Author |
Message |
BobTheDog

Joined: Feb 28, 2005 Posts: 4041 Location: England
Audio files: 32
G2 patch files: 15
|
|
Back to top
|
|
 |
Blue Hell
Site Admin

Joined: Apr 03, 2004 Posts: 24042 Location: The Netherlands, Enschede
Audio files: 276
G2 patch files: 320
|
Posted: Tue Aug 16, 2005 1:21 pm Post subject:
|
 |
|
Did you check the building blocks section ? _________________ Jan
also .. could someone please turn down the thermostat a bit.
 |
|
Back to top
|
|
 |
mosc
Site Admin

Joined: Jan 31, 2003 Posts: 18177 Location: Durham, NC
Audio files: 204
G2 patch files: 60
|
Posted: Tue Aug 16, 2005 1:24 pm Post subject:
|
 |
|
I don't think this is sounding too bad. You are off to a good start, IMHO.
First, you will find the 1 channel mixer a better choice than the level amp for attenuating the LFO.
Second, you need to put some filter inside the feedback loop. Experiment with different kinds. Maybe a low pass filter is a good place to start.
Third, or maybe first, look at Rob's tape echo patch for some ideas. http://electro-music.com/forum/topic-7442.html
I've learned a great deal from Rob's patches, but he is a super master patcher and has a very economical style. Sometimes you have to really ponder his patches to figure out what's going on. It's worth the effort. If you have questions, he'll be agreeable to answer. _________________ --Howard
my music and other stuff |
|
Back to top
|
|
 |
BobTheDog

Joined: Feb 28, 2005 Posts: 4041 Location: England
Audio files: 32
G2 patch files: 15
|
Posted: Tue Aug 16, 2005 2:02 pm Post subject:
|
 |
|
Thankyou both for your answers, I had done a search for Tape Echo and did find the A77 echo which I had a look at but it made my head hurt, I will take a closer (and longer) look at it.
I will also try the 1 channel mixer and placing a filter in the feedback loop.
Thanks for your feedback it has given me new direction.
Cheers
Andy |
|
Back to top
|
|
 |
BobTheDog

Joined: Feb 28, 2005 Posts: 4041 Location: England
Audio files: 32
G2 patch files: 15
|
|
Back to top
|
|
 |
Rob

Joined: Mar 29, 2004 Posts: 580 Location: The Hague/Netherlands/EC
G2 patch files: 109
|
Posted: Fri Aug 26, 2005 1:50 pm Post subject:
|
 |
|
BobTheDog wrote: |
Now I am going to study Robs tape echo.
Thanks again for your help
Cheers
Andy |
Here are some clues:
On first sight the A77 patch is probably a headscratcher. But the idea is quite simple, basically there are just two things needed in the feedback path to get a tape echo sound: an allpass filter and a tape saturation emulation. Regrettably there are no simple G2 modules that directly emulate these two thingies, so they have to be 'build'.
The why of saturation speaks for itself for a tape echo, but the why of an allpass filter might be more obscure.
Allpass filter:
What an allpass filter does is pass all frequencies at nearly unity gain, but shift each frequency a little in phase, and thus in time. This phase shift is not like a 'linear time shift', but is a 'near zero' phaseshift for low frequencies to a 'near 180 degree' phase shift (inverting the signal) for high frequencies. The effect when an echo repeat is repeatedly send through an allpass filter is that the echo gets a little less bright on every time it repeats, which sounds quite natural. Using a lowpass filter takes away lots of the high on the first repeat, but hardly more on the following repeats. So this last one sort of goes like piew-pow-pow-pow-pow. Which is not what a real tape echo does, as a real tape echo has no filtering in the feedback loop apart from the tape-damping caused by the rerecording of the repeats on the tape.
An allpass filter is made by first using a lowpass filter, which creates signal A. Then this lowpass signal A is subtracted from the filter input signal to form signal B, which is like a highpass signal. Now it should be obvious that adding A and B together again would give the original signal back, as signal B is actually everything that the lowpass filter threw away. But instead of adding A directly to B, signal B is first inverted before being added to A, to create an output signal C. For signal C this means that all input frequencies are passed, but the highest frequencies are sort of reversed in phase. The 6 dB slope of the allpass filter makes sure that the phase shift slope curves smoothly, although not linearly, from almost 0 degrees to almost 180 degrees, and pass all frequencies at almost unity gain. This almost is in practice quite nice, as this means that the echo will indeed die away over time when the feedback mixer is fullly opened. In this last case you will hear the repeats gradually deteriorate over a long, but finite, amount of time.
So, when the lowpass filter is set to e.g. 2kHz, the allpass filter will give a phaseshift of exactly 90 degrees to a 2kHz sinewave. For a 100 Hz signal the shift might be only 1 degree and for a 40 kHz signal perhaps 179 degrees.
Saturation:
Tape saturation is a little more difficult to understand. The idea is to create a non-linear gain cell where the amount of amplification actually depends on the input signal level. It should be in the way that; the higher the momentary value of the input level, the lower the gain. This will compress the peaks of the signal, and it is important that both positive and negative peaks are compressed the same way. The common way to do this is to use a VCA circuit that is in its bias set to unity gain (bias sort of means its default setting, so without control modulation signal). Then, the output of the VCA is rectified with a fullwave rectification circuit. The output of the rectifier is first negated to a negative signal and then added to the biasing signal that sets the VCA to unity gain. Rectification can be done with the diode module, and for a compressor one would indeed use the diode module and then smoooth the output before negating it and adding it to the bias signal. But for tape saturation it is even better to rectify the VCA output by taking its quadrature, which is simply done by feeding the output of the VCA to both inputs of a ringmodulator (or on the G2 simply a gain controller). The reason to use this quadrature signal (which is always positive as the quadrature of a negative peak is a positive peak, remember that -1 times -1 is +1) is that this will build up a harmonic series from the ground up in a very smooth way. Lets say that only a sinewave is fed into such a saturation circuit. The circuit will now produce the odd harmonics of this sine wave in a very smooth series where the higher the harmonic the lower its amplitude. If a diode rectifier would be used there would be a bunch of higher generated odd harmonics that is relatively louder than some of the lower generated harmonics, making the sound more distorted, but not in the way that real tape saturation sounds. So, using the quadrature of the output of the VCA gives the closest emulation of what really happens when saturating recording tape.
For a more mathematical explanation one would say that one applies the third order Chebyshev polynomial to the output of the VCA, and by applying it on the output of the VCA the third order polynomial will create the odd harmonic series in a recursive way. (So, when a real nerd asks you how to make a tape echo just casually drop a line that applying a third order Chebyshev polynomial in a recursive way will surely do the job for him. )
A real head scratcher is that the quadrature of the VCA output signal should never clip and needs to be lowpass filtered somewhat. To prevent clipping the output of the VCA should be attenuated to 50% before the quadrature is calculated. Well, on the G2 that is. Filtering should be set to about 5% of the sample rate of the system, which is about 5 kHz for the G2 96kHz sample rate. This also means that odd harmonics above 5kHz will be suppressed, which is actually a good thing to get a 'vintage' sound. This filtering also applies when using an analog VCA and an analog ringmodulator in a 'through the mixing desk' feedback path for a digital delay box, to prevent the mixer becoming a radio receiver or the mixer to log on to your neighbours wifi network.
The effects of both the allpass filtering and the tape saturation created by the non-linear gain cell become apparent when the amount of feedback is increased and there are a lot of repeats. And as the non-linear gain cell also compresses the signal, the feedback can be even more as 100%! Especially in this last case it becomes quite essential to the sound how the feedback is processed. If it is not done properly it will not produce that typical low-pitched tape grunge, like when a real tape echo is repeating endlessly.
Ok that's some clues. |
|
Back to top
|
|
 |
3phase

Joined: Jul 27, 2004 Posts: 1183 Location: Berlin
Audio files: 13
G2 patch files: 141
|
Posted: Fri Aug 26, 2005 7:27 pm Post subject:
|
 |
|
Rob wrote: |
Ok that's some clues. |
Indeed...thanks ... This feedbackloop question is comming up for me for quite a while now , and here are defenetly some clues... Your a77 patch sounds nice..but when somebody knows why it sounds nice there are quite a few more possibilits
I ve another problem thats related to this and i ve to try the mentioned tricks...MAybe you have allready solved it (wouldnt be surprised
I open an own thread for it...its regarding early reflections..
greetz
Sven |
|
Back to top
|
|
 |
BobTheDog

Joined: Feb 28, 2005 Posts: 4041 Location: England
Audio files: 32
G2 patch files: 15
|
Posted: Sat Aug 27, 2005 12:35 am Post subject:
|
 |
|
Hi Rob,
Thanks for the clues, I'm going to have a look at it tommorow when I get some free time (hopefully!)
Cheers
Andy |
|
Back to top
|
|
 |
|