#dd-5

Without regurgitating the Strymon white paper here, I’ll just point out that the DD-5 belongs to the ‘middle era’ the paper talks about -- low-resolution PCM. Released in 1995, it used separate 16-bit (high for the time) ADC and DAC chips, along with the same DSP as the Roland RE-800 rack delay. There are two main limitations to the design that affect how the sound is handled.

First, 16 bits would seem fine, since that’s the audio CD standard resolution -- but quantization noise can be a big problem when you’re dealing with the wide dynamic range of musical instrument signals, especially since an overdriven preamp might add 40dB or more of gain to any preceding noise. The DD-5 ADC uses delta-sigma oversampling to deal with this, so quantization noise is shifted out of the audible spectrum, and noise performance of the pedal is quite good even by modern standards. But the ‘adaptive focus’ method that Boss touted in the user manual is just a compander -- compress on the way in, expand on the way out. Depending on how accurate the companding is, and whether its detector sidechain is frequency-selective, it can affect the dynamics of the delayed repeats (most people talk about this as the ‘feel’ of the pedal). Moreover, the pedal has analog pre-emphasis and de-emphasis circuits (which also affect the dry line), where small non-linearities may creep in.

Second, the DSP has limited on-board RAM; there are no outboard RAM chips in either the RE-800 or DD-5. In order the extract the maximum delay time from the DSP (2 seconds, also the maximum “hold” time), the DD-5 runs the ADDA conversion at 32kHz, rather than the CD standard of 44.1kHz. That means the Nyquist frequency of the conversion is in the audible spectrum (16kHz). To avoid audible aliasing on the delay repeats, the DD-5 uses an active analog Sallen-Key filter on the wet line outputs (incidentally, the same filter topology used on the wet lines of the old Boss/Roland choruses to filter out BBD clock noise). The steep slope makes sure that the output is -12dB at 16kHz, but the filter is also a little bit resonant, as you can see in the graph below:

(click here to enlarge)

Basically, from 5-8kHz there’s a little bump in output, followed by a steep drop-off. Moreover, the phase response of the whole treble area is pretty mangled. (Side note: this isn’t that different from what people claim makes the Echoplex preamp nice -- slight bandwidth limiting, a bit of resonance, and a phase shift in the treble.)

The net effect in the DD-5 is a love/hate proposition. If you run the pedal after distortions or in the FX loop of a gainy amp, the delay repeats will seem to be just a bit brighter or harsher than the dry tone, and may filter/intermodulate the treble in unflattering ways. A lot of people end up modifying the DD-5 to make the repeats darker (just change the feedback caps in the Salley-Key stages). I actually tried modding for darker sound years ago, but didn’t like it, and went back to stock.

Why? I run the DD-5 in front of the amp, delaying my clean sound BEFORE it goes through a massively-distorted high-gain preamp. The slight treble bump in the repeats, and the phase shift from normal, increased the attack and separation of my tempo-synced delay vs my dry signal, so the repeats didn’t muddy things up once everything got massively clipped. I hated the darkened repeats mod because it removed the quality that was making the pedal work for me. Even switching over to a DD-7, with its modern conversion and totally flat output, something was missing -- the repeats just weren’t cutting through like they did with the DD-5.