Keeley Halo

I have my Halo in a loop so I can quickly A/B.

Pretty close, but not identical, the repeats fades into "darkness" faster and more abrupt than the pedal, so you don't get the same reverb-y effect as from the pedal.

But definitely the closest I've heard so far.
You have good ears mate! I had to make a compromise regarding the high cut frequency. Even the 6 dB/oct is a bit too steep, so I had to pick between getting the first repeats right, or the tail. I chose the tail to get the washy effect more accurate. Bump up the high cut 100-200 Hz if you want the first repeats to more closely match the pedal, at the expense of a bit too bright tail.

If the Q parameter was active for the 6 dB/oct slope, it would probably be possible to get the degrading of the highs a bit closer.
 
I have a HALO here and was fascinated by this thread - and very grateful to @guitarnerdswe. So I've made a little demo vid comparing the hardware and the chained delay blocks.


The preset used can be downloaded from http://axechange.fractalaudio.com/detail.php?preset=9751

I like that you noticed the same thing as I did in terms of the "body" of the repeats. There is most likely some EQ on the repeats that's not in the feedback loop (since the added frequencies don't get exaggerated with each repeat). I don't know how to do that on the Axe, because running the delays in parallel to the dry ruins the whole cascading feature of the Halo effect. It would need to be a separate EQ inside of the delay architecture, but outside of the feedback loop.

Here's a slightly different take on it. I've used Stereo Memory Man blocks instead of Stereo Tape Echo. I've applied most of @guitarnerdswe's settings where applicable and tweaked anything else to ear. The end result isn't miles away from @guitarnerdswe's version but to my ears sounds a little warmer. I'm pretty sure I remember Robert telling me early on, during the HALO's final prototype stage, that the Memory Man was involved, hence my attempt here. But before @guitarnerdswe's amazing work, I was trying to replicate the HALO with a single block, and gave up. But now... 😅

Anyway, string these together and see what you think. The preset is here as well. Memory Men on Scene 3. You'll need your own HALO to complete Scene 1.

I can make another video if you think it's worth it.
Oh, nice idea to try the stereo Memory Man! That might be the ticket to get a bit more of that character that's missing. Even with the compander circuit off, the tone is quite different between the 2 types for the same settings. There is just something a bit more "full" sounding about the stereo Memory Man. And down the rabbit hole we go again :sweatsmile:
 
Last edited:
I have my Halo in a loop so I can quickly A/B.

Pretty close, but not identical, the repeats fades into "darkness" faster and more abrupt than the pedal, so you don't get the same reverb-y effect as from the pedal.

But definitely the closest I've heard so far.
Agree that the repeats get darker faster in the Axe version.
If that's a 6db rolloff, maybe try none, but roll off a little before going into the delays.
To do that without affecting the dry, I think you'd have to put the delays in parallel, and EQ that line only.
Not at my axe, and haven't tried any of this, so I may well be out where the trains don't run...
 
Oh, nice idea to try the stereo Memory Man! That might be the ticket to get a bit more of that character that's missing. Even with the compander circuit off, the tone is quite different between the 2 types for the same settings. There is just something a bit more "full" sounding about the stereo Memory Man. And down the rabbit hole we go again :sweatsmile:
Assuming the Memory Man presets are accurate, that would introduce a completely different EQ curve to the repeats. Not saying either one is right, but I suspect you'd get a different character than the Halo. FWIW, I think the EQ you used sounds better. One of the things I tried was two mono Memory Man pedals into stereo cabs, which is what his actual rig was before the Timeline, but I couldn't get that to sound right.
 
So first I have to say these are amazingly good, way better than what I came up with. Have a couple questions/comments if you don't mind indulging me.

Just re-watched the Keeley video and he says that the modulation cross fades between the two channels. Is that what's happening in your blocks? Both the first and second delay block have LFO1 set to target Both with a phase of 90 deg. Would you not want to use both LFO 1 and LFO 2 with each targeting a different channel and a 90 degree phase shift between them?

Regarding diffusion, I feel like this gets mentioned in passing on some of the Halo videos. I liked the sound of a small amount of diffusion with Diffusion Time around 70% and a bit of modulation. The blocks guide says that adding modulation to the duffuser gives a chorus-like effect, which Andy mentions in the Keeley video.

Last is just a comment that I liked the EQ settings you used much better than what I had. I was using a Memory Man like EQ curve with a sharp roll off. I was finding it difficult to make the first delay repeat sit correctly in the mix. You used the 6 dB slopes which are the default on the tape delays. Any idea if the actual pedal is doing something similar?

EDIT: threw this on a patch with a few different variations. I dialed this in in a Strat, neck pickup with the volume rolled off a bit. Scene 1 is the blocks just as above except mix turned up. Scene 2 used LFO1 and 2 as described above. Scene 3 adds a bit of diffusion with modulation to the mix.

I think the issue here is that the Dual Delay is defining one delay time as either the L or R channel. The cross feedback ends up feeding back into the dotted 1/8th, which isn't what happens in the real pedal (or the old Memory Man setup). There was "too much" going on there, which is I think why I started throwing diffusion at it.

Not that I'm doubting what you're saying, but do you have more info on how the phase offsets work? I've found that things like the Blocks Guide and the Wiki tend to be pretty short on technical details, especially on more advanced parameters like this. The way I was interpreting this, it is applying a 90 phase offset to "both" channels. You're saying that there's an additional 90 already baked in there on the R channel? To my ears at least, when I set LFO1 at 90 deg targeting L and LFO2 at 180 deg targeting R, I heard a much wider sound.

Re: diffusion...after using the stereo blocks in series, I no longer felt the diffusion sounded right. I think it was just a Band-Aid I was using on a poor solution.
Regarding the LFO crossfade: When Robert Keeley says that in the video, I'm pretty sure he just means that the modulation isn't the same for left and right. IOW, a phase offset.

When a LFO is assigned to "both" on a stereo delay in the Axe-Fx, the phase parameter only controls the modulation for the right side delay. When using a phase offset of 45-90 degrees, you get this almost panning illusion of the repeats, since one side of the delay is always chasing the other in terms of modulating to a longer or shorter delay time (which is what modulating a delay does). This affects our perception of where in the stereo image it sits. It's more obvious with a chorus, but you can hear it on the Halo too. The repeats "cross" the stereo image at intervals consistent with the modulation rate. So yeah, the LFO "crossfade" is baked into my blocks the way they are set up.

I'm fairly certain that the Halo uses only one LFO for everything (so no secondary LFO modulating at a faster rate, for instance). That LFO might have two or four outputs. In the case of two, both left side delays are modulating in unison, and both on the right side are modulation in unison, but with a phase offset in relation to the left side. If the LFO has 4 outputs, it might be something like 0, 90, 180, and 270 degrees. Since we can't sync the LFOs in two delay blocks in the Axe-Fx anyway, it's kinda pointless to go to far down that particular rabbit hole.

Regards the diffusion: Yeah, I'm totally onboard with your final conclusion: There is no diffusion on the Halo. Some describe diffusion as sounding a bit like a reverb, and some a bit like a chorus. I would say it's both, and neither. Diffusion is a transient schmear, and it does this by adding microdelays to diffuse the sound. It can sound like anything from a static flanger, to a big washy grainy reverb. When done like Lexicon does it (the most famous example of diffusion I can think of is the PCM70 Circular Dlys), it sounds almost fluffy. If a normal delay repeat is a bare fisted punch, adding just the right amount of diffusion is akin to covering the same fist in a big bag of cotton. FWIW, diffusion this way is normally done with the mix at 100 %, and the only control is the diffusion time.

EQ: 6 dB/octave was the closest I could get with that particular delay type. As I mentioned in a previous post, there is something more going on with the EQ that's just missing from the stereo tape type. The high cut is very shallow, and there is probably some EQ going on that's not part of the feedback loop. Like a static low mid bump or something outside of the feedback loop.
 
Assuming the Memory Man presets are accurate, that would introduce a completely different EQ curve to the repeats. Not saying either one is right, but I suspect you'd get a different character than the Halo. FWIW, I think the EQ you used sounds better. One of the things I tried was two mono Memory Man pedals into stereo cabs, which is what his actual rig was before the Timeline, but I couldn't get that to sound right.
The analog delay types have a different baked in character, and respond differently to changes in the EQ section. Even if you set them up the same as a digital delay type, they will sound completely different. His setup with the Memory Mans was completely mono though, so that would explain why you couldn't get the halo thing going.

Agree that the repeats get darker faster in the Axe version.
If that's a 6db rolloff, maybe try none, but roll off a little before going into the delays.
To do that without affecting the dry, I think you'd have to put the delays in parallel, and EQ that line only.
Not at my axe, and haven't tried any of this, so I may well be out where the trains don't run...

I've been scratching my head as how to do that. I can't see a way of making that happen, without Cliff adding a second EQ section inside the delay block, that's either pre or post feedback. But I like the idea. Putting the delays in parallel won't work though, without having a just a mess of a signal flow, like this:

Namnlös.png

You need to have the dry signal flow there between the delays, to get the cascading thing going.
 
The analog delay types have a different baked in character, and respond differently to changes in the EQ section. Even if you set them up the same as a digital delay type, they will sound completely different. His setup with the Memory Mans was completely mono though, so that would explain why you couldn't get the halo thing going.



I've been scratching my head as how to do that. I can't see a way of making that happen, without Cliff adding a second EQ section inside the delay block, that's either pre or post feedback. But I like the idea. Putting the delays in parallel won't work though, without having a just a mess of a signal flow, like this:

View attachment 104997

You need to have the dry signal flow there between the delays, to get the cascading thing going.
Still not at my Axe, but do you need to join the first delay back with the dry path before the second delay?
Could it just be a parallel line with PEQ > Delay > Delay then merge that back with the dry?

EDIT
I haven't studied the Halo at all beyond what I've seen here, and not even that in serious detail, but if need be the first delay could pass some dry to the second one, if that helps.
 
I think the issue here is that the Dual Delay is defining one delay time as either the L or R channel. The cross feedback ends up feeding back into the dotted 1/8th, which isn't what happens in the real pedal (or the old Memory Man setup). There was "too much" going on there, which is I think why I started throwing diffusion at it.
You need to set Feedback R->L to 0%. Feedback L->R would be set to a ratio that is similar to setting the mix level in the first delay block.


Not that I'm doubting what you're saying, but do you have more info on how the phase offsets work? I've found that things like the Blocks Guide and the Wiki tend to be pretty short on technical details, especially on more advanced parameters like this. The way I was interpreting this, it is applying a 90 phase offset to "both" channels. You're saying that there's an additional 90 already baked in there on the R channel? To my ears at least, when I set LFO1 at 90 deg targeting L and LFO2 at 180 deg targeting R, I heard a much wider sound.
It's actually described in the blocks guide:

"LFO1 Phase, LFO2 Phase – Sets the LFO phase offset for the right delay line. This has no effect on Mono
delays (including Reverse, Tape, etc.)"
 
I've been scratching my head as how to do that. I can't see a way of making that happen, without Cliff adding a second EQ section inside the delay block, that's either pre or post feedback. But I like the idea. Putting the delays in parallel won't work though, without having a just a mess of a signal flow, like this:

View attachment 104997

You need to have the dry signal flow there between the delays, to get the cascading thing going.
Eh I wouldn't necessarily call it a mess - it's what you would have to do today with the existing blocks.

Going off what Keely said last Monday in the demonstration session I attended, there's a TiltEQ in addition to the high pass filter in the halo. So maybe a Filter block instead of the PEQ set to Tilt EQ?
 
Still not at my Axe, but do you need to join the first delay back with the dry path before the second delay?
Could it just be a parallel line with PEQ > Delay > Delay then merge that back with the dry?

EDIT
I haven't studied the Halo at all beyond what I've seen here, and not even that in serious detail, but if need be the first delay could pass some dry to the second one, if that helps.
Yes, unfortunately, you have to do it like that.
Eh I wouldn't necessarily call it a mess - it's what you would have to do today with the existing blocks.

Going off what Keely said last Monday in the demonstration session I attended, there's a TiltEQ in addition to the high pass filter in the halo. So maybe a Filter block instead of the PEQ set to Tilt EQ?
And even bigger of a bummer. I've just confirmed that most (if not all) of the EQ happening on the Halo, is done outside of the feedback loop... Which means that the signal flow I posted, is the only way to get those dark repeats that don't degrade. The high pass filter might be in the loop, but the big bass boost and high frequency attenuation is not (the tilt EQ you mentioned).
 
Still not at my Axe, but do you need to join the first delay back with the dry path before the second delay?
Could it just be a parallel line with PEQ > Delay > Delay then merge that back with the dry?
I mean that would involve using a mix less than 100% in the first block so that some dry signal gets into the second delay - which then becomes harder to coordinate (hard to keep a constant dry level).
So I think it's far easier to have both delays as @guitarnerdswe describes and control the wet level independently for each.
 
When a LFO is assigned to "both" on a stereo delay in the Axe-Fx, the phase parameter only controls the modulation for the right side delay. When using a phase offset of 45-90 degrees, you get this almost panning illusion of the repeats, since one side of the delay is always chasing the other in terms of modulating to a longer or shorter delay time (which is what modulating a delay does). This affects our perception of where in the stereo image it sits. It's more obvious with a chorus, but you can hear it on the Halo too. The repeats "cross" the stereo image at intervals consistent with the modulation rate. So yeah, the LFO "crossfade" is baked into my blocks the way they are set up.
I see this now, very odd to me. There are several examples in this thread where parameters either have no effect or don't have the implied effect. I've always taken it on faith that if the parameter is there in Axe-Edit, it does what it says. If the phase offset has no effect in mono delays and only affects the right channel in stereo delays, why have the option of selecting L, R or BOTH as targets?
And even bigger of a bummer. I've just confirmed that most (if not all) of the EQ happening on the Halo, is done outside of the feedback loop... Which means that the signal flow I posted, is the only way to get those dark repeats that don't degrade. The high pass filter might be in the loop, but the big bass boost and high frequency attenuation is not (the tilt EQ you mentioned).
Finally got a chance to listen to @Burgs video. This completely makes sense. The repeats with your EQ setting sit better but the initial echoes are too dark. But opening up the EQ to brighten up the initial repeats messes up all the long repeats. I'm sure you did but what about a higher cutoff with a steeper slope? Another idea I couldn't figure out how to implement was some sort of detune tied to an LFO, sort of like a synth.
 
I see this now, very odd to me. There are several examples in this thread where parameters either have no effect or don't have the implied effect. I've always taken it on faith that if the parameter is there in Axe-Edit, it does what it says. If the phase offset has no effect in mono delays and only affects the right channel in stereo delays, why have the option of selecting L, R or BOTH as targets?

Finally got a chance to listen to @Burgs video. This completely makes sense. The repeats with your EQ setting sit better but the initial echoes are too dark. But opening up the EQ to brighten up the initial repeats messes up all the long repeats. I'm sure you did but what about a higher cutoff with a steeper slope? Another idea I couldn't figure out how to implement was some sort of detune tied to an LFO, sort of like a synth.
Yeah, I agree that is a bummer. I've learned the hard way to fiddle every knob and just see what it does (and IF, it does anything). The assign function does work as intended with stereo delays. It's the phase parameter that only affects the right side. But yeah, I don't really know why assign and phase are there on a mono delay.

I've uploaded a new version with the EQ after the delay, which solves the issue with the degrading delay repeats.
 
Can someone ELI5 whats so special about the Keeley Halo? Lol im ignorant when it comes to delays usually just a wet delay to thicken up my lead sound
 
Skimming through this really interesting thread, reaffirms a consistent bummer I always seem to have with the Axe FX delays in general.. I would like more or wish for more control over the volume / tone of each repeat…
 
Can someone ELI5 whats so special about the Keeley Halo? Lol im ignorant when it comes to delays usually just a wet delay to thicken up my lead sound
2 very dark stereo delays in series, with stereo modulation creating the stereo effect/spread. The EQ is also outside of the feedback loop, which means the repeats don't degrade with each repetition like with normal delays.
 
I've uploaded a new version with the EQ after the delay, which solves the issue with the degrading delay repeats.
Thank you so much for revised version @guitarnerdswe. I tried to create something similar a few days ago, but this is just so much better.

Your efforts are very much appreciated. I love it!
 
Last edited:
Back
Top Bottom