About that 0.06mS Delay

GM Arts

Fractal Fanatic
Here’s the thread where Cliff proposed using a 0.06 mSec delay on one of two cabs, with a recommendation for using different cab models. http://forum.fractalaudio.com/cliffs-notes/91932-my-secret-realistic-cab-sounds.html

Cliff was pretty clear about what it does, but reading through some replies, it seems to have been lost. Even worse, some other threads seem to imply that this delay is now a mandatory requirement for good tone, and if I understand one post correctly, this delay is required alone on a single (mono) signal path!

So here’s what it actually does. Firstly, this is about mixing 2 signals: one without delay, and the other with a very short delay.

0.06mS is way too short to be perceived as a repeat; the effect is filtering caused by mixing these two signals. To keep things simple, we’ll apply an equal mix of the same signal and another delayed by 0.06mS. An easy way to experiment with this in the Axe-FX is with a Flanger block, with depth and feedback set to zero, and mix set to 50%. Adjust the delay to 0.06mS (not 0.6mS) to hear the effect with a mono signal.

This produces a notched frequency response with complete signal cancellation just above 8KHz, with the -3dB point one octave lower at just over 4KHz. The signal is restored over the next higher octave (8KHz to 16KHz), but bear in mind that most Cab IRs will not have much response there anyway, so this effect is mostly a blocking filter over the range 4KHz to 8KHz.

So if you have a cab IR that has some response over this range, it will be perceived as a loss of some treble response. For many, this will remove harshness in a way that’s difficult to achieve with other filters.

Others may find this effect too much. You can soften this effect by decreasing the delay and/or changing the mix ratio. Decreasing the delay raises the frequency at which this cut occurs. For example, a 0.05mS delay blocks response over the octave 5KHz to 10KHz. Lowering the mix % decreases the depth of the notch. Similarly, applying a delay to a different Cab IR than the un-delayed block will “jumble” and reduce the final response to some extent.

If you increase the delay (typically from 1mS and above), you’ll hear the combing effects as multiple notches become low enough to hear in the range of “guitar frequencies”. This sounds like a flanger or chorus without modulation, which shouldn’t be a surprise given we’re experimenting with a Flanger block.

So why does this delay sound produce a tone more amp-like? Most players prefer their amp tone off-axis, meaning that they’re avoiding the direct harsh sound directly in front of the speaker, where high-frequencies are beamed. This filter simulates that effect. It’s also similar to standing slightly off-axis when using multiple speakers. Sound travels at roughly one foot per millisecond, so there is a very short delay between sound from different transducers. As Cliff stated, it also emulates recording techniques with mics placed at different distances from the cab.

How to calculate? To find the frequency where this rolls-off high frequencies at -3dB, it’s simply: Hz = 1000 / 4 /delay in mSec . So for 0.06 mSec: 1000 / 4 / 0.06 = 4167Hz. Complete cancellation occurs at double this frequency, 8333Hz, and builds back to -3dB a double this frequency again, 16666Hz. Bear in mind that with higher delays, there will be audible effects from additional notches above this calculated frequency.

I hope this helps those who are interested.
 
I tried it with two ultra rez cabs (one with greenbacks, the other with d120's)...and of course a Marshall amp. Sounded great. Guess who I was tone chasing with that setup? :)
 
Good breakdown, GM Arts!

I'll add another reason this creates a more in-the-room feel: it simulates the comb filtering you get from multiple speakers in a cab, or from room reflections interacting with the direct speaker sound.
 
Wow. I say wow because I almost understood the OP the first read through. Thanks for that breakdown GM Arts :encouragement:
 
Interesting write up...I found it informative and useful.

Have been playing a lot with a dual cab/different IR's/delayed cab setup lately and like how you can really tailor the overall character blending two IR's with very small changes to the delay...sometimes a little strident/brighter works with a given IR combination and amp, at times a little more mushed out/darker setting is just the ticket...
 
You're sayin it works in mono as well, I don't need a stereo output? I don' t need two monitors for it to work? Is that right?
 
You're sayin it works in mono as well, I don't need a stereo output? I don' t need two monitors for it to work? Is that right?
This effect is strongest in mono. Or in stereo only if you're located exactly the same distance from each identical speaker in exactly identical surroundings. If you're moving around in the stereo field, all bets are off. ;)
As above, there are ways to make this less effective if it's too much in mono.
 
I gotta try this.

I didnt know there was a simple little delay knob that could do this, lol. I have tried before with delays and multiple amps, but nothing as simple as tweaking one knob.
 
Just a quick question - i tend to run my axe through either headphones of pair of monitors. I assume for this to work we are talking parallel cabs - each with their own path - would this work if the cabs path are then merged and then output in a single path
 
I might be forgetting something, but as I recall the reciprocal of the delay in mS equals the frequency in kHz or simply, 1/mS=kHz
 
just been goofing with that delay setting.. it's a really interesting effect..
the strength of it also seems to be more of less obvious depending upon the combinations of cab choices..
it seems to calm down bright cabs in very nice way..
it interacts quite interestingly with the motor drive too.. I like mine set up at 3 or 4 to make the cabs sound a little more like they're 'working'
 
I might be forgetting something, but as I recall the reciprocal of the delay in mS equals the frequency in kHz or simply, 1/mS=kHz

This is correct, Hertz used to be called "cycles per second", so "seconds per cycle" is the reciprocal of Hz. In your example, a single cycle at 1KHz takes 1mS.

In this case we're looking at how two waveforms combine when one is delayed. The notch frequencies occur where the waveforms are at opposite phases. So the first notch frequency occurs when the delayed signal is PI radians (or 180 degrees) out of phase with the original. This is the same as saying the time of half a cycle, so for 0.06mS, the calculation is 1000 / 2/ 0.06 = 8333Hz (the 1000 just converts mS to seconds).

A more useful measure is the -3 dB points, which loosely means "where the effect becomes clearly noticeable". This occurs where the waves are PI/2 (or 90 degrees) out of phase, with the first -3dB point at one quarter of a cycle. Hence the calculation 1000 / 4 / mS.
 
Yall should be with cliff building this stuff, all the numbers and hz,hg,ho,klh or what ever confuse. I think I will just turn knobs and enjoy the great sounds that all those numbers make in side my magical ax fx. thank you smart ones.
 
This is correct, Hertz used to be called "cycles per second", so "seconds per cycle" is the reciprocal of Hz. In your example, a single cycle at 1KHz takes 1mS.

In this case we're looking at how two waveforms combine when one is delayed. The notch frequencies occur where the waveforms are at opposite phases. So the first notch frequency occurs when the delayed signal is PI radians (or 180 degrees) out of phase with the original. This is the same as saying the time of half a cycle, so for 0.06mS, the calculation is 1000 / 2/ 0.06 = 8333Hz (the 1000 just converts mS to seconds).

A more useful measure is the -3 dB points, which loosely means "where the effect becomes clearly noticeable". This occurs where the waves are PI/2 (or 90 degrees) out of phase, with the first -3dB point at one quarter of a cycle. Hence the calculation 1000 / 4 / mS.

Now when you have 2 different cabs, things get interesting.
 
Back
Top Bottom