Valhalla Room Reverb

If you caught the '1000 Years of Reverb' talk with Sean Costello (Valhalla DSP) [link here] earlier this year, you'll remember when Sean was discussing classic Lexicon reverbs and there was a connection made with a member of our audience, René Jaeger, who was working at Lexicon when the 224 and PCM60 were being developed.

When Sean covered reverbs from the '90s such as the Alesis family of products, Rick Rodriguez joined the discussion, as he worked on the hardware of some of those famous reverberators. That exchange inspired a riveting panel-like discussion with these three folks, and we wanted to continue the magic at our next meeting.

We were exceedingly fortunate to be able to present this event in the gorgeous 'Stone Room' at Robert Lang Studio, a space which is worthy of a meeting by itself. The story of its genesis is remarkable enough; when you consider the talent that has recorded there, you know you are in a special place. The many videos and TV shows recorded there have shared it with the world.

Our presenters for the Evening of Reverb:
Sean Costello
Sean started out experimenting with Csound at the University of Washington in the late 1990s. He was hired by Staccato Systems in 1999 and worked developing physical models for video games. Staccato Systems was acquired by Analog Devices in 2001 and he began working on audio algorithms and development tools for the SHARC and Blackfin DSPs. In 2007, he began working as a consultant, developing a wide variety of audio algorithms for a variety of clients. He started Valhalla DSP in 2009 with the release of ValhallaFreqEcho. Sean approaches algorithms from a psychoacoustic perspective and enjoys creating and selling tools directly to artists.
Rick Rodriguez, Hardware Design Engineer:
Rick's audio pedigree begins with Quad Eight Westrex in 1983 and spans two decades of electronic hardware design and development for the musical instrument and professional audio markets with Fender Musical Instruments, Alesis Studio Electronics, and consulting work under Hugsley Audio Research. He graduated from California Polytechnic State University, San Luis Obispo with a Bachelor's of Science Degree in Engineering.
René Jaeger
René worked at dbx, designing audio compressors and noise reduction systems. He then moved on to Lexicon doing A/D and D/A converter development as well as the legendary PCM-60 digital reverb, culminating in the even more legendary 480L reverb. A brief stay at New England Digital ended with a move to California to the Grass Valley Group, after which he moved to the PNW, where he worked for Mackie Designs/Loud Technologies designing mixers and a class D power amplifier.
In addition to the above, René has done consulting work for instrumentation, medical and professional audio companies, managed the development of the Pacific Microsonics Model One HDCD mastering converter, and partnered in the founding of Berkeley Audio Design. After retiring in September 2014, René is devoting himself to the perfection of audio reproduction at home and other household activities.

About the Studio
Forty-one years ago, Robert Lang purchased a plot of land with a one-room garage on it in Richmond Beach, WA. With big vision, good friends and a few shovels, he dug out the hillside to create this one of a kind, highly sought after approx. 9,000 sq. ft. music recording facility. From Nirvana to Peter Frampton to the Foo Fighters to Macklemore & Ryan Lewis, the unparalleled vibe and sounds you can achieve at RLS attract musicians, engineers and producers from all over the globe as well as students and aspiring music professionals to learn in the Education Program offered at the studio. Recently featured in Dave Grohl and the Foo Fighter's Emmy Award winning HBO documentary series, Sonic Highways, RLS continues to capture the hearts and minds of the world's music lovers and professional music-makers. Visit Robert Lang Studios on the web

The evening began with hardware engineer René Jaeger describing his move from dbx to Lexicon in 1979. He began working on the A/D D/A pcbs for the now classic Lexicon 224, later going on to work on the 200. He continued with the story of the design effort to bring a more economical (<$2k) reverb to market, which would become the PCM60. After René pointed out some of the sourcing challenges, Sean Costello asked about the ALUs (Arithmetic Logic Units) used, and a discussion on their limitations ensued. Sean remarked that some algorithms did well with limited ALUs while others did not, noting that the hardware informed the choice of algorithm and design.

The discussion moved on to the user interface of the PCM60, noting it was a preset-driven interface (rather than parameter driven). René pointed out that this was one of many decisions to keep the cost down allowing Lexicon to bring it to market for $1500. Next, he discussed the design of the 480L based on 4 sets of the chipset used in the PCM60, allowing for novel signal paths.

Sean pointed out that the PCM60 did not use a micro controller (eg Z80); he asked what would have been different if they had used one. René referred to David Griesinger's site (and paper), going into some background of Dave's experience with real halls and some of the key aspects of the perception of reverberation. Specifically, he noted how important the first 50ms (initial reflections) are in defining the reverb and how the later echoes are 'smeared'. An optimization possible with a microprocessor is moving the delay taps of the tail to simulate this 'smearing'. Also, the microprocessor in the 224 and 480L allowed for hot swapping parameters, which meant the LARC (remote control for 224XL) allowed the user to make changes in real time.

Next, René pointed out how limited bandwidth was actually a good thing for reverberation since it mirrored high frequency absorption in physical rooms.

A question about Alesis' role in the development of reverb technology prompted a change in focus of the discussion to Alesis. In particular, how Keith Barr was influenced by Lexicon when he released his digital reverberator — the first sub $1000 unit at $799. Sean pointed out that the Midiverb 2 was still in use today particularly for its so-called 'bloom' patches - algorithms that had a slow build up (around 5 seconds) and a long delay (around 15 seconds). This slow build up was a consequence of many simple all-pass filters routed in series, and became a key component of the signature sound of many artists of the time.

At this point, Sean revisited the bandwidth issue, pointing out that higher bandwidth can be a bad thing for reverb algorithms. Essentially, high bandwidth exposes problems with an algorithm that may not be apparent with lower bandwidths. In particular, the noise may be masked by limited bandwidth. The panelists pointed out that this noise became part of the sound, and that users may have come to expect it as part of the reverb sound.

This led to a discussion on bit depth which led to a slight digression on dynamic range, with more discussion on noise and psycho-acoustics, A/D and D/A converters, and various EE anecdotes.

After a brief break for refreshments and circulation, door prizes were awarded. Ssd data erase.

• 4-Gig Thumb-drives courtesy Robert Moses, Executive Director, AES
  • Bob Miller
  • Matt Stegner
• Dan Mohr won a Pomona Volt-Alert, courtesy Rick Rodriguez and Fluke.