Are you using different components than you have traditionally?
B.J. Buchalter: Not really, although the components of the types that we have been using continue to evolve and we continually evaluate new products to see if they fit with upgrades or new designs.
what are they and in what areas are you seeing improvement (ease of design, performance aspects, etc).
B.J. Buchalter: Generally in terms of noise and distortion specifications. Unfortunately, this is always in competition with quiescent power, which is a real issue for the sorts of high-density designs we do.
If there are components that you have traditionally used (analog or digital, ancillary components other than chips) that are becoming hard to source; how are you adapting? What new approaches are you taking?
B.J. Buchalter: It is a struggle; in the past we have tried to focus on suppliers that have a good track record of part availability. This past year we’ve had to work harder than usual to find parts when we need them. As a result, we have been increasing our stocking position to be less sensitive to market availability.
Do your designs include surface mount components and is this usage increasing?
B.J. Buchalter: Yes; our designs are (and have been) almost completely surface mount; we generally only use through-hole parts for bulk capacitors, transformers, connectors and other electromechanical parts (like potentiometers, encoders and switches). SMT parts are easier to source these days, generally easier to rework/service and allow for (much) higher density designs.
What criteria determine your selection of A/D and D/A conversion parts.
B.J. Buchalter: First and foremost subjective sound quality evaluation. Secondarily is noise and distortion specs, and third is latency. Other factors like power and packaging are not really considered in our design process.
Have you adopted any new conversion parts this year? If so, which and why?
B.J. Buchalter: Yes. They are not new to the industry, but new to our products. AKM AK5394A and AK4395. Why? Because they sound so good.
Are you using SRC parts in your designs? What parts are you using, where and why?
B.J. Buchalter: Yes. On our older products CS8420 on the AES/SPDIF I/O (with SRC available on the input). On the 2d Card, we utilize the SRC block in the ADSP21364 for TOSLINK input. In both cases to allow the user to get input from a wild source device.
Formats such as USB, Ethernet, AES50 and FireWire are being increasingly used for audio purposes, alongside AES3 and ADAT optical. Are you incorporating new protocols into your designs? If so, which protocols are you adopting?
B.J. Buchalter: We have been using FireWire (and AES/SPDIF) since we began as a hardware company. We are investigating other protocols for future products.
If you are using digital interface protocols, what devices are you using for interface?
B.J. Buchalter: We use the TI PHY and LLC parts for Firewire (TSB41AB2 and TSB12LV32), with the protocol layer implemented in FPGA and DSP (our own design). We are happy with it; we believe that it has stood the test of time and freed us from reliance on system vendors and their release/bug fixing schedules. It has also allowed us to implement novel new solutions based upon existing hardware (with simply firmware changes) and support the addition of significant expansion capabilities to older products – thus preserving our customer’s long-term investments. For AES/SPDIF, we use AES transceivers from Cirrus. For ADAT, we use the Wavefront semi ADAT transmitter and receiver parts.
With some of the formats listed in Q7, devices are powered off of the interface bus, or for other reasons like portability, lower voltage supplies are employed than in traditional audio designs. Does this low voltage approach present challenges in the analog circuitry proceeding or following conversion?
B.J. Buchalter: It would if we didn’t utilize switching converters to get reasonable analog supply levels in our products.
Do these challenges mean performance compromises or present restrictions in analog performance?
B.J. Buchalter: Not in our products, but they do present significant challenges in the power supply and distribution engineering within the products.
If applicable, what techniques are you employing inside your products to raise these smaller voltages to higher supply rails?
B.J. Buchalter: In general SMPS technology (either boost/buck or flyback) generally with high-frequency filtering and linear post-regulation to provide clean analog supplies at a wide variety of required pro-audio rails sourced from a wide-range unregulated input power source. This is definitely one of the largest challenges we faced in designing our products.
If you are involved in DSP design, why have you chosen the components you use?
B.J. Buchalter: We use the Analog SHARC parts + some FPGA processing. The SHARCs have a really nice balance of floating point and fixed point processing as well as overall control/microcontroller type instructions. The SHARC is easily programmed in assembly code for optimal or nearly optimal algorithm implementation. The SHARC has a rich set of peripherals with dedicated DMA and a good memory model. Finally the SHARC has a well documented and easily understood machine programming model that made it possible for us to implement dynamic compilation and dynamic loading over FireWire – features that are critical to our +DSP environment.
Do you see advantages in one family of processors over another?
B.J. Buchalter: At the end of the day, if you can get the chip to do what you want, that’s all that matters. That being said, if the chip architecture and instruction set allows you to focus on DSP rather than managing the chip, you can get more done with less effort. The SHARC seems to be a near perfect blend. Also, since we have lots of code for the SHARC at this point, it has a natural advantage for us as we don’t need to recode if we continue moving forward with the SHARC.
Do you see advantages in the design tool sets available for DSP programming for particular families of components?
B.J. Buchalter: The VisualDSP tools are quite nice, but we don’t really exercise the tools very much. We have found that even with modern optimizing compilers, we can still do a much better job of optimizing DSP code by hand. With other DSPs, we may be more likely to utilize the compiler, but since the SHARC has an easy to manage execution model, we can routinely beat the compiler for code speed and size. In small machines where you want to get the most performance possible, this is a huge win.
Have you experimented with, or employed, FPGA processors for DSP tasks (and if you are employing them, why)?
B.J. Buchalter: We have experimented with it, and we have some small processing block implemented. We expect to add more in the future. The benefit is massive performance for specific tasks. The drawback is that they are much more complicated to develop for, and much less flexible at run-time.
Are their advances in native processing that are now allowing you to perform DSP tasks formerly relegated to dedicated hardware inside a computer’s CPU?
B.J. Buchalter: We actually started as a Native processing company. Native processing has come a long way, and is clearly better suited to a wide variety of processing tasks than DSPs (large memory, large processes, latency is not an issue types of process – like mixdown). Even though you can run current native DAWs on current native hardware with reasonably low latencies, you give up much of the advantage of native processing when running with very low buffer sizes (e.g. the available processing power drops dramatically as you approach single-sample buffers). We still find that the best approach is a native/DSP hybrid – where critical low-latency processing is done with singe-sample processing loops on DSP feeding and being fed by a native processing engine that can handle the tasks that are less latency sensitive.
Are there advances in DSP technologies that you are particularly excited about?
B.J. Buchalter: The new SHARC family has a number of exciting architectural features that we look forward to taking advantage of (FIR, IIR, and FFT coprocessors, delay line DMA, link-ports and DDR memory support along with a fast SHARC core).
What devices are you employing for system control—do you integrate control into DSP or other processes or are you using dedicated hardware for human interface, i.e. button commands and parameter display?
B.J. Buchalter: We utilize DSP + FPGA for system and UI control. The DSP does FireWire transport and system control and the FPGA implements lower-level UI functions like display controller, encoder decoding, debouncing and message formatting. Our newest product implements a large scale multi-channel PWM controller to allow color and brighness control of the 495 bi-color LEDs that are on the front-panel of the device. This allows software control over the usage of the indictators and has allowed us a great deal of flexibility in implementation and UI functionality.
What advice would you give a consumer who is trying to intelligently assess a purchasing decision—as a designer, do you have any guidance to share? Trends you’d like to comment on? Devices that we haven’t asked about that you are excited about or are turning to for more of your designs?
B.J. Buchalter: One critical thing is that parts don’t necessarily equal performance. Two designers can utilize the same basic components and come up with designs that have dramatically different performance characteristics. It is a mistake to say, “this box uses this ADC so it must be good”. The components (in some cases) may provide an absolute limit on the performance of a product, but they certainly do not guarantee a specific level of performance. Processors keep getting faster and cheaper, serial communications are the future, our audio devices will continue to get smarter, and audio quality has become extremely good at reasonable prices. This doesn’t mean that you can get world-class quality and functionality for a few hundred bucks, but you definitely can without having to spend tens or hundreds of thousands of $ these days. These are all good things.