Inputs Question (soft clip, normalize)

nikki-k

nikki-k

Experienced
I am looking for information so as to better understand "what is happening" with the inputs on the Axe-Fx II please.

I could swear I read somewhere that input 1 (front) has a soft limiter? All inputs?

In the wiki, this appears:
All inputs are normalized, allowing the Axe-Fx II to operate with unity gain irrespective of the trim settings. The sole purpose of the input trims is to optimize A/D drive level for the best signal-to-noise ratio and distortion performance. Each input has its own dedicated A/D. The INSTR input jack parallels the rear inputs for improved signal-to-noise performance

In what manner are signals "normalized"? Could/would/does this result in any sort of "compression" effect (the term "normalized" would lead to me to believe it is (a?) dynamic effect of some sort, be it compression, companding, etc). IME, it feels and sounds (to me) as if there is some *compression-like* action occurring, especially in regard to the initial attack. Utilizing the Amp Block exacerbates this, more so with higher gain settings within the Amp Block.

I was about to experiment a bit through the use of various (physical) routings, and the use of a unity gain, shunted patch, and then with (at least one) enabled block. I then realized that if this information already exists, or could be provided (please!), I would not really need to do all of that... except for curiosity's sake ;)

Apologies if this has been asked and answered prior. I scoured the wiki, but could not find a complete answer.
 
Nikki,
as far as I understand this, It is a compensation at the A/D. So as input trim is increased, the A/D output is decreased, and vise-verse, to maintain unity gain. But haveing said that, I do notice a difference in input gain to the amp block. Maybe that's normal.
 
The unity gain terminology does seem to throw folks.

My take on that is, in regards to the AxeFx and the unity gain comment, the voltage level in is exactly the voltage level out on a preset with all shunts. No more no less. Turn up the input, voltage goes up at the output and vice versa. In this scenario the output level is a cut only. So the output level knob would have to be maxed to achieve the exact input voltage.

The only processing on the front input that I remember reading was a treble boost pre A/D and the exact opposite treble cut post A/D. But with the variable Z feature, I'm not sure if there is more special sauce in there or not?

Richard
 
Last edited:
"I could swear I read somewhere that input 1 (front) has a soft limiter? All inputs? "

Input 1 front does. This is just to keep the A/D converter from clipping (which sounds really bad). The soft limiter allows you to run pedals in front with out the fear of clipping the A/D converters.

What is meant by normalized and unity gain is that as the front input pre A/D converter is boost or cut with the input level, it is inversely boost or cut by an equal amount post A/D. It only goes so far, at extremely low input level settings, you will notice a drop in gain. At normal levels you should not. No compression or companding is used. It is NOT dynamic.

there is also additional processing in the front to lower the perceived noise floor. This is done sort of like dolby noise reduction (its exact method is proprietary - I imagine developed from Cliff's experience with SONAR systems). I believe more detail information can be found in the WIKI on this.

 
Last edited:
I find it pretty clever...Input gain does not alter incoming voltage at the output, it just trims the A/D. Boost/Pad does the same thing, or the opposite, at the D/A side: boosts digital level hitting the D/A and pads the output by the same amount.

javajunkie said:
there is also additional processing in the front to lower the perceived noise floor. This is done sort of like dolby noise reduction (its exact method is proprietary - I imagine developed from Cliff's experience with SONAR systems). I believe more detail information can be found in the WIKI on this.
Click to expand...


Unfortunately it doesn't work all that well to me...I lose quite a bit of top end in that front input. The rear is fine.
 
Trazan said:
I find it pretty clever...Input gain does not alter incoming voltage at the output, it just trims the A/D. Boost/Pad does the same thing, or the opposite, at the D/A side: boosts digital level hitting the D/A and pads the output by the same amount.


Unfortunately it doesn't work all that well to me...I lose quite a bit of top end in that front input. The rear is fine.
Click to expand...

I don't believe you have enough information to detemine that is the cause, especially since it is the same as the Ultra (which has been tested by a number of users as being flat).
LMO tested his on the II and it was flat.

You may have another issue. For all we know you could have an issue with the impedance selector which would cause that high end to drop like that and it would only happen on the front.
 
Last edited:
Trazan said:
Unfortunately it doesn't work all that well to me...I lose quite a bit of top end in that front input. The rear is fine.
Click to expand...

That doesn't sound right.

With all shunts, IIRC, you should get a ruler flat response using the front input.

You might have a hardware issue.

Richard
 
Trazan said:
Unfortunately it doesn't work all that well to me...I lose quite a bit of top end in that front input. The rear is fine.
Click to expand...

Probably unnecessary to ask, but you do switch between Front and Rear in I/O > Input Select?
 
javajunkie said:
He has went through all this in another post.
Click to expand...

Yes, I posted on it earlier. I checked all impedances, and they all sounded different. I chose the 1MOhm setting, which should be the same as the rear input. I of course chose the appropriate input, else I would not pass sound...

I'm comparing going straight guitar>amp to going guitar>AFX>amp. I'm curious if anyone else has, or can do the same test. If nobody get the same result (short cable, single coils into a crisp amp - to make things clear) then I must assume something's not right with my unit. At least don't tell me there's not an obvious high end loss in part 2! 8)

Part 1 - Axe rear input
Part 2 - Axe front input
Part 3 - straight guitar to amp

 
I was referencing freq response graphs that were showing the input to output frequency response as measured by some test software HarBal?

So I'm guessing your guitar -> amp, Axe -> amp test is apples and oranges to my comment about a possible hardware defect.

Richard
 
Trazan said:
Yes, I posted on it earlier. I checked all impedances, and they all sounded different. I chose the 1MOhm setting, which should be the same as the rear input. I of course chose the appropriate input, else I would not pass sound...

I'm comparing going straight guitar>amp to going guitar>AFX>amp. I'm curious if anyone else has, or can do the same test. If nobody get the same result (short cable, single coils into a crisp amp - to make things clear) then I must assume something's not right with my unit. At least don't tell me there's not an obvious high end loss in part 2! 8)

Part 1 - Axe rear input
Part 2 - Axe front input
Part 3 - straight guitar to amp

Click to expand...


Just because the impedance selection sounded different, does not mean there might not be a problem. I'm not saying that is the problem either. However, your assumption that is the processing in the front input is problem may not be correct.

Try this to remove the amp out of the equation (if you are isolating the problem exclusively to the front input) - input 1 front to out 1 of axefx, input 2 rear to front of axe-fx.
I will do the same.
 
Last edited:
javajunkie said:
Just because the impedance selection sounded different, does not mean there might not be a problem. I'm not saying that is the problem either. However, your assumption that is the processing in the front input is problem may not be correct.
Click to expand...

True. I only know that going through the AFX using the rear input sounds like going straight to amp, while the front input chops away the high end. Dunno why 8) I'll check some more.
 
Trazan said:
True. I only know that going through the AFX using the rear input sounds like going straight to amp, while the front input chops away the high end. Dunno why 8) I'll check some more.
Click to expand...
Just did a check to output 1. no real difference between front and rear. I'll post a clip in a bitI could get a difference if my input 1 mode was set to l+r, the front and input 1 level settings seem to interact with the post A/D boost/cut in that mode.But when set to left only the were the same as far as I could tell. Not like your clip for sure.Here is a clip of both front and rear:
 
Last edited:
Here's skipping everything but the AFX. Using internal amp sim (forgot to turn that reverb off!), recorded straight through USB. Playing a pattern twice on Input 1 REAR - Input 1 FRONT - Input 2 REAR...in that order. Input 1 and 2 set to L only. Short cable, bright pickups and amp - just to make the difference more obvious (believe me, it's much more noticeable at live loudness!). Top end sizzle attenuated somewhere on the way through Input 1 front.

 
javajunkie said:
Just did a check to output 1. no real difference between front and rear. I'll post a clip in a bitI could get a difference if my input 1 mode was set to l+r, the front and input 1 level settings seem to interact with the post A/D boost/cut in that mode.But when set to left only the were the same as far as I could tell. Not like your clip for sure.Here is a clip of both front and rear:
Click to expand...

Impossible to tell with your clip, Sean, since the sound doesn't contain those higher frequencies...
 
No offense or anything to anyone, but all of this does not answer the main question, and is distracting from it.

The questions remain unanswered:
Is there some type of normalization (as per the wiki, quoted in OP) in operation upon one or more of the inputs?
If yes, which inputs? And, in what manner is it performing normalization (method, and placement in the input chain)?
Is there (also?) a soft clip in place across one or more inputs?
If yes, same follow-up questions as for normalization.
Is there any other non-user-defeatable processing occurring?
If yes, what?

All of that with a huge Please? Pretty please? :D
 
nikki-k said:
No offense or anything to anyone, but all of this does not answer the main question, and is distracting from it.

The questions remain unanswered:
Is there some type of normalization (as per the wiki, quoted in OP) in operation upon one or more of the inputs?
If yes, which inputs? And, in what manner is it performing normalization (method, and placement in the input chain)?
Is there (also?) a soft clip in place across one or more inputs?
If yes, same follow-up questions as for normalization.
Is there any other non-user-defeatable processing occurring?
If yes, what?

All of that with a huge Please? Pretty please? :D
Click to expand...

http://forum.fractalaudio.com/axe-f...-question-soft-clip-normalize.html#post633963
 
nikki-k said:
No offense or anything to anyone, but all of this does not answer the main question, and is distracting from it.

The questions remain unanswered:
Is there some type of normalization (as per the wiki, quoted in OP) in operation upon one or more of the inputs?
If yes, which inputs? And, in what manner is it performing normalization (method, and placement in the input chain)?
Is there (also?) a soft clip in place across one or more inputs?
If yes, same follow-up questions as for normalization.
Is there any other non-user-defeatable processing occurring?
If yes, what?

All of that with a huge Please? Pretty please? :D
Click to expand...

1. I answered this. normalization is bad choice of words (in the sense that you are referring to). read my earlier post for what it does do.
2. I answered most of that earlier. Boost/cut pre A/D, Inversely applied post A/D. Look in the manual for the signal chain where the A/D converts. All inputs do this.
3. Yes, I answered this. There is a soft clip on the front input
4. I answered this as well. On front input 1 there is processing applied akin to dolby noise reduction (it isn't exactly that and is proprietary info). the sum effect is a lowered perceived noise floor.
 
Pink noise- that should cover any valid frequencies (up to the point mp3s cut them off):
Sorry for the horrible SNR, in a hurry

They match up with a spectrum analyzer.

Just realize the amp/cabs were on, on that. Here is with bypassed
 
Last edited:
You must log in or register to reply here.
Back
Top Bottom