through all audio signals.

Although the present invention is not limited to use with any particular type of noise reducing compensating system and improves the operation of all types of compensators, including wideband compensators, it is particularly useful for use with sliding band noise reduction systems. Do. Examples of such systems are described in VS-PS Re 28,426 and VS-PS 3,757,254.

High frequency audible compression or stretching in this known sliding band circuit is achieved by applying a high frequency boost (for compression) or a high frequency cut (for extension) by a high pass filter having a variable lower node frequency.

As the signal level increases in the high frequency band, the filter node frequency slides upward to narrow the boost or cut band and exclude useful signals from the boost or cut band.

In the sliding band arrangement, the influence of the uncertainty of the recording and playback response causes modulation of the low level medium frequency signal when a high frequency signal in the recording and playback uncertainty region is present at the compressor input.

In a dual path sliding band device of the type as described in VS-PS Re 28,426, this effect can be suppressed by providing rapid high frequency attenuation in the sidepath of the device. Sliding band type (VS-PS Re No. 28,426) and fixed band type (VS-PS No. 3,846,719) generally meet the objectives of the present invention while this configuration provides rapid attenuation only at medium and low frequency signal levels. It is preferable.

Hereinafter, an embodiment of the present invention will be described in more detail with reference to the accompanying drawings.

1 through 4 are schematic block diagrams illustrating the various locations in which the spectral distortion network according to the invention may be located.

FIG. 1 shows the simplest and preferred embodiment, in which the spectral distortion network 2 having the low frequency or high frequency part or the network 2 having both the low frequency or high frequency part is located in the input signal path and the compressor ( 4), the output of this compressor is applied to transmission channel N.

On the reproducing side of the transmission channel N, a complimentary expander 6 operates as a reproducing signal, which is used for any spectral de-skewing to compensate for the characteristics of the compressor input path network 2. To the network 8.

This network location has the advantage, in particular, that the compressor and the extender each comprise two or more series devices such as those described in Audio Papers, pages 20-26, published in May 1980.

FIG. 2 shows an equivalent configuration for the circuit of FIG. 1, which in practice is not preferred because such a configuration is very complicated, requires additional circuitry and is expensive.

In such an equivalent configuration, the spectral distortion network 10 is arranged in the compressor control circuit, and the additional spectral distortion network 12 is arranged in the output path of the compressor signal. If any non-distorted network is used on the reproduction side, the compensating non-distorted network 14 is used in the input path of the expander signal and any spectral distortion network having the same characteristics as the network 10 of the compressor control circuit ( 10 is disposed in the control circuit of the expander 6.

The properties of the networks 4 and 12 may be slightly different from those of the network 2 and the properties for obtaining the same overall result as the network 2 are also different.

This consideration also applies to the networks 14 and 10.

Here, each of the compressor 4 and the expander 6, which contains a series of devices as described in the audio paper described above, requires a spectral distortion network 10 only within the control circuit of the first compressor device, and the first compressor The network 12 is only required in the output path for the serial device of the controller and only in the input path for the serial device of the final extension device having the network 10 in the control circuit of the expander 6. Excreted.

3 and 4 show the arrangement of the spectral distortion network for the side paths of the dual path compressor and expander. The construction of such compressors and expanders is well known to those of ordinary skill in the art, and further detailed description is considered unnecessary.

One such form (Figs. 7 and 8 of VS-PS No. 3846719) is a filter with a control limiter that gradually restricts it as a rectified and uniformized control signal as the signal level increases.

Another form (VS-PS, RE 28,426) is a sliding band highpass filter whose passband is gradually narrowed by a control signal to exclude large signal components from the filter output. The preferred node frequency for the variable filter is about 375 Hz in static state, but the high pass band gradually narrows in response to the control signal. Double-path compressors and expanders (alone or in series) may also be used in the configurations of FIGS. 1 and 2.

The spectral distortion network 18 of FIG. 3 is arranged in the input signal path of the noise reduction side path circuit 20 for the compressor 22 and the arbitrary expander 24. As shown in FIG. The equivalent configuration of FIG. 4 is illustrated by connecting the sliding band compressor 26 and the singer 27. In this equivalent configuration, the spectral distortion network 28 is arranged in the side path control circuit 30 that controls the frequency variable stage, that is, the sliding band filter circuit 32.

An additional spectral distortion network 34 is arranged at the sidepath output. Optionally, the same network is located in the expander side path. The arrangements of FIGS. 1 and 2 are preferable to the arrangements of FIGS. 3 and 4 because the former operates at all signal levels and also reduces channel overload, i.e. tape saturation, across the band. Although only a single side path is shown in FIGS. 3 and 4, multiple side paths (eg, multiple paths as described in VS-PS 3,846,719) may be used.

The side paths may also be configured such that the compressor side path has a feedback configuration and the expander side path has a circulation configuration, as in VS-PS 3,903,485, for example.

Although a configuration of FIG. 1 may be preferred where a series of double-path devices of the type described in Audio Papers, pages 20-26, published in May 1980, are used in compressors and expanders, the spectral distortion network may be It is possible to use for any final series of serial devices and expanders.

It is desirable that the spectral distortion characteristics have the following matters.

1) The falling node frequency within the transmission channel or tape recorder area, including the region of the expander bandpass filter, shall be able to be properly subordinated. That is, a slightly lower or higher frequency for the transmission channel or tape recorder's response should be uncertain or start roll off.

2) Sudden attenuation that provides a reasonably specified limit for the frequency controlling the circuit.

3) Properly defined form with regeneration to maintain the overall flat frequency response, ie attenuation to easily characterize each other during regeneration.

4) A form with an optional characteristic of the low-level noise sensitivity characteristic of the human ear-that is, as quickly as possible attenuation and deep frequency response drop without accompanying a noticeable increase in noise level when using contradictory characteristics during reproduction.

If the frequency is above about 8 KHz or below about 50 Hz, while the spectral distortion characteristic effectively reduces the noise reduction operation above or below the sharp node frequency, especially if the noise level is extremely low, as the present invention is used for tape recording comparator As a result, the increased noise becomes irrevocable given the human ear's response to low-level high and low frequency noise.

Such characteristics of the present invention have been demonstrated experimentally.

The judging method for this process can be seen in the type of CCIR noise reduction curve shown in FIG.

Here, the curve involves the low level noise sensitivity of the human ear.

The sensitivity is low at low frequencies and peaks at about 6-7 KHz and then drops sharply thereafter. Thus, there is a need for reduced psychoacoustic to maintain substantially noise reduction at frequencies above 8-10 KHz.

This is equivalent to the high frequency of the noise reduction system's ability to observe, which inherently provides a large amount of noise reduction even if low frequencies are rarely processed.

The preferred technique of the present invention can eliminate hum noise, which is the only low frequency noise that has inherent problems with this hum in cassette tape recording.

In professional noise reduction systems, a low frequency noise reduction system is provided with guarantees for unexpected hum noise, but it is generally necessary for noise reduction below the lowest hum noise components (e.g. 50 Hz) that tend to intersect. It is not necessary.

The use of spectral distortion networks, especially at the high end of the spectrum, sometimes does not eliminate or replace the entire band limiting filter, also known as multiple filters ("MPX"). The reason is mentioned above. As discussed, band limiting filters typically used for recording and playback have several functions associated with this discussion.

Therefore, even in an ideal signal channel, it is desirable to have the following in echoing and decoding.

1) Bandwidth Limit Filter

2) spectral distortion and spectral non-distortion networks.

The spectral distortion networks 2, 10, 12, 18, 28, 34 result in a sharp attenuation as shown in FIG.

The dotted line shows that the trailing high frequency (low frequency response) does not have to be as accurate as shown by the solid line.

One suitable type for the spectral distortion network (2, 10, 12, 18, 28, 34) is a lowpass (tradepass) filter with a steep slope of 18 decibels per octave and the CCIR noise decay curve (figure 6). It has a nodal frequency within the abrupt decrease and below the upper cutoff frequency of the transmission channel. A cutoff frequency of about 8 to 10 KHz or a cutoff frequency of 50 Hz at low frequencies is useful for high performance tape decks with useful but uncertain responses at about 15 KHz (30 to 60 Hz at low frequencies).

The network also takes the form of a shelf network with about 10 decibel floors as shown in FIG.

Another suitable form of network is a notch filter having a center frequency of about 16 KHz (20 Hz) and a node frequency of about 8 to 10 KHz (40 Hz) and a depth of about 10 decibels. In addition, a twin-tuned notch filter is used, particularly in professional applications, to make the second notch an octave fraction (ie 1/3 octave) above the first notch to provide a wider overall notch. .

It has been experimentally demonstrated that a depth of about 10 to 15 decibels is suitable to eliminate the mid-modulation effect, which is the most difficult case. However, depths as small as 6 decibels can substantially improve the mid-modulation effect, especially when the slope is steep, such as 18 decibels per octave.

If a flat overall response is required, the same network and compensating networks 8, 10, 14, 18, 28, 34 are used for the regeneration portion of the system.

In particular, in consumer portable small cassette radio tape recorders and playback apparatuses, in order to more clearly understand the present invention, various typical small cassette recorders have a typical high frequency response curve estimated at a sufficiently low input level to prevent tape saturation. Reference may be made to FIGS. 7 through 9. Figure 10 shows representative high frequency response curves at various input levels for another typical small cassette recorder.

It should be noted that the recorder response curve drops sharply above 10 KHz in FIG. In FIG. 8, the response curve rises at about 10 KHz, resulting in a significant peak at about 17 KHz. The response of FIG. 9 has a high frequency peak at 15 KHz and the response curve drops sharply above this frequency. For decibels to avoid saturation, the response in FIG. 10 is nearly ideal and is substantially flat except for 20 KHz. However, such a good response is rare. Therefore, the selection of the nodal frequency of the high frequency spectral distortion network of about 10 KHz is almost lacking in the response below 10 KHz, which is a typical recorder that is easy to adjust for the response curves of FIGS. 7 to 10 below the saturation level. Very preferred. Thus, the spectral distortion network is secured at most levels so that there is no level mismatch of the reproduction control signal caused by the uncertainty of the high frequency response. In addition, a nominal frequency of about 10 KHz is a good choice for this network and the frequency will be on the steep slope of the CCIR noise increase / decrease curve (figure 6), so a slight loss of noise reduction may be acceptable to the human ear.

In selecting a suitable node frequency for the spectral distortion network, the circuit designer can choose an approximate frequency different from 10 KHz, based on the system parameters. For example, higher node frequencies may be allowed for high performance transport channels. For a small cassette device whose characteristics are of the type shown in FIGS. 7 to 10, the allowable high frequency node frequency may have a range from about 8 KHz to about 11 to 12 KHz.

Also, while a filter of about 18 decibels per octave is desired to prevent high level high frequency (low frequency) signals from controlling the compressor, a small slope of about 12 decibels per octave for most signals is consistent with the purpose of the present invention. .

A much sharper slope than 18 decibels per octave makes it difficult to supply compensating non-distortion and costs much more.

The required filter attenuation rate depends in part on the sensitivity of the compressor to signals above the filter cutoff frequency. For example, considering a dual-path sliding band compressor such as that described in US-PS Re 28,426, in this apparatus, if a signal such as that shown in Figure 11 is applied to the compressor (the signal may High frequency pre-emphasis is used in the compressor control circuit, which causes this signal to have an energy spectrum as shown in FIG. This pre-emphasized signal spectrum has a peak value. After rectification, these peaks supply a DC control signal that controls the sliding band operation of the compressor.

Figure 13 illustrates the uncertain frequency response of the tape recorder channel shown for four representative small cassette tape recorders a, b, c, d. The effect on the spectrum of FIG. 12 causes four different spectrums to appear in the control circuit of the expander (decoder) and generates four different DC control signals. It is obvious here that an error occurs in the decoding process.

In this case, the preferred spectral distortion network characteristic causes the expander (decoder) to generate the same DC control signal in each case as shown in FIG. Network characteristics with node frequencies at about 10 KHz as shown in FIG. 5 are highly desirable. The network does not remove the sliding of the frequency band and it just decreases in small increments indeed. However, the sliding (compression as in the band-separation system of US-PS 3,846,719) can be recovered immediately during regeneration.

In Fig. 15, it should be noted that the basic purpose of the present invention, that is, the same peak value in the spectrum of the AC control signal is present at the commutation point of the compressor and the stretcher.

For the sliding band system of the type just mentioned, the spectral distortion network is particularly useful for suppressing the intermediate frequency modulation effect caused by the high level of the high frequency signal of the compressor which is not recovered and not applied in the expander. This rare frequency modulation effect, which occurs very rarely in a real music source, is the basic operation of a sliding band device with an incomplete signal channel, that is, the primary signal effectively controls the sliding band frequency characteristics and does not recover during playback. It relates to an operation that can cause an audible effect at high frequencies.

When the main signal is sufficiently high above the frequency of the low level intermediate frequency signal, the intermediate frequency modulation effect may be audible if the high frequency signal is intermittent, e.g. scratching impact sounds and cymbals, and not reproduced at the same level by the tape recorder. Do not. In such a case, the intermediate frequency signal is modulated even after decoding because the high frequency signal causes the sliding band frequency response to change without compensating extension in the playback decoder.

It should be noted that this effect is not fundamentally a tape saturation effect, i.e. it can be caused by incorrect biasing and inhomogeneity or gap loss and bad azimuth. However, if there is saturation in the control frequency region, the effect is obviously worse.

This problem of low-level, mid-frequency modulation effects can be better understood with reference to Figure 16, which uses a 15KHz signal at levels below -60 decibels at 0 decibels and sweeping low levels at -65 decibels. A series of probe tow curves using probe tone are shown when recording and playing back on a sliding band system.

When the significant 15KHz signal increases in amplitude from -60 decibels to -50 decibels, the compressor's output changes by about 2 decibels. Because of this significant signal coin region of the intermediate frequency, the 2 decibel change must be maintained in the recording process (a change of 10 decibels is produced in the 1 KHz region). Thus, some error on reproduction in the control frequency domain is generally increased in the middle frequency domain.

If the noise reduction system handles low frequencies, there will be a matching effect at low frequencies. Extremely low frequency rumble at the input signal to the compressor can operate the compressor circuit. If the recorder does not regenerate the rumble component, the signal modulation effect at the expander output is evident.

Claims (1)

The relative signal amplitude response in the transmission channel between the compressor and the expander is uncertain in the high and low frequency ranges of the useful bandwidth of the composite signal applied to the compressor, where the compressed signal is applied to the transmission channel, which in turn is applied to the expander. In a compressor in which the compression is controlled by a control circuit which is used together with a complex signal in the response to the frequency and amplitude characteristics of the complex signal, the compressor operation is significantly less affected by the complex signal above the abrupt roll off frequency, To significantly reduce the influence of the high and low frequency domains in controlling the control circuit so that the adverse effect on the compensation caused by the error in the amplitude response occurring therebetween is significantly reduced and to attenuate the high and low frequency domains in the compressed signal. The compressor A signal comprising a circuit arrangement comprising filter means (2 or 10, 12) in which an audible signal component is present and rapidly distorts the spectral balance of the signal with respect to frequencies in its frequency domain but is less sensitive to low level noise components. compressor.

Images