Rss

Eventide H3000 as a formant synth

Image copyright: John R. Southern

Here’s an interesting “Formant Morph” processor i’ve recently designed using the Patch Factory (aka algorithm 111). As you may know, Patch Factory algorithm as itself is not an effect, but more an effect designing mini studio which is why it unfortunately gets overlooked by many of the Eventide users, since (by default) it does not produce any special sound.

What we will build will be a primitive version of a vocal Formant filter. Proper vocal emulation requires at least three bandpass filters, while Eventide H-3000 unfortunately offers only two BPFs in the Patch Factory. Still, better something than nothing. Here is how it will sound on various settings:

External_In.mp3 – Here the effect is first applied to a saw wave. You can hear sample&hold morphing of the vowels as they shape “aaahh”, “eeehh” and the others. Then i processed the beat with the same settings. Then speaker’s voice with some variation on the settings.

Stereo_Version.mp3 – More exotic sound can be obtained by building a stereo effect using the second delay line. See the bottom of the text for this optional mod. Listen to this one on the headphones.

Stand_alone.mp3 – If you will use noise generator included with the patch factory, here are some of the effects you can achieve by doing small variation on the parameters. Much more can be created of course, only limit is your imagination.

Building the structure

First thing to do, once you load algorithm 111 is to go to Expert parameters (via Parameter button) and choose Patching. The goal in here is to patch everything to Null. Goes very fast, you don’t need to dial the value, just type 99 and press enter, this will select Null Input and press Parameter button until you browse the full circle through patch destinations (Filt1 In, Filt2 In, Delay1 In…etc.). At this point everything is disconnected. Now we will build the structure:

  • Browse via Parameter button until to get to the Scale1 In and patch a Noise Gen to it by selecting the Noise Gen as a Patch Source. IMPORTANT: If you look at the picture below, you will see that it shows a little bit different situation – instead of the noise, left input is connected, however that is how the final product should be. But at this point, you should connect the Noise Gen instead! This is for calibration purpose that we will talk about later. Scaler is also a very important element and please don’t omit it, just because you can do it without the scaler. Trick is about the correct levels which will later have to be dialed on the Scaler1, to get the correct harmonic enhancement for our formant shaper without going into distortion. This is why we patched the loud white noise source. Using it, you will calibrate the scaler later on. For now, forget about this.
  • Now it’s time to split the signal into two filters. Go to Filter1 In page and select Scaler1 for the patch source. Repeat the same with Filter 2 In.
  • Browse now to the Sum 1a In and select Bandpass1. Browse to Sum 1b In and select Bandpass2. This is the core of our effect. Two bandpass filters have just been connected into series.
  • We will spice our effect with one short delay line to get a nice flanging, which always sounds killer on Eventide machines. For that job we will need to split the signal again. First go to Delay1 In and select Sum 1 as its patch source. Then go to Sum 2 to join the delayed signal with original by doing the following: Sum2a In: Sum1 and then Sum2b In: Delay 1.
  • IMPORTANT! Prior to doing the next step please reduce the level of your amplifier or headphones to MINIMAL! Serious hearing damage might happen to you and you might damage the speakers as well!! This is not a joke! A lot of loud noise might be produced at this stage!
  • Now connect select L Output and choose Sum 2 as its source. The Sum2 is now connected to the output and we’we built the structure.
  • Its time to properly set each of its parameters. You will do that by exiting the Expert function.

Back to the Parameters now. We have to do some calculation first. Our goal is to build a formant morph that will go in between two vowles: “Eeeh” and “Oooh”. As said earlier, we only have two filters at our disposal, so we will use the upper two peak frequencies (out of three). Quick google shows us the following:

“Eeeh” 3010Hz and 2290Hz as its peaks
“Oooh” 2410Hz and 840Hz as its peaks

Since the LFO modulation is alternating between positive and negative values we need to calculate 1)the center frequency from which we will apply the LFO to sweep between the 3010Hz and 2410Hz; and 2)the LFO amount deviation needed to reach these two values:

Peak 1 morph calculation
3010 + 2410 / 2 = 2710Hz center frequency; from which
2710 – 2410 = 300Hz LFO modulation amount

Peak 2 morph calculation
2290+840 / 2 = 1565Hz center frequency; from which
1565-840 = 725Hz LFO modulation amount

Setting the Parameters

  • Go to Scale 1 and set it to 15%. (now you can increase the headphones / speaker volume).
  • Go to Parameter Cutoff1 and set it to 2710Hz. Go to Cutoff2 and set it to 1565Hz. This sets the filters right in the middle between the upper and lower points of the two vowels.
  • Go to the Q Factor1 and set it to 950, and set Q Factor2 to 950.
  • Go to the Delay 1 line and set it to 0.40ms. This is the center frequency from which our flanger will operate.
  • Now we’re done with the parameters.

Setting the Modulation

  • Push the FUNCTION button twice. This gets you into the modulation and LFO page.
  • Chose FuncGen and then select Function from the menu. Select Sample & Hold from there. Set Rate to 4Hz and Amount to 100.
  • Now select Patch from the Modulation of Parameters screen.
  • Turn the main knob until you select delay 1 and set its source to Function Generator. Press Done. Set range to -0.30. This will sweep the flanging effect from 0.10 to 0.70ms, since we previously set the center of the delay at 0.40 miliseconds.
  • Now we just need to input the LFO amount for the formant filter. Go back to Patch again and turn the knob until you select cutoff 1. Move the knob wheel and set its source to Function Generator. Press Done. Set the range to 300.
  • Repeat the same but chose cutoff 2 now and set the range to 725

At this point you should already hear sample&hold doing “formant synthesis” to the white noise source and vowels going out of your speakers (or what’s left of the vowels, given they’re made of just two bandpass filters). Only thing left to do is to replace Noise Gen with the Left Input at the Scaler1 patch. You will do that by going back to expert settings via Parameter button and press it until you see Scale1. Then just select Line Input as its Patch Source.

Now it’s time to go to your console and route Aux into the Left Input. This is where you will have to “calibrate” the Scaler1 to avoid clipping and distortion. Simply go to Parameter settings and adjust Scaler1 in between 1-100% volume, depending on your console’s or sound source’s volume. You will notice distortion as unpleasant artifacts, each time a peak hits into the Eventide. Most simple is to just set the scaler to 100% and regulate the amount of volume being sent into Eventide via Aux pot at the console. However (!) for white noise effects (stand alone) you will have to set the Scaler1 to the low volume, else the loud noise will start distorting once it enters the resonant bandpass filter. This is why i’ve said Scaler is important. In fact will will even add a soft function for its control, so that you don’t have to enter the patch settings every time you want to readjust it.

Stereo fun and Soft Function

For more stereo width and to put some exotic vibe into our effect, we will make its stereo version. First thing to do is to go to Expert settings and enter the Patching.

  • Got to delay2in and select sum 2.
  • Got to r output and select delay2 there.
  • Press Return to go back to standard parameters and navigate to the delay 2 and set it to 20ms.
  • You should already hear chorus like stereo width that has been just produced. However we will do some more fun now.
  • Go to the Functions module (press function button twice) and go to Patch.
  • Turn the knob until you select delay 2.
  • Select Function Generator as its Source and set range to 10.00. Done.
  • Last step is to add the soft function the the scaler. First go to Parameter settings and set scale 1 to 0% (at this point formant synthesizer should become silent).
  • Press Function button and select SoftFunc.
  • Press Name and rename it to Preamp. Set Sensitivity to 100 and Pol to +. Press Done.
  • Press Function button and press Patch. Select scale 1 with the knob and press Source. Select Soft Function 1 “Preamp” and press Done. Set range to 100.
  • Save the patch i.e Formant Synth.
  • This completes our job.

Tip: To match the beat you’ll need a calculator (not all the beats can be matched). For 120 bpm, (which is essentially 2Hz) you want to chose values that are either multiples or the same as the beat frequency i.e. 2.00Hz, 4.00Hz, 8.00Hz, etc. or you can halve it to 1.00Hz or 0.50Hz. Enjoy the fun! If you have any tips, feel free to add them to the Eventide H3000 Club thread.

If your sequencer slightly gets out of the beat for a small amount, you can go to the Function Generator (via Function button) and simply press the Trigger at the beat start. This should correct the problem.

Comments (3)

  1. Hello Don, I liked a lot your article about H3000. I own one. I bought for basic reverb, flanger, delay tasks but recently I passed my voice through preset… If I remember well was “319 Gregorian Chant” and I felt was something similar to the effect used on many old acid house records like “Bam Bam – Where is your child”. My question is, do you think with the eventide can I easily recreate that voice fx or is better to use someting else. I always loved that effect but never found information on the net about how to reproduce it.
    Thank you for your time.
    Cheers
    Edgar Boada

  2. Mark

    Hi, Don, I wonder what you would not do with a yamaha FS-1 and E-MU Morpheus!

Leave a Reply

Your email address will not be published. Required fields are marked *