The world is changing, and it is time, once again, to reconsider the design of a standalone Crescendo system for live performance use.
For the past decade, I have been using a Muse Receptor dedicated to running Crescendo on any audio that you feed it. But Muse is no more, and the versions that I have running in my Mac DAWs seem to suffer a bit more latency than I’d like for live sessions – thanks mostly to OS X and the I/O buffering it provides.
[When I practice guitar, I definitely have to skip using my DAW Crescendo system. The delay between playing and hearing makes it impossible for live situations.]
I looked around for some DSP solutions, since that is where it all began about two decades ago. And DSP’s are capable of performing well with very small latency. They don’t need Operating Systems, just access to the raw audio stream in and out, and uninterrupted performance without the need to manage display screens and mouse / keyboard input.
Alas, the world has moved mostly beyond DSP’s, which are now used (cheaply) in active crossover systems or other very simple applications. Yes, of course there are some really great products out there still using dedicated DSP’s – the Kemper Profiler is one good example. And there are many others.
But each of these dedicated systems built around DSP’s also had to provide a ton of I/O capabilities, A/D conversion, and D/A conversion. A big investment of time and money and engineering effort. You can only afford to do something like that if you have a large enough market for your products. In those situations, they make use of much more powerful DSP’s than are used in the typical DIY projects that you can find, along with PIC, Arduino, and ARM processors. None of these DIY level products are capable of performing Crescendo calculations, nor of offering high-quality audio.
So I am concluding that for small quantities of Crescendo systems for live performance, probably the best we can do is a Mac Mini attached to a small MOTU or other audio interface. That allows us to perform all the hairy Crescendo calculations with time to spare, and we can augment the system with hyper-recruitment pre-equalization plugins, and headphone / speaker / room correction plugins after the Crescendo processing.
This offers standalone, high quality audio processing for a price somewhere around the $1600 mark (about $800 each for Mac Mini, and MOTU Ultralight audio interface). Yes, that is somewhat expensive, but still less than 1/4 what high-end hearing aids would cost, and provides far better than any hearing aid ever possibly could.
Latency will still be an issue, but perhaps a Mac Mini dedicated to only this purpose can run with very small buffering. We need to keep the latency down around 10 ms or less for live performance.
The Crescendo algorithm, itself, produces around 3-5 ms of latency. This Crescendo latency is imposed by the laws of physics, where we need to channelize the spectrum to within 200-300 Hz resolution. It has nothing to do with processor performance.
So buffering will have to be around 128 samples (in & out at 48 kHz), or fewer, to keep the additional I/O latency to around 5 ms. Fortunately, recent vintage Intel processors can handle that with relative ease. (not like the early days of Crescendo when they could barely manage)
Once all the plugins have been calibrated and adjusted to my hearing, we can just run the Mac Mini as a headless workstation. Set and forget. Just needs to boot on its own, without a keyboard, mouse, or display screen. During setup we can use screen sharing from one of the studio DAW systems.
So, looks like that will be my next project. I don’t want to get involved in a manufacturing line with DSP’s and audio I/O design. That is too much effort as I get older, and life is becoming too short.
I will attempt to create a standalone Crescendo system based around a Mac Mini (recent vintage or new), and a MOTU Ultralight-level audio interface. I will attempt to squeeze down on the playthrough latency to see if we can get an acceptable Crescendo for live performance.
[ PS: for listening, I will probably continue to use my Cranesong Solaris DAC. I have never heard anything better than this, and it has output gain just sufficient to allow more than 20dB of playthrough headroom. For serious levels of hearing impairment, you sometimes need that much headroom on symphonic dynamic range material, to get playback without any clipping events.
The Mac Mini can use the MOTU Ultralight for audio input, and playback for portable situations, but the Cranesong Solaris can use a sister USB port on the Mac Mini for highest quality playthrough.]