The Diffusion (DIFF) section of ValhallaDelay was designed from the ground up to be as simple to use as possible, while sounding good at all settings. Ideally, I wanted a diffusion network that would allow for slight amount of smearing of delays, all the way up to massive reverbs. The Diffusion went through a lot of iterations over the last 3 years. Here’s what we landed on:
- Two controls, Amount and Size. I had originally designed a Diffusion section with 6 controls, but the vast majority of these controls would be useless for (say) 90% of all intended uses. I decided to “kill my darlings” and limit the controls to the most commonly used controls.
- No on/off switch. Turning Amount below 1% turns the diffusion off, which results in CPU savings.
- Hardwired modulation depth. In the past, I’ve had Mod Rate and Mod Depth for diffusion sections. For ValhallaDelay, I wanted something that just “sounds right” at all settings. So I figured out a modulation scheme that results in no net pitch change, and is virtually inaudible at low Diffusion Amount settings. The modulation depth is also scaled by the diffusion size, with the idea being that a short puff of diffusion for a smeared delay should have less modulation than a longer diffusion network used for reverb.
- DENSE diffusion. A diffusion network uses a bunch of cascaded allpasses. I was originally going to allow for selectable orders for this network, but the higher order network always sounded better to my ears.
- HUGE size. The diffusion network can span 20 seconds of delay space, without sounding sparse or grainy. This allows for massive reverb sounds.
A deeper dive into the diffusion controls:
- Amount: Controls the coefficients of the diffusors. The larger the coefficient, the longer the decay of the diffusors, and the shorter the attack. A few good starting points:
- 0% is OFF (the diffusion section is bypassed)
- 68%: the diffusion attack is the exact length as the decay. Nice for reverbs that fade in and out.
- 91%: A good setting for a “pure diffusion” reverb that doesn’t rely on FEEDBACK for its decay.
- Size: Controls the length of the diffusion network, as a percentage of the delay length for each channel.
- 0% is the shortest
- 100% is the longest, and corresponds to the diffusion network being the exact length of the displayed delay
- In almost all modes, the diffusion length and “straight” delay length will add up to 100% of the displayed delay. So, as the diffusion gets longer, the “straight” delay gets shorter. This allows for the diffusion size to be adjusted without affecting the overall delay.
- RevPitch is the exception to the above rule. The diffusion is added onto the base delay. This is necessary, as the RevPitch delay is constantly changing from the displayed delay to the shortest delay as the signal is reversed.
Most of the Valhalla Delay modes have “standard” diffusion, where all frequencies have the same amount of phase delay applied. The Ghost mode has a unique diffusion in the Present and Future ERA settings, where higher frequencies have less phase delay applied. The results of this are really weird, with the sounds ranging from ghostly voices from parallel planes, up to bouncing ball delay effects when the Diff Size is high and long delays are used. We’re really proud of the diffusion in the Ghost mode, and I’m really proud of Don Gunn for naming this mode!!!