MEMO - BRYSTON CUSTOMER FEEDBACK SUBJECT Nice Letter from Mr. Burke Strickland on SP2
April 13th 2007 Hi James, Just wanted you to know that I accepted delivery on my new SP2 (serial number 000371) from Viet Tran at Sound World of Houston yesterday evening.
Once I got it home, it didn't take very long to get it set up and playing music as compared with the time it had taken to set up the demo unit a few weeks ago. (Practice makes perfect. :>) Of course, your previous advice on the Test function helped streamline the process, too.) Viet is also kindly letting me try out a 9B power amp, and of course the combo is wonderful. I can see (actually, that would be hear) why Stereophile gave them both an "A" rating! It is especially revelatory to have the same kind of ultra-highquality amplification all the way around when playing SACD or DVD-Audio. However, I think with my Maggie 3.6's I'm ultimately going to want to go with a 6B or some 7Bs up front. But I digress from the main points I wanted to make. An important reason Im writing you today is to let you know that in addition to having a great product line, you also have a great local dealer in Houston, Viet Tran and his Sound World of Houston store. Besides being a very personable and trustworthy businessman, his willingness to work with me on this upgrade decision process played a big part in my making the right choice to go with Bryston this time around. In contrast, I cannot stand to even be in the same room with the local Classe dealer, and "support" from Integra Research (the SP2 replaces an Integra Research RDC-7) is nonexistent, so they weren't even in the running despite very favorable reviews on their current products. Which brings up the final, but no less important, point, which is that your timely and helpful responses to my emails these past few weeks clinched my feeling that I am dealing with a company that is worthy of my support as a customer. I have really appreciated the information you have shared, as well as the quick and candid way you shared it, and I look forward to a long and positive business relationship with Bryston. Many thanks, Burke Strickland

T HE A IM

OF THE I NTEGRA

R ESEARCH RDA-7

Our goals for the RDA-7 were ambitious: to produce an amplifier suitable for either home theater or music system applications with sufficient power and dynamics for even the most demanding system installation. The RDA-7 design also required the anticipation of future standards in home theater multi-channel sound reproduction. Superb sound quality, of course, was the overriding goal. We are bringing something new to this market niche by providing an amplifier with sound quality that approaches the level of a typical high-end design. With these lofty goals in mind, the design objectives for the RDA-7 were as follows: 1. To provide a multichannel amplifier that offers superb sound quality in either home theater or traditional music reproduction. 2. To provide a robust design under the most demanding home theater application by ensuring full THX Ultra certification. (With intelligent protection circuitry to guarantee trouble-free performance under adverse conditions.) 3. To offer elegant engineering in both the packaging and design of this world-class product offering.
BALANCED AUDIO TECHNOLOGY
4. To offer a progressive design that meets the THX Surround EX standard for the reproduction of the full home theater experience. (Indeed, the RDA-7 blazes new ground by offering seven channels of high-power output in one convenient package.) 5. To provide a balanced design that is consistent with one of the founding principles of Integra Research and Balanced Audio Technology. 6. To offer an amplifier that provides true synergistic performance with the full Integra Research system. Of course, in this elegant yet very tight package featuring so many channels, one cannot fully achieve the sound quality of a cost-no-object high-end amplifier. Nonetheless, the RDA-7 substantially bridges the gap. THX Ultra certification was also critical to ensuring the robust performance of the RDA-7 under even the most demanding home theater reproduction. Stable high-power output is a critical factor in this regard, and here the RDA-7 does not disappoint. The RDA-7 offers 150 watts minimum into eight ohms across a large number of channelsseven, to be exact, in one chassis. This, to our knowledge, is an industry first.

OVERALL TOPOLOGY

The main focus of the RDA-7 design was to create a superb sounding amplifier. This was accomplished by minimizing the number of gain stages. The fewer number of gain stages employed, the more truthful and musical the final sound. This is why the RDA-7 employs a robust threestage design. Its driver stage can even be considered a single gain stage. In most amplifiers, there is a voltage gain stage and an output stage, and then several gain stages in between. At the very least, there is a buffer stage between the driver and output stage. This standard approach complicates an amplifier circuit and invariably degrades the sound. In the RDA-7, we were able to eliminate these buffer stages. This was achieved by utilizing a high-current, high-power voltage gain stage. While quite uncommon in the audio industry, this has been a hallmark of Balanced Audio Technology amplifier designgain stages that run at unusually high power and current levels. This provides several advantages: 1. We can eliminate the buffering stages altogether, without any problem with frequency response, bandwidth, or slew rate. 2. The circuit becomes more linear when it is run at high currents. This occurs in most solid state devices as their operating currents increase. 3. We were able to manage power dissipation effectively by carefully selecting the devices used for this novel circuit topology. In the RDA-7, a synergistic combination of devices is used to achieve the elegant simplicity of this design. For example, high-power-dissipation
TO-220 cased parts are used in the driver stage. They work at a very high quiescent current of 50mA, and that creates a low impedance point that directly interfaces with the output stage without any problem. This is an important element of the RDA-7. For the driver stage of the RDA-7, the design objective was to provide a very linear circuit topology. With this as the goal, we wouldnt need a large amount of overall negative feedback to provide a linear response. For this reason, each element in the driver as well as the output stage of the RDA-7 had to be quite linear. Since both gain blocks were designed from the outset to meet this objective, only a small amount of overall feedback was required. In fine-tuning the design of the RDA-7, we had several elements responsible for setting the overall gain, and we tried many different combinations until the best one presented itself. We used our finest solid-state and tube designs as a reference, and the result for the RDA-7 is an outstanding wide-bandwidth sound that compares quite well with these stellar benchmarks. Great care was also taken to optimize the bias current for the output stage. The bias current does not have any simple rule for its setting. A balance between power dissipation, linearity, and output impedance must be struck, because they are all interrelated. We identified a setting that yields a wonderful balance in this regard, one that is conservative with respect to heat dissipation, yet delivers a fast, open and clean sonic signature. The RDA-7 architecture is completely based on bipolar transistors and achieves a bandwidth of over 200 kHz. We decided that bipolar technology was the most direct and efficient way of achieving the robust performance required for home theater application. There was also the question of how much power the channels could actually deliver. Every channel delivers 170 to 180 watts during testing, compared to the RDA-7s conservative 150 watt per channel specification. Indeed, the output power rating could be increased to a higher level and still be safe. This is in part due to the proprietary protection circuitry, which is completely outside of the amplifiers signal path. Needless to say, this is quite a versatile amplifier. The RDA-7 is a widebandwidth design that uses a straightforward signal path with minimal feedback, yet provides seven channels of amplification with over 1000 watts of total output power. With the RDA-7, you arent limited to using it only for home theater applications. You can also use the RDA-7 for simply outstanding music reproduction.

APOGEE LOW-JITTER CLOCK

Onkyo has licensed the Apogee Low-Jitter Clock for use in Integra Research. For over a decade, the name Apogee has meant the highest quality digital audio equipment available, and Apogee equipment built around its renowned Low Jitter Clock has become standard in the recording industry. The Apogee Low Jitter Clock is the very heart of the equipment that professional recording engineers use to make the master recordings for the music that you listen to. Now, Onkyo and Apogee are giving you the opportunity to use the same technology when you play these recordings back at home.
A BOUT A POGEE E LECTRONICS CORPORATION
From the very beginnings of digital audio, there were some who strongly held the belief that, despite the lack of noise and other benefits, there was a certain something about digital recordings that wasnt quite right. Two opposing groups emerged: those who believed that digital audio was fine, and those who thought it was inferior to traditional analog techniques. As is often the case, the truth lay somewhere in between. Apogee Electronics was established by in 1985 by Betty Bennett and Bruce Jackson, both veterans of the professional audio industry. They formed the company for one purpose: to deliver the highest possible quality from digital audio. They both were aware that the digital technology of the mid-Eighties was not perfect, but also knew that it was not inherently inferior to analog. Instead, they recognized that digital audio introduced a whole new set of problemsbut were convinced that they were problems that could be solved. Through intensive research, Apogee engineers have proceeded to identify, tackle, and overcome the problems associated with the digital conversion process. Apogees first products were retrofit filters designed to replace the standard filters in a wide range of digital audio recorders. These were extremely well received, and ultimately there were few major artists or studios who did not request Apogee-equipped digital multi-tracks. In 1988, Apogees filters received the prestigious TEC Award for technical excellence in the audio industry. Filters were just the beginning: several other parts of the conversion process also needed to be addressed, and this prompted Apogees decision to build completed converters. This required years of development work, as each step in the conversion process required entirely new solutions to new problems. Special power supply modules were developed to ensure adequate isolation between digital and analog power circuits. Quality analog circuitry was as important as digitala factor often neglected in early designs. Clock jitter was also shown to be a major factor in conversion degradation: the Low-Jitter Clock module required a major development effort on its own.

Apogees AD-500 A/D and DA-1000 A/D provided true 18-bit resolution and become landmark products in the history of digital audio (the AD-500 received a TEC Award in 1992). Apogee was already at work developing its next generation of 20-bit resolution converters, and in anticipation of the move to 20-bit technology in the production environment, Apogee began working on a way to translate the added detail and clarity of a 20-bit recording into the 16-bit domain of Compact Disc while minimizing any loss of quality. The usual way of achieving this was by adding dither noisebut the penalty was a higher noise floor and noticeable hiss. UV-22 encodes the detail of 20-bit data flawlessly into 16-bit data using a unique proprietary algorithm instead of dither. It is used on over 80% of hit albums released in the United States and is virtually the industry standard for word-length reduction. Apogees is continually working on new products, and its latest converters offer 24 bits of resolution, and sampling frequencies up to 192 kHz.

J ITTER : E FFECTS

S OLUTIONS
The very heart of a digital recording and playback systems is the system clock, and Apogee quickly identified clock jitter as being a major factor in conversion degradation and the primary cause of an out-of-focus sound stage. Jitter is defined as any irregularity in the timing of samples being received at a D/A converters input or in the sample timing of an A/D converter, and it is the enemy of all digital systems. In the analog to digital conversion process, the timing regularity of the sampling processin other words, exactly when each voltage sample is digitizedis controlled by a crystal clock when the converter is in internal sync mode, and/or a phase-locked loop or PLL (which essentially causes the clock to lock to the timing provided by an external signal) when the converter system is being clocked from an external sync source. The signal from the clock tells an A/D converter when to take each sample measurement, and a D/A converter when to convert the digital value back to analog. Therefore, the accuracy of the timing between each sample determines how faithfully your digital recording represents the analog signal, and how faithfully it is played back. In a digital audio recording system soundwaves are converted to a varying analog voltage, and this voltage is then converted into a sequence of numbers that represent the sound waveform. These numbers have to be big enough to accurately capture the smallest and biggest details in sounds, and also need to be changed fast enough so our ear is not aware of them stepping by. You are probably aware that film animations consist of a sequence of individual drawings that change fast enough to give the illusion of motion. If we slow the sequence of drawings down, the image starts to flicker like the old movies and motion becomes jerky. To fool our

eyes into seeing fluid motion, the images need to change from one to the next at around 25 per second. The frozen visual images of individual movie frames are analogous to the individual numbers of digital audio. Our ear doesnt get fooled into thinking that these numbers sound real until they change at around 32,000 times a second. The individual numbers are called samples and represent audio in narrow slivers of time. The rate at which these frozen slices of audio change is called the sample rate. The movie film analogy is helpful in gaining an understanding of the effects of jitter. Imagine filming a horse running a constant speed with a film camera operating at 24 frames per second. If the time between each shot is the same, the motion of the horse looks fluid and correct. Each shot would have the horse advance by the same distance, say two feet. If the time between each shot is not the same, the motion of the horse would look jittery and unnaturalthe image would not be correct. In one shot the horse might have advanced two feet, in the next it could have advanced four feet, and in the next one foot.
The time difference between shots causes the motion of the horse to look jittery. Although it seems unnatural to our eyes,the effect of the jitter on our ears is even more noticeable. 10
This is an exaggerated example, but it is a good illustration of what jitter does. In the case of digital audio, jitter is the measure of how closely the samples arrive to the ideal time for each sample. Although the digital audio samples are changing very quickly, it is still very important that the spaces of time between them are the same. Jitter messes up the timingbut on a subtle microscopic level, and, unfortunately, it turns out that the human ear and brain are very sensitive to these tiny timing irregularities, and jitter of just a few nanoseconds can compromise digital audio performance. In terms of the listening experience, Jitter interferes with the brains ability to perceive a stereo soundstageto gain an impression of the relative positions of instruments and effects when a recording is played back. Jitter smears the audio soundstage; the sense of width and depth is skewed, narrowed or even lost altogetherbecause the arrival time of individual samples is being skewedsmeared across time. Louder, high frequency sounds are the first to be affected by jitter. These high-pitched sounds carry the fine sound detail that contains subtle cues to help us locate the source of sounds. Tiny echoes and decaying reverberation tails are blurred, and high harmonics are less sharply localized. It takes a lot of jitter to affect mid range sounds. Even more jitter is needed to affect the bass which causes the high harmonics of these sounds to be affected first. Jitter also comes in different flavors. Some is totally random in nature and some contains energy clumped at specific frequencies.

Originally recorded waveform.
Accurately sampled with the Apogee Low-Jitter Clock.
Produces a digital disc waveform that is exactly the same as the original recording.
Inaccurately sampled with a jittery clock.
Produces a digital disc waveform that is substantially distorted from the original recording, creating unwanted noise and distortion.
Apogees patented Low-Jitter Clock audibly cleans up digital signals and delivers extremely low jitter performance. The Apogee Low-Jitter Clock takes in erratic or jittery timing signals and puts out a family of cleansed, ultra pure timing signals. This allows optimum A/D and D/A conversion for uncompromising sonic results, and, in fact, the coveted Apogee sound is due in large part to the accuracy and low jitter characteristics of its clock circuitr y. Recording industry professionals have long acknowledged this fact, and in professional recording studios it is almost standard practice to use an Apogee converter as the master clock for the entire studio, and distribute the benefits of the Apogee clock to all of the other digital gear. The below printout is taken from a high-speed digital oscilloscope. The bottom trace shows the timing jitter in an optical (consumer) digital audio link. The top trace shows the recovered timing after the jittery input has passed through the Apogee Low-Jitter Clock. This jitter cleaning module is taking an already low 4-5 nano-seconds peak to peak input jitter and reducing it to below the oscilloscopes 100 pico-seconds resolution. This cleansed output is better than 30 pico-seconds RMS and represents an improvement of over 50 times.
Printout from a High-Speed Digital Oscilloscope
WIDE RANGE AMPLIFIER TECHNOLOGY
To accurately reproduce source signals from equipment using new digital formats such as DVD-Audio, an amplifier is required to have a flat frequency response that extends to 100 kHz. While achieving this bandwidth figure presents no particular problem on its own, doing so while maintaining excellent characteristics for other critical performance criteria is a serious design challenge, particularly with regard to the dynamic signal-to-noise ratio. Integra Researchs solution to this issue is a set of three refinements in amplifier circuit design that we collectively refer to as WRAT (Wide Range Amplifier Technology). Our first goal was to reduce the amount of negative feedback (NFB) to an absolute minimum and thereby eliminate the effects of counter electromagnetic force from the speakers. The second goal was to eliminate fluctuations in ground potential through careful layout of the components and design of the printed circuit boards to avoid open earth loops in the circuit. The third goal was to improve the ability of the amplifier to supply high levels of instantaneous current to the load.

L OW NEGATIVE FEEDBACK DESIGN
In a typical audio amplifier, the amplifier output signal is partially returned to the amplifier input via a negative feedback (NFB) circuit that has a gain characteristic. This common technique is employed to as a means of improving the static frequency response and distortion ratio characteristics. However, the load that an audio amplifier drives is a loudspeaker that will typically include components such as a vibration plate, voice coil, magnets, choke coils and capacitors. When a speaker is driven by the amplifier, a counter-electromagnetic force is generated in the speaker, and this is also returned to the amplifier input via the speaker cable, output terminal and the NFB circuit. Accordingly, if the amount NFB is large (i.e. if the NFB circuit has high gain), the amount of counter-electromagnetic force that reaches the amplifier from the speaker will be correspondingly large, and this will disrupt the drive mechanism of the amplifier, and adversely influence its transient response characteristics. To avoid this situation, our approach was to improve the open-loop frequency response (i.e. before application of NFB) of the amplifier, and reduce the amount of NFB to a level at which the influence of the counter electromagnetic force from the speakers on the performance of the amplifier becomes negligible. Our WRAT amplifiers have an extremely broad open-loop frequency characteristic, and extremely low distortion right across the frequency range. As a result, with just a minimal amount of NFB they can accommodate input from the new digital formats while maintaining superb dynamic characteristics. In addition to being unaffected by speaker reaction, they also provide improved speaker drive and braking, and reproduce audio with an extremely high degree of clarity and accuracy.

Circuit Diagram of NFB

E LIMINATION
G ROUND P OTENTIAL FLUCTUATIONS
Signal amplitudes in an audio amplifier are referenced to ground. In other words, the signal amplitude at any point in the circuit is the difference in potential between the signal level and its associated ground level. Accordingly, if the ground potential is fluctuating, even minutely, due to the presence of electrical noise, the amplifier will not have a stable reference point, and will not be able to reproduce the signal from the input source accurately. This can lead to output of objectionable noise within the audible spectrum, and adversely influences musical dynamics, accuracy of sound imaging, musicality, sound stage and depth. It is obvious that a stable ground potential is important for an amplifier, and conventional amplifiers employ a variety of techniques to prevent external factors from influencing the ground potential in the amplifier. Our engineers have taken this an important step further by focusing on the fact that when amplifiers operate with a transient signal, they actually cause the ground potential to fluctuate themselves. Conventional amplifiers employ an open-loop earth configuration in which the physical arrangement of the electronic components and the printed circuit board layout are such that the ground for the circuit is spread over a large area. When such an amplifier is operating with a transient signal, there will be many places where noise can be superimposed onto the ground potential and cause it to fluctuate. In addition, fluctuations in the ground potential in one channel also influences the ground potential of other channels. With Integra Researchs WRAT amplifiers design philosophy, the component and PCB layout are carefully designed so that the earth circuit branches form closed loops, and are all connected

to a common point. This prevents spurious current flow between different ground points, and effectively eliminates fluctuations in ground potential.
H IGH I NSTANTANEOUS CURRENT CAPABILITY ( HICC )
Typical audio signals are not simple varying voltages. In general, they are extremely complex, and consist of a wide range of frequencies and amplitude levels. When an amplifier outputs such a signal to a loudspeaker, the mechanical and electrical components in the loudspeaker, such as the diaphragm and crossover networks, accumulate energy. When the direction of flow of this accumulated energy is out of sync with the direction of energy flow from the amplifier, the amplifier must be able to supply large amounts of instantaneous current to cancel it out. Further, at times during music reproduction, the impedance of a speaker can become very low for brief periods, and the amplifier will accordingly be required to supply a large amount of current. In order to provide a solid level of speaker drive under the various conditions that arise during the reproduction of a musical signal without risking destruction of the speaker, it is essential that an amplifier have the capacity to supply large amounts of current for brief periods. The instantaneous current capability of Onkyos WRAT amplifiers is far better than that of conventional amplifiers with comparable power output, and also offer better speaker control capacity.

VECTOR L INEAR CONVERTER

VECTOR LINEAR CONVERTER (VLC)
Designers of digital audio equipment devote much of their attention to obtaining good frequency characteristics. In reality, however, the frequency characteristic of a system does not have a defined temporal space, and only gives an indication of the response characteristics of the equipment with respect to the temporal dynamism of the original sound. In other words, a digital disc player can have excellent frequency characteristics, but be inferior in terms of audio quality due to temporal distortion caused by phase shift (among other factors). The result is inconsistency between the relative timing of different frequencies at the output. The philosophy behind VLC is to focus attention on maintaining the temporal dynamism of the original sound, in addition to providing good frequency characteristics, and to completely eliminate the sonic unevenness that is caused by the temporal distortion inherent with conventional D/A conversion methods.
CONVENTIONAL D/A C ONVERSION M ETHODS
Conventional ladder-type and 1-bit (modulation) D/A converters discretely convert the digital value at each sampling period to an analog quantity, and produce an output waveform that includes a large quantity of highfrequency image spectra at multiples of the sampling frequency. In this state, the waveform is considerably different from the original sound waveform, so a high-order analog filter is essential at a subsequent stage to remove these high-frequency components and obtain a smooth analog waveform. High-order analog filters cause deterioration in the phase characteristics of the reproduced signal, and result in temporal jitter that is not present in the original waveform. This results in an appreciable drop in the quality of the reproduced audio signal.

T HE VLC D/A C ONVERSION CIRCUIT
The VLC technique employs a unique D/A conversion circuit to overcome this problem. Unlike conventional methods, which simply convert the sampled data into discrete analog values, the VLC circuit converts the data between the sampling points, and joins the discrete sampling points, with analog vectors in real-time to produce a smooth output waveform. Other manufacturers may use powerful digital processing circuits to calculate interpolation values between sampling points, and then pass this data through a high-speed D/A converter to produce the analog output signal. However, this is just equivalent to increasing the sampling frequency, and the output is still in discrete analog steps. In addition, the increase in noise caused by high-speed digital processing has an adverse effect on sound quality. VLC requires neither complex digital processing or high-speed D/A conversion circuits, and does not increase the sampling frequency. It uses an extremely simple analog circuit (Fig. 3) to interpolate a linear
vector between samples, and produces a continuous rather than discrete (staircase) analog output waveform. The D/A conversion time is kept as short as possible in order to faithfully trace the original sound waveform, and reproduce the musical expression of the original sound without omission.
Fig. 1 Conventional D/A conversion Fig. 2 VLC D/A conversion
Fig. 1 shows the staircase output of a conventional D/A converter superimposed with the same signal after it has passed through a high-order, low-pass filter. The staircase waveform is produced by discrete analog conversion and includes an appreciable highfrequency component that must be removed by a high-order, low-pass filter to obtain a smooth analog output waveform. As a result,there is appreciable lag between the filtered waveform and the original staircase waveform due to the phase characteristics of the analog filter, and overshoot distortion also occurs.

Fig. 2 shows an example of VLC D/A conversion. Analog vectors corresponding to the continuous variation between the sampling values are formed at the initial stage of the D/A conversion to produce a smooth analog signal. VLC directly provides a smooth analog signal that has an extremely small high-frequency component. As a result,the degree of filtering required is significantly reduced,and the output waveform has no significant temporal distortion or overshoot.

P RINCIPLE

OPERATION
The digital input signal (sampling frequency: 44.1 kHz) is passed through an over-sampling filter and over-sampled at eight times the original sampling frequency. This signal is DGA in the circuit diagram below. The DGA signal is the digital input for the multi-bit D/A converter DAC-A. The digital signal DGB is formed one sampling period after DGA, and this signal is the digital input to another multi-bit D/A converter, DAC-B. The analog outputs of these two D/A converters are VA and VB respectively.
Fig. 3 VLC circuit diagram
The VLC circuit converts the difference between the outputs of these two multi-bit D/A converters (VA and VB) to a current. It uses this current to charge a capacitor, and form a voltage vector, and then superimposes this vector onto the VB voltage level to smoothly join the adjacent sample levels and form a smooth analog waveform. The current that flows in resistor R6 in the circuit diagram is proportional to the difference in voltage levels between sampling points. This current charges capacitor C2, and the resulting voltage output is a linear vector that joins VA and VB (the R6/C2 circuit performs a simple integration operation). This vector is passed through the buffer OP-2 and continuously added to the VB level. This operation is repeated at each sampling point to directly obtain the smooth analog output VC. Because this circuit converts the DC levels (between samples) to an analog vector, it also makes it easy to obtain a broad frequency characteristic.
OPTIMUM GAIN VOLUME C IRCUITRY
OPTIMUM GAIN VOLUME CIRCUITRY
Optimum Gain Volume is a new refinement in amplifier technology that provides a significant improvement in signal-to-noise ratio (S/N ratio) performance. The S/N ratio of an amplifier is a key indicator of performance. However, in general, the value quoted by amplifier manufacturers is measured under conditions that are nothing like typical listening conditions, and, accordingly, it does not provide a true indication of amplifier performance. At high volume levels noise is masked, and difficult to hear, but at low volumes it is readily apparent. The design objective for our Optimum Gain Volume circuit was to improve the S/N ratio of an amplifier under realistic listening conditions, and for this reason we have focused on achieving a higher S/N ratio at realistic user listening volume levels. In order to objectively evaluate the performance of our design under realistic listening conditions, we chose to perform the S/N ratio comparison under the test conditions specified by the IHF A-202* international standard. The reason for this is that the standard provides a means of objectively quantifying the S/N ratio of an amplifier under typical operating conditions, and provides a more honest and realistic indication of amplifier performance. When measured under the test conditions, typical modern audio amplifiers yield an S/N ratio of between 80 and 85 dB. In contrast, our amplifiers with Optimum Gain Volume can achieve a S/N ratio in excess of 100 dB through optimization of the gain balance. In addition, we have maintained complete compatibility with conventional amplifiers with regard to gain to speaker output.

B ASIC OPERATION

The schematic diagram in Fig. 1 shows a conventional amplifier block diagram, and the gain balance configuration employed in the Optimum Gain Volume circuit. The IHF A-202 standard requires the volume to be adjusted so that a 0.5 V line input produces a 1 W speaker output (with a speaker impedance of 8 ohms). In the case of a conventional amplifier, to produce 1 W of speaker output from a 0.5 V input, there is generally 30 dB attenuation at the volume control, and the signal is then amplified by around 16 dB in the preamp, and a further 29 dB in the main amp. In contrast, as you can see from the block diagram of the Optimum Gain Volume circuit, the input signal first passes through an attenuator where it is attenuated by 14 dB. It then passes through the Gain Adjustment Amplifier (a 0 dB to 16 dB variable-gain amplifier), a tone control circuit, and the main amp, where it is amplified by a further 29 dB for output to the speaker. To satisfy the IHF measurement conditions, the input signal is attenuated by 14 dB in the volume circuit, passes through the Gain Adjustment Amplifier with 0 dB of gain, then through the tone control
*IHF A-202 is a measurement standard defined by the Institute of High Fidelity Inc. (IHF) of the U.S.A. This standard defines the typical volume (listening) level that users operate HiFi amplifiers to be that which produces 1W of output for a 0.5V input. The IHF A-202 standard is widely used in the U.S.A., Japan, and many other countries as well.
circuit, and on to the main amp where it is amplified by 29 dB to provide 1 W of speaker output. Comparing the two circuits, it is clear that after passing through the volume control there is difference of about 15 dB in the signals. Assuming that the thermal noise generated by the volume control and the input conversion noise of the tone control blocks of the two circuits are the same, the Optimum Gain Volume circuit provides about a 15 dB improvement in S/N ratio.
Fig. 1 Amplifier Topology for Digital Audio
Fig. 2 Improved Gain Configuration for Digital Audio: 15 dB Improvement in S/N Ratio
S PEAKER L EVEL COMPATIBILITY
It is possible to achieve speaker level compatibility by appropriate selection of the gain value of the gain adjustment amplifier. The gain adjustment amplifier has a maximum of 16 dB of gain, and the main amp has 29 dB of gain, so it is possible to adjust the gain to the same level as that of a conventional amplifier. The relationship between the attenuation of the attenuator and the gain of the gain adjustment amplifier is shown in the table below.
Total attenuation 0 dB -5 dB -10 dB -16 dB Attenuator 0 dB 0 dB 0 dB 0 dB Gain adjustment amp 16 dB 11 dB 6 dB 0 dB Main amp 29 dB 29 dB 29 dB 29 dB
NON -SCALING CONFIGURATION
NON-SCALING CONFIGURATION
The Dolby Digital specification allows users to specify the configuration of their Dolby Digital receiver so that the Low Frequency Effect signal (LFE) is distributed to the front left and right speakers instead of to the subwoofer. This flexibility is provided for the benefit of users who do not own a subwoofer, and for users who have high-performance front speakers, and feel that, with certain soundtracks, the reproduction of the LFE signal by their front speakers is superior to that of a subwoofer. This configuration requires that the LFE signal be added to the signals for the left and right front speakers at some point in the receiver circuit.

S CALING THE CONVENTIONAL A PPROACH
Given the ease with which digital signals can be handled, it is perhaps unsurprising that most manufacturers of Dolby Digital receivers choose to add these signals together when it is most easily accomplishedwhile the information is still in digital form at the step prior to conversion to analog (Fig. 1)

Fig. 1

CONVENTIONAL CIRCUIT
With this arrangement, the 20-bit* data words of the LFE signal are arithmetically added to the 20-bit data words for the left and right front speaker channels before conversion to analog for outputa very simple process. However, this simplicity comes at a price. Because there is the possibility that the sum of two 20-bit words can exceed 20 bits in length (and result in harsh clipping), the data must be scaled before it is added to ensure that the result fits into 20 bits. This is done by discarding the least significant bit of each word to make room an overflow bit in the result word, and the obvious outcome is that a significant amount of audio information is lost (Fig. 2).
*The word length for Dolby Digital and DTS data is 20 bits.

Fig. 2 Scaling

Addition of two 20-bit words before scaling: + 10010100100101001001 + 11101011011010110110 Clipping Waveform = 01111111111111111111
Addition of the same two 20-bit words after scaling: Discarded Data + + = 10111111111111111111
NON-S CALING CONFIGURATION
These bits that other manufacturers discard contain low-level information (reverb tails, subtle reflections, and other nuances) that is important in creating a sense of three-dimensionality in your recordings. Your brain uses these cues to determine relative room size and instrument placement to create a well-defined sound stage. Integra Researchs philosophy is not to mess with your music by taking shortcuts. With our non-scaled circuit design, the signals are added using
analog circuits after the analog-to-digital conversion process. This complicates the design a little, but ensures that every single bit of the music data is faithfully preserved for your listening pleasure (Fig. 3).
Fig. 3 Non-Scaling Configuration
Due to a policy of continuous product improvement, Onkyo reserves the right to change specifications and appearance without notice. THX is registered trademark of Lucasfilm Ltd. & TM. all rights reserved. Surround EX is a trademark of Lucasfilm Ltd. and Dolby Laboratories Licensing Corporation. Dolby and the double-D symbol are trademarks of Dolby Laboratories Licensing Corporation. DTS is a trademark of Digital Theater Systems, Inc. All trademarks, registered trademarks, copywrites and images are the property of their respective owners.
ONKYO CORPORATION Sales & Product Planning Div.: 2-1, Nisshin-cho, Neyagawa-shi, OSAKA 572-8540, JAPAN Tel: 072-831-8008 Fax: 072-833-5222 ONKYO U.S.A.CORPORATION 200 Williams Drive, Ramsey, N.J. 07446, U.S.A. Tel: 201-825-7950 Fax: 201-825-8150 e-mail: research@onkyousa.com ONKYO EUROPE ELECTRONICS GmbH Industriestrasse 20, 82110 Germering, GERMANY Tel: 089-849-320 Fax: 089-849-3226 ONKYO CHINA LIMITED Units 2102-2107, Metroplaza Tower 1, 223 Hing Fong Road, Kwai Chung, N. T., HONG KONG Tel: 852-2429-3118 Fax: 852-2428-9039 ONKYO INDIA PVT. LTD. No. 1, Sriram Nagar, South Street, Alwarpet, Chennai 600 018, INDIA Tel: 91-44-467-2152 Fax: 91-44-467-2304