What are the benefits of using "bucket brigade" delays/fx plugs?

[Numanoid]

A bucket brigade or bucket-brigade device (BBD) is a discrete-time analogue delay line, developed in 1969 by F. Sangster and K. Teer of the Philips Research Labs. It consists of a series of capacitance sections C0 to Cn. The stored analogue signal is moved along the line of capacitors, one step at each clock cycle. The name comes from analogy with the term bucket brigade, used for a line of people passing buckets of water.

In most signal processing applications, bucket brigades have been replaced by devices that use digital signal processing, manipulating samples in digital form.

Bucket brigades still see use in specialty applications, such as guitar effects.

Why ?

[justin3am]

Why not?

[resynthesis]

Probably due to the degradation in quality as the delay repeats and the perceived warmth

[thecontrolcentre]

It sounds good.

[valhallasound]

Bucket brigade delay chips are really noisy, horrible things. In order to get them to work at all, there is a ton of supporting circuitry that needs to be in place:

- BBDs work by sampling the signal. Similar to digital delays, but using stored analog charges, instead of numbers.
- The longer the delay, the lower the sampling rate. So the circuit needs to be designed to avoid aliasing at the lowest sampling rate. Which can be somewhere around 6-8 kHz or so for longer delays.
- The low sampling rate requires a TON of lowpass filtering, both in front of and after the delay, in order to avoid aliasing. These filters often have a sharp-ish cutoff around 2 to 3 kHz. DARK.
- In order to get a decent S/N ratio, companding is used. This is essentially a compressor in front of the delay, followed by an expander after the delay. This helps control the noise floor. It also does weird things to transients, and has VERY strange effects when the feedback is turned up to self oscillation.
- BBDs also have their own distortion that happens internally.

In my opinion (and I've spent a LOT of time thinking about this lately), most of the magic of BBD-based delays can be attributed to the supporting circuitry, and not the BBD chip itself. A 10-bit digital delay, with the supporting circuitry being identical to a BBD delay, will sound really close to a BBD delay. There is at least one pedal out there (of which I forget the name) that uses this approach: a PIC is used to create a low resolution digital delay, and the supporting circuitry is the same as that found in BBD pedals.

Sean Costello

[Numanoid]

most of the magic of BBD-based delays can be attributed to the supporting circuitry, and not the BBD chip itself.

Interesting point

[.jon]

What is best in life?

To crush your signals,
to feel them driven before you,
To hear the lamentation of the buckets.

[Numanoid]

To crush your signals

Good 'un

But signals are not my "enemies", they are my Friends

https://www.youtube.com/watch?v=6PQ6335puOc

[Chris-S]

Bucket brigade delay chips are really noisy, horrible things.

After reading this I expected much worse, but this sounds good IMHO.
https://youtu.be/oniORILA7lw

[justin3am]

There is at least one pedal out there (of which I forget the name) that uses this approach: a PIC is used to create a low resolution digital delay, and the supporting circuitry is the same as that found in BBD pedals.

There are a number of pedals and eurorack modules which use that approach with a princeton PT2399 or a Holtek HT8955a. I think they sound noticeable different from BBDs when modulated or when using long delay times (low sample rates). Different, mind you, not necessarily better or worse. It may still be down to filtering but the PT2399 based delays I've used sound much crunchier (maybe the aliasing is more obvious?) then the BBDs I've used. That said, I've only experimented with a small fraction of the delays you have probably measured.

I have spent a lot of time with a Doepfer A-188 using a variety of different variations on the MN3x0x series chips, applying different filter configurations and gain compensation within my eurorack modular. I've also been building my own primitive digital delay based on the PT2395, HT8955a and echo chips/"voice changers" designed for Karaoke machines, and designing my own supporting circuity (or "special sauce") along the way. I've come to enjoy the characteristics of all of them. It's also interesting to experiment with different types of clock sources. Divided down crystal oscillators are the most stable but it can be pretty awesome to clock a delay with a 4046 (as an oscillator or a PLL!), or a 555 timer.

[metamorphosis]

What is best in life?

To crush your signals,
to feel them driven before you,
To hear the lamentation of the buckets.

Awesome !

[someone called simon]

Is it only the analog pedals that do the thing that analog delay pedals lovers like, i.e., twiddle the delay knob knob and hear the pitch of the delays go up and down?
I have a TC pitchfork pedal with an 'analog' mode, but it doesn't do that. I suppose it must be able to be modelled digitally, I wish mine did that...

[jens]

There is at least one pedal out there (of which I forget the name) that uses this approach: a PIC is used to create a low resolution digital delay, and the supporting circuitry is the same as that found in BBD pedals.

Not sure, but my Carl Martin Red repeat might work like this...

[Chris-S]

Are there good BBD delay freeware plugs?

Played around with ArcDev ET-301 yesterday, seems good to me.

[whyterabbyt]

Bucket brigade delay chips are really noisy, horrible things.

You say that as though its a bad thing.