Tube Amps / Music Electronics
|For current discussions, please visit Music Electronics Forum.||The sunn still shines online!|
|1/3/2003 11:40 AM|
||Re: Barberpole Phasing( was Re: Digitech Hyperphase)|
Sean - do you need a special reader to see the patent details on the site you indicated? I get the initial description OK, but when I click on IMAGES, I just get a blank screen.
|1/3/2003 8:46 PM|
A special reader is needed to view the images. Info on how to configure your system can be found at
|1/3/2003 8:57 PM|
I'm sure that it would be impossible to exactly clone a PH3, as it probably uses custom Roland chips for the DSP, ROM, etc. Many of the Boss pedals seem to have custom Roland chips. Whether these are actually made by Roland, or are in fact Atmel chips that are repackaged, is another question. Someone else once pointed out how closely the Roland Sound Canvas modules resemble the application notes for the Atmel Dream Sound synthesis chips.
Anyway, even though it would be impossible to exactly clone a PH3, it WOULD be possible to design a DSP-based pedal that would run algorithms similar to the Boss PH-3. Even a cheap 16-bit fixed point DSP would do a pretty good job with the PH-3 algorithms:
- very little memory needed (program memory, and a tiny amount of memory for the allpass filter states)
- 1st order allpasses work well in 16-bit fixed point.
Of course, higher bit resolution systems are always nice, but you could run a 16-bit DSP in extended precision mode if need be. A 32-bit floating point chip (such as the Analog Devices SHARC or TigerSHARC) would have the advantage of being easier to port existing computer plugins, such as VST.
|1/3/2003 2:28 PM|
I was always under the impression that all that was really needed for Barberpole effects were two versions of the same effect sidechain (phase or time shifted), driven by quadrature ramp waves (ascending or descending). The ramp part was what I assumed to be essential since something always had to be moving in the one direction. Indeed, that's why it is called "barberpole" since that's how the barberpole visual illusion works too.
I'll check the patent and compare it to my muddled thinking.
|1/3/2003 8:42 PM|
The technique you describe is another method of obtaining barberpole effects, and is a good way of making a "barberpole flanger." It is also how simple pitch shifters work.
The frequency shifter can be looked at as a form of quadrature crossfading, but with some weird phase rotation also going on. I am not that mathematically inclined, so I leave the math to someone else.
A few other ways of obtaining barberpole effects:
- Crossfading between 4 or more bandpass filters, that are being swept in frequency. The crossfading should have minimum gain at the beginning and end of the frequency sweep, to hide the "glitches" when the filter returns to its starting frequency. This is very similar to the Risset "endless glissando," but with filters instead of sine oscillators.
- Use FFTs to convolve the sound of a Risset "endless glissando" with the input signal.
- Use FFTs, and a graphic sonogram->sound program such as Metasynth, to create the barberpole effect. This is done by "filtering" the image with a graphic that looks like many parallel diagonal lines.
For digital applications, I prefer the frequency shifter technique for barberpole phasing, as it produces a nice sound, and you don't have to get into fancy math to compute the sweep function for the allpass filters (the allpass filters are fixed). The Boss PH-3 sounds IDENTICAL to the frequency shifter based barberpole phasers, including the perceived pitch shift at high sweep rates.
Even though the Bode patent is from 1983, using a frequency shifter to obtain barberpole phasing dates back to at least 1973. In the Pink Floyd song "Time" from Dark Side Of The Moon, the phasing sound on the female vocals was obtained by routing feedback around a frequency shifter that was originally used for feedback supression in PA systems.
|1/3/2003 9:12 PM|
BTW, I listened to the Barberpole phasing effect on the Digitech website for the Hyperphase. To my ears, it sounds somewhat different than the Boss pedal. While I am fairly confident that Boss uses the frequency shifting technique, I am not sure that Digitech is using this technique. It sounds like they might be crossfading between two phasers, driven by ramps. Since they have a "Triggered" phase shift, which sounds like it triggers a downwards ramp, having a ramp/crossfade based system is not out of the question.
|3/12/2003 9:43 PM|
So stage 1 would seem to be a decent panner. Now I have a circuit for one of those - the Accessit panner long since defunct but worked very well. It's here somewhere in one of serveral piles of papers!
|<<First Page||<Prev||Page 2 of 2|