Better Noise Gate algorithm

[BrianCarroll]

IMO the noise gate is the only weak link in the AxeFX - Cliff mentioned in a topic that he thought about using the Gate/Expander block algorithm which is better. I'd really like to have a better noise gate even if it's a little bit more CPU hungry. CPU usage has been vastly improved in the last firmwares, I guess that the extra CPU could be use for this...

The Axe-Fx has no "noise reduction". If he means the noise gate, then, admittedly, it could be better. The dedicated Gate/Expander block uses a better algorithm and I debated using that algorithm for the global gate but it uses more CPU so I decided against it.

from http://www.fractalaudio.com/forum/viewtopic.php?f=16&t=7019&start=10

 

[Secret80'sMan]

Brian,
Have you tried the Gate/Expander block? I haven't but if you try it and have somewhat better results than all is well no?

 

[BrianCarroll]

AxeFX Standard here, not the Ultra - No Gate/Expander for us.

 

[Beefcake]

Try "the Clawfinger technique": put a vol/pan block where the gate would go. Then set the envelope to control volume. With the lowest value at 0 and mid + high at almost the same level near the top, it works great, great, great for me. Fiddle with the slope value for the attack / release curve and I bet you'll like it too.

A big thanks to Clawfinger for bringing up this nifty trick, I have foregone using the Axe-FX gate and exclusively use this technique. This "gate" is also routable, so no more gate eating the sustain of my cleaner sounds - it's only where I want it, before the HI GAINZ.

It works so great, I voted no since it would waste memory. Unless, maybe, a super light block could be added with this exact functionality and nothing else, maybe even one that processed in mono to save CPU?

 

[BrianCarroll]

Try "the Clawfinger technique": put a vol/pan block where the gate would go. Then set the envelope to control volume. With the lowest value at 0 and mid + high at almost the same level near the top, it works great, great, great for me. Fiddle with the slope value for the attack / release curve and I bet you'll like it too.

A big thanks to Clawfinger for bringing up this nifty trick, I have foregone using the Axe-FX gate and exclusively use this technique. This "gate" is also routable, so no more gate eating the sustain of my cleaner sounds - it's only where I want it, before the HI GAINZ.

It works so great, I voted no since it would waste memory. Unless, maybe, a super light block could be added with this exact functionality and nothing else, maybe even one that processed in mono to save CPU?

Yeap, it's actually what I do, but I still think that the AxeFX deserves a better dedicated noise gate.
And adding the Vol/Pan block + Env. controller adds CPU load, maybe more than a new algorithm - Only Cliff can tell.

Thanks Beefcake.

.
.
.
Two thumbs up Jocke !!!

 

[FractalAudio]

If you want the better noise gate it's around a 2% increase in CPU usage. So factor that into your responses.

 

[BrianCarroll]

Thank you Cliff, since the Vol/Pan block + env. cont. is something like 1%, a better noise gate would be just 1% more than "the Clawfinger technique".

IMO it's worth it.

 

[FractalAudio]

I guess I have to ask "What is wrong with the noise gate?". I just tried the Clawfinger technique and it sounds the same to me.

I hear people complaining but no one has spelled out what the problem is.

 

[mitch]

I voted yes...of course it's a little easier to sacrifice some cpu for ultra users. Would this cpu hit be in effect at all times or only when the gate was engaged? I don't use a gate all that often, so would be good if I could reclaim some cpu when it's bypassed.

 

[Beefcake]

And for me, since "TCT" works pretty much perfectly and is routable - and I don't typically have a use for a non-routable gate anymore because of how I build my patches - it isn't worth it. More power to everything else.

(Just my two cents, of course.)

 

[Beefcake]

I guess I have to ask "What is wrong with the noise gate?". I just tried the Clawfinger technique and it sounds the same to me.

I hear people complaining but no one has spelled out what the problem is.

At least for me, it's fast high gain palm mutes. The gate just doesn't cut them out fast enough, there's a quick fading noise after each note and it sounds slightly like butt and detracts from the feeling of chunking along. I find "TCT" a lot more satisfying in this regard. I suspect Clawfinger thought the same.

For anything else, IMHO, the stock gate is just fine, but then again I don't use gate for much else, like I said.

 

[porieux]

My ideal scenario would be to not need a noise gate at all. How about a balanced connection between guitar and Axe-FX? I always wanted to try one of those Les Paul Recording guitars :ugeek:

 

[BrianCarroll]

At least for me, it's fast high gain palm mutes. The gate just doesn't cut them out fast enough, there's a quick fading noise after each note and it sounds slightly like butt and detracts from the feeling of chunking along. I find "TCT" a lot more satisfying in this regard. I suspect Clawfinger thought the same.

For anything else, IMHO, the stock gate is just fine, but then again I don't use gate for much else, like I said.

+1

 

[hp/mp]

Try "the Clawfinger technique": put a vol/pan block where the gate would go. Then set the envelope to control volume. With the lowest value at 0 and mid + high at almost the same level near the top, it works great, great, great for me. Fiddle with the slope value for the attack / release curve and I bet you'll like it too.

Nice! I raised the release in the Envelope too to hear fading delay/reverb.

However when I hit a (distorted) chord dry, the milliseconds afterwards when it fades, you can still hear a glimpse of overdrive noise.. Or when I turn down the guitar volume knob and play with a gain patch, the backround fizz is very apparent. Is this just going to be an inevitable effect of gain? I tried tweaking the slope, threshold, and release but can't get it to where it can fades perfectly - where it's not too awkwardly short and cuts off effects, and yet no creeping overdrive artifacts. I get close though, and is better than the global gate. Maybe I have to use the global gate combined with this technique? Or a gate block?

Care to post your settings?

 

[javajunkie]

If you want the better noise gate it's around a 2% increase in CPU usage. So factor that into your responses.

No, I don't want 2% increased cpu usage unless it can be shut off. I already use the gate block, that would be a double whammy.

I use the SC select on the gate block if I use a gate.

It sounds like a lot of people just want to be able to turn the release time lower than a better algorithm?

 

[m lebofsky]

If you want the better noise gate it's around a 2% increase in CPU usage. So factor that into your responses.

Improve the noise gate, but when users set the threshold at minimuman unload the gate to free-up CPU resources. YES ?

 

[javajunkie]

If you want the better noise gate it's around a 2% increase in CPU usage. So factor that into your responses.

Improve the noise gate, but when users set the threshold at minimuman unload the gate to free-up CPU resources. YES ?

personally, I wouldn't have a problem with that. So long as my ambient cpu usage doesn't go up.

 

[dk_ace]

I don't want the increase in CPU usage unless it can be turned off. Resources are already getting tight, and I don't think a 2% increase across the board is worth it unless virtually everyone wants it. If there was no workaround it would be one thing, but seeing as you guys already have a way to achieve the response you want I don't think it's worth it.

D

 

[mranonimouse]

If you want the better noise gate it's around a 2% increase in CPU usage. So factor that into your responses.

No, I don't want 2% increased cpu usage unless it can be shut off. I already use the gate block, that would be a double whammy.

I use the SC select on the gate block if I use a gate.

It sounds like a lot of people just want to be able to turn the release time lower than a better algorithm?

I'm in the same boat as JJ, except I don't use the gate at all so I don't want a 2% increase to ambient CPU use.

Apologies to those who want the gate!

 

[mortega76]

Personally I would love a gate that used up all my cpu, forcing me to buy an amp, cab and fx processor... But at least it will be an awesome $2000 noise gate!