Author: mattwaddell

“The PIX240i came out a year after the original PIX240. The only real difference between the two was a much better looking IPS display. The original PIX240’s display exhibited jello-like wobble when a camera panned and was displaying it on the screen. It was a rolling-shutter-like artifact, but on the screen instead of a camera sensor. This seems like a somewhat insignificant change, but like many changes that look simple to the outside world, it was far more difficult on the engineering side. It’s rare that you can buy a part – like a screen – from one vendor and it simply drops into the product with no changes. If memory serves, we had to change our injection molds to accommodate the new screen as well as the driver circuitry, both of which are pretty huge changes that took many months to do. The new IPS screen from Mitsubishi was brighter, had better scan time, and didn’t exhibit any of the jello effect. We went on to make the PIX240i for several more years, and this improved screen was a big part of its popularity. The PIX220i was simply a lower-priced version of the PIX240i in which we omitted the SDI in/out. As the parts needed to implement SDI in and out are very expensive, this did allow us to sell the HDMI-only version PIX220i at a lower price, but it was never as popular as the PIX240i.”

“The PIX 240 was our first video product. We (perhaps naively) thought we could parlay our digital recording and timecode expertise into a new market: video. We fielded calls almost daily about timecode and how to sync up sound with video. We figured – why not record both on one device? I could never convince camera manufacturers to integrate decent audio inputs. We decided to sort of be the camera ourselves and record our own high-quality video along with our very high-quality audio, eliminating the need for timecode. There were tons of nice cameras with great sensors spitting out uncompressed video via SDI, but the internal (tape or card) recording on the cameras was terrible. I knew that we could record the rates we needed to 2.5” drives reliably, but just needed to get a license from Apple to implement ProRes on an FPGA. I approached them several times but was told that Aja had an exclusive. I hired the engineering firm Atterotech (Fort Wayne, TX), whose engineers knew both audio and video. They found a company who manufactured an odd 360 core processor which somehow had an Apple license, so we went that route. Atterotech and I designed the hardware together – I drafted behind these very talented engineers until I learned all the ins and outs of designing video. Mike Lawson (Mechanical Engineering) designed the chassis and the docking 2.5” HDD caddy. He and I spent some time working on the cooling of this product and utilized a large-diameter, slow spinning fan on the back. It was utterly silent and cooled the massive multi-core processor well. The video topped out at 1080p30 and we recorded in either ProRes or DNxHD. We introduced the product thinking we had a hit on our hands, and we did. People still tell us they think the PIX 240 was one of the best products we ever produced.”

“The original MixPre came out in the year 2000, and in the intervening years, USB digital audio had grown in popularity. This new version, called the MixPre-D, is in some ways one of the coolest products we’ve made, as it was the last of our transformer-based input products. I took the digital guts (Blackfin DSP) from the USBPre-2 and fitted them into the original MixPre chassis along with some other enhancements, like an AES digital output and an M/S matrix. The analog section stayed the same as the original MixPre with the Lundahl transformers, opto limiters, and conductive plastic pots. The power supply was another one of my super-efficient flyback varieties featuring several supply rails (+15V, -15V, +48V, +5V, +3.3V, +1.8V, and +1.35V for the DSP core) with my custom-designed multi-tap transformers that we hand-wound at our headquarters in Reedsburg, Wisconsin. This design was truly old meets new… which is now old, but so am I 🙁 We made this product for several years, but it was eventually eclipsed by the modern generation of MixPres which featured even more features and higher quality audio for less money.”

“After 10 years of making the original USBPre1.5, we had learned a lot and solicited lots of customer feedback. Also many of the parts from the original USBPre1.5 had gone obsolete. I designed the entire USBPre-2 with new improved mic preamps – class A long-tail pair with discrete transistors, better metering, and used the Blackfin DSP which was may more powerful than the original bizarre Phillips processor used in the USBPre1.5. The entire code stack for this product was written by Mark Ketilson (Software) and has not really changed since day 1. We are still manufacturing this product in Reedsburg and it stays popular even today. I can’t count the number of FoH and acoustical measurement folks I’ve met who know our company primarily through this device.”

“The CL-9 was the first of our flat-panel fader panels in the growing 788T ecosystem. Like the CL-8, this was also an ‘after-the-fact’ development. Since I knew that the USB port would work for this function, I designed the CL-9 around this port. Jason McDonald (Mechanical Engineering) as usual did the mechanicals, and Francois Morin (Field Programmable Gate Array expert) wrote the VHDL (Very High-Speed Integrated Circuit Hardware Description Language) code for running the entire CL-9. One unique feature is the fader caps in which Jason integrated neodymium magnets to pull down on the faders to eliminate an audible “click” sound on quiet movie sets when touched.”

“The 552 (and the 788T) was the start of the melding of mixers and recorders for Sound Devices. Of course we now have both the MixPre line and the 8-Series line, both of which are mixer-recorders. But before the 552, these were physically two different devices, and it was quite popular to carry the 442 mixer along with the 744 recorder. The 552 was our first hybrid unit which combined both a mixer and a recorder in one power-efficient unit. The 552 was mostly an all-analog mixer featuring 5 Lundahl input transformers. I just looked over the schematic for the first time in years – each input had dual opto-based limiters along with one stage implemented with depletion-mode MOSFETs — a total of 3 limiter stages per channel. To this, I grafted a small 2-channel recorder based on an ARM926 microcontroller from Atmel. Jason McDonald (Mechanical Engineering) designed the top and bottom panels out of injection-molded carbon fiber to save weight. The 552 was also the first (and last!) time customers met Sven, our much-maligned voice assistant. This was an effort to make navigating the blinking LEDs from the 442 that controlled settings easier to use. The microcontroller in the 552 had just enough oomph to run a primitive voice synth, but small power-efficient displays or pre-recorded voices were out of the question. A funny story: Sven had an internal setting to be a female or male voice. Foreshadowing the movie “Her”, one of our programmers had it set to female during development and said that after hearing it so much, he “sort of fell in love with her.” We all agreed that he needed to get out more…!”