Founded in 2006, BURL Audio is a pro audio gear engineering and manufacturing company based in Santa Cruz, California.
Based out of Paradise Recording and spearheaded by Rich Williams, BURL Audio’s gear is designed in a commercial recording studio, by professional recording engineers, for recording engineers. Rich’s experience in music is extensive. He has a Bachelor of Science in Computer Engineering from the University of California, Santa Cruz. He then furthered his education, graduating from the California Recording Institute. While at UCSC, he majored in DSP based hardware design, digital and analog electronics design, music synthesizer software design, and audio signal processing. He also studied music theory composition and recording arts. Rich is also an accomplished musician. He is the lead singer of the psychadelic, funk-rock band Burlacticus Undertow, who have garnered rave reviews in the Bay Area.
We were particularly interested in Rich’s view’s on the analogue audio elements of converter design, as Burl Audio appear to have been pursuing a design philosophy where converters can add warmth and detail, rather being designed solely with transparency as the goal.
First we asked Rich to comment on the really important elements within successful A/D and D/A designs;
“The entire circuit path is critical starting with the AD/DA chips. Each chip has a completely different sound. I have been researching AD/DA chips for the past ten years and we use only the cream of the crop in our Burl converters.
“The analog sections are crucial and the weakest part of most AD/DA designs. Off the shelf op-amps don’t cut it here at Burl. We use only the highest quality transistors running in class-A. One other unique feature of Burl Audio AD converters is the use of a proprietary transformer, the BX1, on the front end. This is not an ordinary transformer. The BX1 is extremely transparent (especially compared with off the shelf op-amps), has razor flat phase response beyond the audio range, and gives you the desired saturation with hot signals and large peaks. If your analog section isn’t right, it won’t matter which AD/DA chip you use. The key is to get the most out of the chip.
“Lastly, the clock has to be clean and jitter free. It’s a MAJOR misconception that external clocks are going to give you a better result. Converter chips ALWAYS run better off of an internal clock crystal. When using an external clock, you must either use a PLL (phase locked loop) or a re-sampler in your design, both of which are sub-par to an internal crystal. Almost all converter chips run off of a 24.576 MHz master clock for 48kHz, 96kHz, and 192kHz sample rates and 22.5792 MHz clock for 44.1kHz, 88.2kHz, and 176.4kHz. This is the actual clock used for sampling, not the sample rate word clock. So when you use an external clock it is run through a PLL to get this high frequency clock. The PLL adds its own self jitter.
“In most large multi-channel systems, it is impossible to use internal clock on all converter boxes because they would be out of sync. One can be on internal while the rest must be on external. In this case, one external clock usually works best so that all channel will have a consistent sound. The Burl Audio B80 Mothership addresses this problem by having a very large channel count in one box where all channels are running off of one internal crystal. All aspects of the design must be considered in order to have a pleasing sounding converter.”
What about the correct use of dither?
“In floating point systems, dither should only be used at the end of the process during mastering.”
Okay, so still from a converter design perspective, are there particular issues in getting digital audio in/out of PCs/MACs successfully?
“USB and Firewire should be avoided at all cost. MADI via PCIe is the way to go, or Digilink with Avid systems.”
So coming back to jitter? “Get a good converter and run on internal clock.”
So shouldn’t converters be transparent?
“There is no such thing as ‘transparent’ in any converter. Use what sounds the best.”
At 24 bit/96kHz, are we now approaching a point where further improvements in performance will be negligible?
Rich explained that this was a very big topic on it own but added that “Higher sample rates are desirable for a number of reasons, such as a plug-in’s ability to filter out harmonics generated beyond the nyquest frequency.”
Thanks to Rich and to Will Khan for contributing to this feature.
You can review the entire Burl Audio product range at www.burlaudio.com