Input Clipping Thread

[flitz]

Clip warning at around 68%/70% with hard strumming and DiMarzio Tone Zone/bridge, DiMarzio PAF Joe/neck.

 

[State of Epicicity]

Well, when I first installed the new beta I first found that I had to set my Saturday Night Specials, set pretty low, to around 47%, but after reading more people's experiences, I tried again hours later (I had been playing the whole time, so the unit was on), and, with the same playing-as-hard-as-I-can, I was able to bump it up to the 70s. It made me wonder if it could be change as the unit is on for a few minutes. I.e., if maybe there are background things still running at first that make the clip warning inaccurate if you try it when first booted. I don't know, and I don't know when can play next to be able to test.

 

[jamesmarshall]

The new input clip warning should, based on the way clipping is described in the manual, show up when you're pushing 6dB higher than tickling the red. I know for sure I was not at that level when I was playing. On the other thread I referenced, @mr_fender and @GlennO recorded signals to test. For Glenn's, he only tickled the red and found the input peaks flattened; I'm not sure of Fender's LED meter with his clip. If that's right, than the original guidance that tickling the red still have you more room was wrong. I'm going to test myself when I can find time to play again.

I think there is a huge misunderstanding here!

Tickling the red is the input gain and not the A/D sensitivity. Earlier in the thread here I copied the section from Yek's guide that explains the A/D sensitivity. This is not a "gain" setting but it can affect gain if set way too low (like setting it to 0.5% on a PAF, try it). This is just setting the A/D sensitivity or how much you are sending to the A/D chip, which you do not want to clip. It is safer to set this lower for your highest output pickups (say EMG, BKP, etc.) and leave it since the rest of your guitars will be fine with it at the same setting since this is not gain.

 

[Fixated]

On my Suhr modern terra with the stock Suhr pickups I have to set the input to 6% and can still get clipping.

My Ibanez with bare knuckle aftermaths has to be set below 5%.

 

[AlGrenadine]

On my Suhr modern terra with the stock Suhr pickups I have to set the input to 6% and can still get clipping.

My Ibanez with bare knuckle aftermaths has to be set below 5%.

Axefx mk1 or mk2 ?

 

[Fixated]

Axefx mk1 or mk2 ?

Mk2

 

[State of Epicicity]

I think there is a huge misunderstanding here!

Tickling the red is the input gain and not the A/D sensitivity. Earlier in the thread here I copied the section from Yek's guide that explains the A/D sensitivity. This is not a "gain" setting but it can affect gain if set way too low (like setting it to 0.5% on a PAF, try it). This is just setting the A/D sensitivity or how much you are sending to the A/D chip, which you do not want to clip. It is safer to set this lower for your highest output pickups (say EMG, BKP, etc.) and leave it since the rest of your guitars will be fine with it at the same setting since this is not gain.

Thanks James. I just reread the wiki link in your first post in this thread. Are you saying that the LED red is not an indicator of where the A/D chip clips? It almost seems like you're saying the only indicator of where the A/D clips is the input clip warning. That just does not sound right at all. I know the input gain is something different, where you could, eg, use that parameter in the input block to match
the output level of a standard Strat to that of a Les Paul to make a preset more usable.

 

[Patzag]

There may also be differences in measurement form. Possibly the clip indicator comes on when there is even a single sample above -0.5dB, whereas the LEDs measure RMS. This would explain the difference.

 

[GlennO]

The new input clip warning should, based on the way clipping is described in the manual, show up when you're pushing 6dB higher than tickling the red. I know for sure I was not at that level when I was playing. On the other thread I referenced, @mr_fender and @GlennO recorded signals to test. For Glenn's, he only tickled the red and found the input peaks flattened; I'm not sure of Fender's LED meter with his clip. If that's right, than the original guidance that tickling the red still have you more room was wrong. I'm going to test myself when I can find time to play again.

Just to be clear, I would not use the word "wrong" to describe the "tickling the red" instructions for setting the input sensitivity. I would say those instructions were vague, but necessarily so because the Axe-FX lacked an input clip indicator.

Because of that vagueness, there is there has always been uncertainty about whether the input A/D converter is being clipped. Yes, that means it's possible that someone could be clipping even though they followed the instructions, but it also means that many people have not been clipping when they followed the instructions. Without a clipping indicator, it was difficult to know for sure, but the instructions were as good as they could be, given the circumstances.

In summary, we now have the tools to use a more precise procedure for adjusting the input sensitivity: Turn it up until you are just shy of getting the input clip warning.

 

[jamesmarshall]

Thanks James. I just reread the wiki link in your first post in this thread. Are you saying that the LED red is not an indicator of where the A/D chip clips? It almost seems like you're saying the only indicator of where the A/D clips is the input clip warning. That just does not sound right at all. I know the input gain is something different, where you could, eg, use that parameter in the input block to match
the output level of a standard Strat to that of a Les Paul to make a preset more usable.

Think of the A/D sensitivity more like the signal from a good ole DI box going into a console. Back in the day some would put a boost on the DI signal to get a signal they liked before hitting the console and tape which they could reamp later if needed. This was one of those things that you had to watch very closely so that it would never clip or distort, so that you could have a clean signal for reamping later. Again sometimes with a single coil strat you had to add boost before the DI box so you could have enough level to reamp it later, too little and it won't get you enough input to amplify. I have a feeling the A/D sensitivity is just like that. It is just setting the "level" going into the digital domain and not the actual signal level of the input, per se. Hence the name of it.

I could be totally wrong, but that is what I think I have figured out from all of my research on it.

 

[sprint]

Because of that vagueness, there is there has always been uncertainty about whether the input A/D converter is being clipped. Yes, that means it's possible that someone could be clipping even though they followed the instructions, but it also means that many people have not been clipping when they followed the instructions. Without a clipping indicator, it was difficult to know for sure, but the instructions were as good as they could be, given the circumstances.

In summary, we now have the tools to use a more precise procedure for adjusting the input sensitivity: Turn it up until you are just shy of getting the input clip warning.

+1 - "Tickling the red" is
somewhat vague and
open to interpretation. The new clip indicator is brutally clear / honest.

 

[State of Epicicity]

Thanks @GlennO and @jamesmarshall. I think that still does not account for the 6dB disparity. If tickling the red meant you actually had 6dB to go before actually clipping, then you really were totally fine as it was before, and the new clip warning is just activating at a much lower threshold, or with, as @Patzag mentioned, a possible difference in the way its measured. If however, the input clip warning is actually right, then the original guidance is just wrong outright.

And your testing that you posted Glenn, totally supports that second case. If you got that measurable clipping just tickling the red, you definitely did not have 6dB to go!

 

[Shahar]

Axefx mk1 or mk2 ?

I think you are on to something, seems MK II has to be set significantly lower.

 

[State of Epicicity]

I think you are on to something, seems MK II has to be set significantly lower.

This might be worth setting up in a poll. I wonder if there's a way to do that without creating a new thread. But then again, maybe in that case a thread would be better than a poll buried on page 4 or 5 of this thread.

 

[GlennO]

If tickling the red meant you actually had 6dB to go before actually clipping,

I think your confusion stems from this statement. That simply is not true. I posted an explanation above:
https://forum.fractalaudio.com/threads/input-clipping-thread.193293/page-4#post-2403054

 

[jamesmarshall]

Thanks @GlennO and @jamesmarshall. I think that still does not account for the 6dB disparity. If tickling the red meant you actually had 6dB to go before actually clipping, then you really were totally fine as it was before, and the new clip warning is just activating at a much lower threshold, or with, as @Patzag mentioned, a possible difference in the way its measured. If however, the input clip warning is actually right, then the original guidance is just wrong outright.

And your testing that you posted Glenn, totally supports that second case. If you got that measurable clipping just tickling the red, you definitely did not have 6dB to go!

I see what you are getting at however, the original wording was always vague on purpose because of so many variables.

The thing is we are talking full scale here so 0 is digital clipping but, I am thinking that is not measuring the A/D input itself, but the overall signal level and -6 for overall signal is still good and you won't clip the input section however, most A/D converters are a little more sensitive and one stray peak that an LED or your ear may not catch can still clip the A/D chip well before overall signal ever reaches 0DBFS.

 

[sprint]

"Hard Strums" is also somewhat subjective which can also contribute to some confusion.

I can't hear input sens clipping at all on distorted or EOB tones, but on pristine cleans I can hear it as a pronounced clip (not like a compression), but the clipping "pop" volume is very low compared to digital clipping at output.

 

[yek]

I think you are on to something, seems MK II has to be set significantly lower.

Absolutely not.

I have a Mark I and a Mark II here.
The Input Sensitivity and metering work the same, using the public beta.

I get an occasional on-screen clipping warning with my Gibson with PAFs and Instrument Level at 20%, on both units.
So I turn it down to 15%. No more on-screen warnings but a very occasional red LED (both units). Let's call it ticking the red.

It's a good thing that Mr.Fender's reports en Glenn's contributions have lead to adding the on-screen warning. But it doesn't mean that every Axe-Fx III user has been running the input too high and that all recordings since 2018 have a clipped guitar signal, etc.

My feeling is that OCD is getting the better of some around here.

 

[Patzag]

I just ran a very simple test:

1. Create a preset with Synth block going to Out 3. Also add an input 1 block. Does not need to go anywhere.
2. Set synth to sine wave, no tracking.
3. Out 3 on full. Route cable from Out 3 to front input.
4. Set the AD sensitivity to 100%
5. Increase the output of synth until the red "tickles". in my test it was +10dB
6. Increase output until clipping indicator comes on. In my test it was 15.5 exactly.

So there you go. Red light comes on at -6dB. Clipping indicator at -0.5.

End of story.

Set your sensitivity accordingly.

 

[State of Epicicity]

I think your confusion stems from this statement. That simply is not true. I posted an explanation above:
https://forum.fractalaudio.com/threads/input-clipping-thread.193293/page-4#post-2403054

Okay, I see. I was speed reading many posts and missed this:

the Axe-FX lacked an input clip indicator

So you're saying for sure that it lacks an input clip indicator, that instead the LED is actually

the overall signal level

But aren't you guys just guessing that too?

My feeling is that OCD is getting the better of some around here.

I mean, isn't the point of public betas to find potential bugs? I don't think that's OCD; I think it's a cool way to find little things that might've been missed. And maybe this input clip warning is correcting the vagueness of the wording in the manual, but I think it's a cool thing in this case to help clarify any confusion.