Electrical audio signal processing systems and devices – Noise or distortion suppression
Reexamination Certificate
1998-11-06
2003-04-29
Nguyen, Duc (Department: 2643)
Electrical audio signal processing systems and devices
Noise or distortion suppression
C381S094500, C381S094800
Reexamination Certificate
active
06556685
ABSTRACT:
BACKGROUND
1. Field of the Invention
The present invention is in the field of audio signal processing, more specifically in the field of audio dynamics processing utilizing variable gain elements to compress and expand audio frequency signals for noise reduction purposes.
2. State of the Art
In the 1970's, companding noise reduction systems—that is, audio systems employing complementary compression and expansion processes—became popular as a means to reduce the noise floor and increase the headroom of the analog magnetic tape recording process for higher fidelity, wider dynamic range recordings. By compressing (encoding) an input signal exhibiting a dynamic is range of 120 dB, for example, by a 2:1 ratio, the signal could then be passed through a transmission path, medium or “noisy channel” having a limited dynamic range of 60 dB. Then, by complementarily expanding (decoding) the signal at the output of the transmission path or medium by a 1:2 ratio, the 120 dB dynamic range of the original signal would theoretically be preserved. This type of process was successfully applied to other “noisy channel” transmission systems exhibiting limited dynamic range such as FM broadcasts and telecommunications. It should be understood that a transmission path, medium or channel is being used to refer to any type of audio processing system having a limited dynamic range.
In the last twenty years, the two most popular trade names associated with companding noise reduction systems have been dbx® and Dolby®. A particularly important patent in this field by David E. Blackmer, U.S. Pat. No. 3,789,143, granted Jan. 29, 1974, describes a method of dbx-type noise reduction wherein compression and expansion are complementary in time response. Principles in this patent form the basis of complementary noise reduction systems used by dbx®, Dolby®, and others. The typical design of these systems, as illustrated in the aforementioned Blackmer patent, consists of two separate circuits-an encoder circuit and a decoder circuit. Each of these circuits, in their most simple form, includes a main audio path consisting of a variable gain element having an audio input port and an audio output port, and a detector path or “sidechain” having a circuit that detects the audio signal level and creates a control signal. Typically, there are other circuits in the sidechain which are “downstream” from the detector, whose function is to shape or process the control signal—this processed control signal being used to control the gain of the variable gain element. More complex noise reduction schemes may include additional circuits in the sidechain that pre-process the audio “upstream” from the level detector.
In each of these well-known systems, there exists two sidechains-one for the encoder and one for the decoder. Typically, the encoder uses a feedback circuit topology wherein the input signal to the encoder sidechain is the audio output signal of the encoder. Conversely, the typical circuit topology of the decoder is feedforward wherein the input signal to the decoder sidechain is the audio input signal of the decoder. These sidechains are substantially identical other than the fact that they create control voltages which cause opposite reactions in their respective variable gain elements. That is, when one variable gain element is adding gain to the audio signal passing through it, the other is complementarily attenuating the audio signal by the same amount.
The main advantage of this feedback encoder and feedforward decoder arrangement is its complementary nature because the input signal that the decoder sidechain “sees” is substantially the same input signal that the encoder sidechain sees having only the noisy channel between the inputs of the two sidechains. Of course, the noisy channel will add some noise and distortion and may exhibit frequency response anomalies which would alter the “clean” signal that the encoder is acting upon, thus presenting an altered signal to the decoder, but with properly designed sidechain circuitry, the sidechain detectors react appropriately to the signal. Thus, with substantially identical encoder and decoder sidechains acting upon substantially the same signal, the decoder operates in a substantially complementary manner to the encoder.
A person unfamiliar with noise reduction systems might naturally question the need for having two separate sidechains in the system, especially when the circuitry is substantially identical in each, and each detector is intended to react to substantially the same signal. The answer is, that for traditional uses of noise reduction it is not possible to share sidechain circuitry without adding unnecessary complexity to the system. For example, in analog tape recording, encoding occurs during the recording process making use of the encoder sidechain. Because playback can occur in a different location than where the recording took place, the decoder must have its own sidechain electronics in the playback system. A similar need for the separate decoder sidechain is obviously needed for noise reduction on radio broadcasts as well. One way around this is to record the decode signal onto tape in parallel with the audio signal or, for broadcasts, to transmit the decode signal along with the audio. Both methods would add unnecessary complexity to the encoder as well as require additional bandwidth or an extra channel in the transmission path, neither of which are desirable nor required to obtain reasonable performance.
However, as is well known to those skilled in the art, one of the biggest problems with these types of noise reduction systems is that they require components of the decoder sidechain to be closely matched with corresponding components of the encoder sidechain to avoid frequency and gain errors. Also, the operating levels of the system comprising the “noisy channel” must be calibrated to the same operating levels to which the compander is calibrated to avoid mistracking and gain errors. If close matching is not maintained, this “semi-complementary” processing may produce very audible and unpleasant artifacts. This is not easy to do in manufacturing and components may drift over time and through aging and temperature cycling. Other mistracking errors occur when the noisy channel does not behave in a predictable manner. For instance, analog tape may have “dropouts” in level and various brands and types of tape, or even various levels of quality within one type of tape, will cause errors and mistracking due to inconsistent frequency response and other non-linearities.
Another problem associated with noise reduction systems, due to the minimum compression and expansion ratios-typically 2:1 or greater-required to achieve satisfactory levels of noise reduction and increased headroom through the limited dynamic range channel, is that the noise level is audibly modulated by the level of the audio signal, a phenomenon known as “breathing.” Other problems include dynamic distortion of low frequency waveforms if the level detectors are too fast and thus track these waveforms, and distortion of fast, high-level transients if the level detectors are not fast enough to react to the audio. Lastly, the sheer number of components in good noise reduction systems makes them costly to implement on a per-channel basis.
Turning the focus to the present invention, the inventors considered the problem of how to achieve a wider dynamic range from a system that many audio professionals never really considered to be the “weak link” in the audio chain. For years, in most systems, either analog tape or the broadcast channel was the limiting factor in overall system dynamic range even when employing a noise reduction system such as dbx®. The electronic noise from solid state devices such as equalizers was a minor factor and was often completely ignored. But as digital recording brought increased dynamic range to the audio signal chain, people began paying more attention to the electronic noise contributed by other equipment and began looking more closely at the equip
Johnsen Roger T.
Urry Robin M.
Harman Music Group
Nguyen Duc
Thorpe North & Western LLP
LandOfFree
Companding noise reduction system with simultaneous encode... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Companding noise reduction system with simultaneous encode..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Companding noise reduction system with simultaneous encode... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3049729