Data processing: speech signal processing – linguistics – language – Speech signal processing – For storage or transmission
Reexamination Certificate
1998-07-29
2001-05-15
Dorvil, Richemond (Department: 2641)
Data processing: speech signal processing, linguistics, language
Speech signal processing
For storage or transmission
C704S270000
Reexamination Certificate
active
06233548
ABSTRACT:
BACKGROUND OF THE INVENTION
The present invention pertains to a method and apparatus for performing level compensation for an input signal. More particularly, the present invention pertains to a method and apparatus that adjusts the output level of an input signal to compensate for level changes made to the input signal by a signal processor.
A category of devices referred to as ‘signal processors’ are known in the art. These processors can be used to modify an input signal, such as music (or other audio sound) that has been converted to electronic analog or digital signals. Common audio processors include tonal modifications (e.g., an equalizer or EQ, vocoders, distortion effects, chorus effects, flanger effects, ring modulators, wah-wah effects), dynamic modifications (compression, expansion, tremelo, vibrato,etc.), reverberation, delay, many others, and combinations of these. Other processors are known in the art such as common video processors (e.g., color filters) and other processors used in a variety of fields. These processors all have a potential “side effect” of changing the original input signal's absolute peak level in relation to other signals or devices present. This change in peak level is usually not desired, is often not readily apparent to an operator, and can easily affect the judgement of an operator of the device, often without the operator knowing it.
In the audio field, there are three common reasons to “equalize” a sound: 1) to get rid of a problem noise (e.g., cut out an annoying air-conditioner buzz), 2) to tailor a pleasant musical sound (e.g., make a singer's voice sound pleasantly ‘husky’ by adding a certain low range), and 3) to create an effect (e.g., make a singer sound as if singing through a 1920's megaphone, as in the Beatles' “Yellow Submarine”).
When using an equalizer to boost or cut a portion of the sound for any purpose, the volume of the sound that is equalized actually gets louder (when boosted) or quieter (when cut). It is also understood in the art of sound engineering that it is important to allow a comparison between the original sound and the equalized sound as part of the process of using an equalizer (or other sound processor). This is accomplished with a “Bypass” switch, which chooses whether the sound goes through the equalizer circuit, or bypasses it.
Psycho-acoustic Concerns and Results: The study of how a person hears is commonly referred to as psychoacoustics. A well-known phenomenon in this field was first codified by Fletcher & Munson, and is most commonly displayed as a graph called the “Fletcher-Munson curves” or “equal loudness contours.” The data shows that the human ear-brain system interprets loudness as a function of frequency with two particular effects:
1—A person hears unevenly at different frequencies (at the same volume level, low and high frequency sounds seem quieter than middle frequencies), and
2—As the entire sound gets louder and louder, one will hear more and more evenly. With very loud sounds, the above effect becomes small or disappears. Thus, as one turns up the volume, a person seems to hear more bass and treble, which is why some people listen to music so loudly.
A human being typically is capable of hearing sound in the frequency range that is made by the human voice (from about 100 Hz to 4 kHz) more easily than sounds that are at higher or lower frequencies. A low bass guitar note and a high violin note both need to be played with a lot more energy than a voice, if one wants these instruments to sound as loud as a voice.
The particular effects outlined above interact with one another. As the sound gets louder, frequency differences affect hearing less and less; at very loud levels, it takes about the same amount of energy for the bass, the voice and the violin to sound as loud as each other.
This is commonly experienced when using a stereo. With the volume at a normal listening level, a compact disc will have an acceptable sound. If the volume is turned down a lot, the music will suddenly seem as if there is not enough bass, and also not enough cymbals or other high frequency sounds (the Fletcher-Munson curves describe how much less bass and high frequencies). Stereo systems are often supplied with a “Loudness” control (sometimes an on/off switch, sometimes a variable knob) to try to compensate for this (with very limited success). The Fletcher-Munson curves are averages of data empirically derived from many people, thus they are approximations for any given individual's experience.
Audio engineers deal with this phenomenon on a regular basis. The following are two examples of how this issue is addressed in the art.
A first example involves an instrument that produces only one sound, a tom-tom on a drum set. Given a recording of the tom-tom, the audio engineer is looking to make it sound better using an equalizer. One may hear a pleasant sound near the drum's fundamental resonance, for example between 400-800 Hz. There is also a very unpleasant sound just above this frequency range. An engineer can use a low pass equalizer (which cuts out high frequency signals) that has a control to select frequency (this lets one choose the point above which sound is cut). This should allow the engineer to cut out the unpleasant area and then fine tune the lower frequencies to find a spot where the drum sounds the best.
First, an engineer may cut out the unpleasant frequencies above the nice range. But now, the sound is much too quiet, so the volume is turned up significantly. Next, the engineer fine tunes for the nicest sound. The drum itself has varying degrees of loudness at each frequency, and the nicest place to set the control may be in a frequency range where the drum is relatively quiet. If the nicest part of the drum's sound is very quiet, it may go unnoticed (again because of the Fletcher-Munson effect). From experience or training, the engineer will (again) turn the volume up significantly, to help hear the best sound. Next, a bypass switch is used to see if the equalized signal is an improvement over the original (unequalized) sound. The much louder unequalized sound makes it impossible to tell what the engineer has accomplished (the louder one almost always sounds fuller and better just because it is louder, as described by the Fletcher-Munson curves) and can hurt ones ears because the sound is so much louder (the equalized sound had to be turned up a lot so that one could hear it). This increased volume could also damage the speakers because the sound is so much louder.
In a second example, the recorded track of a singer is mixed with the rest of the band (drums, bass, guitar, piano, etc.). As stated above, the music sounds “bassier” when the volume is turned up (or less “bassy” when the volume is turned down), as described by the Fletcher-Munson curves. As more bass is added to a sound with an equalizer, the total volume is being increased.
If the recording of the voice sounds “thin” because the microphone was poorly placed, the engineer may try to compensate by turning up the bass equalizer. The voice sound improves, but the extra bass has made the voice louder than the rest of the band, so it is turned down to an appropriate level. The result is that, the voice doesn't have enough bass anymore (due to the Fletcher-Munson effect). So the process is repeated a few times, until a satisfactory result is achieved.
A known technique for solving the problems addressed above is set forth in the following complex procedure.
1—Select a channel to be equalized (“EQ Ch.”).
2—Set up a separate channel that duplicates the unequalized original input (“UNeq Ch”).
3—Listen only to the EQ Ch. (i.e. ‘SOLO’ the EQ Ch.—listening only to that channel, turning off all other channels).
4—Adjust the equalizer settings on the EQ Ch. until satisfactory.
5—Solo the EQ Ch. and the UNeq Ch. back and forth to:
a—match the EQ Ch.'s output level to the UNeq Ch.'s output.
b—adjust the equalizer settings on the EQ Ch. if necessary.
6—In the mix (with the other sound channels tu
Osmand John H.
Schwartz Stephen R
Dorvil Richemond
Kenyon & Kenyon
Schwartz Stephen R.
LandOfFree
Method and apparatus for performing level compensation for... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method and apparatus for performing level compensation for..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method and apparatus for performing level compensation for... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2545201