Selecting DSP audio equipment
These days there are more and more DSP (digital signal processor) audio products available. You know the type of product with analog inputs and outputs on the back panel with some RS-232 and/or network control, a suite of audio processing capabilities including equalization, mixing, matrixing, and often including acoustic echo cancellation for conferencing applications, and perhaps even a digital bus interconnect to link the device to one or more other devices.
There are two primary styles of products – open architecture and fixed architecture – depending on whether you can choose which DSP algorithms are in which order (open) or whether you adjust parameters on DSP algorithms that are mostly in a fixed sequence (fixed). There are even some hybrid architecture products available to give you the best of both options.
Commonly used DSP products are available from Polycom, Biamp, CleareOne, and many others including Atlas Sound, Crestron, Extron, Rane, and Symetrix.
With so many options, how does someone choose which product to use? And who chooses? The partner or the end user? Does the end user care? My experience has been it’s mostly the partner who selects the product because the end user is typically looking for the functionality and isn’t as aware of the nuances of the different products.
While the selection often comes down to a partner’s experiences and skill levels with the particular devices, here are a number of additional factors that can, or should, play into the decision:
Do you need acoustic echo cancellation (AEC) capabilities? Better make sure you have enough inputs with the AEC processing capability to support the current number of microphones needed and perhaps even think a bit about future needs.
Need feedback elimination? What’s the latency between analog input/analog output? You want to make sure that latency is no more than 20 msec or so to minimize the delay between the person speaking and the reinforcement in the room.
Need VoIP? What are the options for adding VoIP? Plug-in card (common) or an extra device that requires its own rack space real-estate?
How easy is it to create a design? How good are the tools to create the configuration file? How quickly can you create a design (or reuse an existing design) for your needs. The software tools should make it easy to create a design so you can spend your time on understanding the needs and the application rather than on the tool.
While most products these days have such clean audio processing paths that it’s hard to find a device that sounds bad due to poor analog input/output sections, you should understand what input gains are done in the analog domain and which ones are done in the digital domain.
Too much digital gain is like too much digital zoom on a digital camera – the audio quality suffers because of it.
Are there design templates you can use to create your configuration file?
Is there documentation that describes how to create your design or use the features you want to use? While most people using the DSP products don’t read the documentation, there is a huge amount of information available in the documentation of most of the these products.
How capable is the support team?
Are there third-parties that can create the design file for you if you are busy or simply need some help?
How easy is the system to control? Is there a decent user API syntax that is human readable? Is it cryptic (hopefully not in Hex!) or is it more readable? Can you control the device with command names and parameter names (easier to understand) or do you need to assign control numbers and control the numbers (potentially more confusing, but often used with the flexible open architecture products).
Are there Crestron/AMX/other control modules that are available that make the integrators job easier?
While the price of the DSP isn’t supposed to be an issue due to the cost of all the other components in the system, the reality is the price of the DSP does matter either due to other integrators proposing lower cost alternative solutions or because when multiple devices are used, the DSP cost can be a larger portion of the overall system.
Keep in mind the price of the system is not just the price of the hardware, but the integration cost with a control system, the amount of time it takes to create a configuration file, and waiting on customer support while you are in the field.
Do you need a physical phone dialer? Or volume control knobs? Or simply don’t want to solder to an DB25 connector?
Make sure you know what the options are because it may this job, or the next job, easier.
How easy will it be to add a few more microphone inputs? Or how easy will it be to scale to two or more devices? Can you put additional devices basically anywhere due to networked audio, or are you limited to be within 30 feet of the first unit?
Do you need to integrate natively with third-party Dante or AVB devices? If so, make sure you know what you can and can’t do with the digital networking, such as the number of channels of audio that can be shared, whether automixing information can be shared, etc.
What happens when you have multiple devices and a device needs to be replaced? Does the whole system go down or is there still some part of the system operational?
Most systems are not redundant do to the cost/complexity, but if the whole system goes down when one device needs to be replaced, you may want to keep a spare unit on site ‘just in case’. That will also add to the cost!
Wrapping it up
There are many dimensions to choosing a DSP audio product. Make sure you know what’s important to be successful with your project and then prioritize your needs and make the best choice based.