Input gain and metering

[GlennO]

One impediment to using the factory presets or sharing presets is there is no way to know what kind of input signal strength the preset author was using when they created the preset. This sometimes leads to confusion and disappointment when sharing presets with people who have input signals of varying strengths.

What if, when you create a preset, it could be tagged with your guitar input signal strength. When you share it with somebody, that tag could be used to optionally adjust their input gain to match the signal coming from your guitar. That might make the experience of the first time user browsing factory presets more satisfying. It could also help improve the preset sharing experience among AxeFX users.

All modelers, including the AxeFX, have a global input level adjustment. This idea is basically a suggestion on how to make use of this existing feature in an intelligent and optional manner to help solve a common problem. No, this isn't how a tube amp works, but most interesting innovations in digital modelers come from doing things that tube amps can't do.

Edit: rewritten for succinctness

 

[FractalAudio]

I have to admit I don't understand the rationale for the input design on the AxeFX. It seems to me the goals are:

1) Have a full dynamic range to the ADC. In other words, your signal should be as loud as possible to minimize noise introduced by the ADC.
2) Avoid clipping. You don't want it so loud that it clips.

So, why not just have a single gain adjustment before the ADC, and a meter that shows if you clip? In other words, get rid of the compensation that is causing confusion and have a meter that lights up the top LED only when you clip?


This would have a significant benefit: if all AxeFX users adjusted the input levels so the meter showed a strong signal just short of clipping, this new design would mean all AxeFX users would be using a similar signal strength at the start of the digital processing chain. This would make it a lot easier to get good sounds when playing the factory presets or shared presets because everyone would be using a signal strength that matches what the preset authors used. As it is currently, the compensation gain ensures everyone is using different input signal strengths and getting inconsistent results when using the factory or shared presets.

Look at the confusion that occurred when Cooper demonstrated factory presets. That revealed that he gets very different sounds out of those presets than many other people because of differing input signal strengths. I wouldn't be surprised if getting rid of the compensation gain and adjusting input gain to a nominal level as I suggest above would lead to a revelation for many people when they try the factory presets (or shared presets) again with this new system.

People always like to complain about factory presets and there are other factors of course that will affect how a preset sounds with a particular guitar, but getting rid of the compensation gain would at least put everyone on a level playing field when it comes to input signal strength.

That said, it's possible there is something I'm overlooking that makes the current design better than what I'm suggesting.

Then the amp gain would vary.

 

[GlennO]

Then the amp gain would vary.

The idea is: everyone would adjust their input gain (pre-ADC) to achieve a nominal level at the start of the digital processing chain, thus helping to ensure consistent results from the digital processing. In other words, the preamp gain that works for one person would work for others when that preset is shared. The current compensation gain makes it hard to get consistent results. Does that make sense?

 

[FractalAudio]

The idea is: everyone would adjust their input gain (pre-ADC) to achieve a nominal level at the start of the digital processing chain, thus helping to ensure consistent results from the digital processing. In other words, the preamp gain that works for one person would work for others when that preset is shared. The current compensation gain makes it hard to get consistent results. Does that make sense?

No, it doesn't make sense. The signal out of the A/D is normalized so that the amp models (and drive models, etc.) respond correctly. Without that normalization the amp gain would be dependent upon the A/D input level.

This is the whole problem with plug-ins. There's no normalization because the gain of the audio interface is an unknown. A Deluxe Reverb model might have 10 times the gain as the real amp because the plug-in has no idea how much gain the interface has.

 

[Gilmourizing]

The idea is: everyone would adjust their input gain (pre-ADC) to achieve a nominal level at the start of the digital processing chain, thus helping to ensure consistent results from the digital processing. In other words, the preamp gain that works for one person would work for others when that preset is shared. The current compensation gain makes it hard to get consistent results. Does that make sense?

You are trying to make the same amp settings to have the "same" gain in different guitars (by changing a/d input level), but that is not what happen in the real world. It is expected to have more gain when you plug a guitar with hotter pickups.

 

[GlennO]

No, it doesn't make sense. The signal out of the A/D is normalized so that the amp models (and drive models, etc.) respond correctly. Without that normalization the amp gain would be dependent upon the A/D input level.

This is the whole problem with plug-ins. There's no normalization because the gain of the audio interface is an unknown. A Deluxe Reverb model might have 10 times the gain as the real amp because the plug-in has no idea how much gain the interface has.

Well, the normalization with a plugin happens when the guitarist watches the meters and adjusts the input level to achieve the recommended input signal strength . Every guitarist knows how to do that, so why not leverage off of that knowledge to achieve consistent input?

It just seems IMHO the current compensation gain is causing more problems than it is solving. It's great for optimizing signal/noise without changing the tone. But the cost is it makes presets sound inconsistent when shared. You can achieve the same signal/noise optimization without the compensation gain by normalizing your input once for your guitar, then dialing in presets to taste. The advantage would be that you can share those presets, or browse factory presets, without the current inconsistency.

 

[FractalAudio]

Well, the normalization with a plugin happens when the guitarist watches the meters and adjusts the input level to achieve the recommended input signal strength . Every guitarist knows how to do that, so why not leverage off of that knowledge to achieve consistent input?

It just seems IMHO the current compensation gain is causing more problems than it is solving. It's great for optimizing signal/noise without changing the tone. But the cost is it makes presets sound inconsistent when shared. You can achieve the same signal/noise optimization without the compensation gain by normalizing your input once for your guitar, then dialing in presets to taste. The advantage would be that you can share those presets, or browse factory presets, without the current inconsistency.

I guess we'll have to agree to disagree.

 

[IronSean]

Well, the normalization with a plugin happens when the guitarist watches the meters and adjusts the input level to achieve the recommended input signal strength . Every guitarist knows how to do that, so why not leverage off of that knowledge to achieve consistent input?

I would argue only guitarists who do a lot of recording with plugins know this. There's entire threads and videos on the topic of gain staging plugins because lots of people don't know this.

It just seems IMHO the current compensation gain is causing more problems than it is solving. It's great for optimizing signal/noise without changing the tone. But the cost is it makes presets sound inconsistent when shared. You can achieve the same signal/noise optimization without the compensation gain by normalizing your input once for your guitar, then dialing in presets to taste. The advantage would be that you can share those presets, or browse factory presets, without the current inconsistency.

What is does is it allows an amp to behave like a real amp. If you plug in your strat to a fender model, and set the knobs in the same ballpark as the real thing you'll get a similar level of headroom to the real thing. If you compensate that strat so it's as hot as a humbucker then your fender is overdriving like you plugged in a humbucker guitar. That would cause more confusion for sure then anything we see currently. And it would encourage people to leave the input compensation low for their strat so it acts like a strat, which will hurt the Signal-to-Noise ratio.

You want your presets to work independant of the guitar, but many people want their guitar to work like it does on real amps, and in real amps the guitar level varies wildly.

The biggest confusion I see people have with the input compensation is overthinking it because they don't think it's as smart as it is. They think turning it up actually increases volume when it doesn't.

 

[Dave Merrill]

Well, the normalization with a plugin happens when the guitarist watches the meters and adjusts the input level to achieve the recommended input signal strength . Every guitarist knows how to do that, so why not leverage off of that knowledge to achieve consistent input?

It just seems IMHO the current compensation gain is causing more problems than it is solving. It's great for optimizing signal/noise without changing the tone. But the cost is it makes presets sound inconsistent when shared. You can achieve the same signal/noise optimization without the compensation gain by normalizing your input once for your guitar, then dialing in presets to taste. The advantage would be that you can share those presets, or browse factory presets, without the current inconsistency.

The current Fractal paradigm is more like real world amps. Hotter pickups hit the amp harder, that's part of how they sound. Input level adjustments only optimise s/n in the input convertor.

It's not clear that that's the best strategy, since it does make sharing presets, including the "shared" factory ones, less consistent than if users adjusted the effective input gain to match their guitars.

EDIT: I see others posted the same thing while I was typing.

 

[FractalAudio]

The current Fractal paradigm is more like real world amps. Hotter pickups hit the amp harder, that's part of how they sound. Input level adjustments only optimise s/n in the input convertor.

It's not clear that that's the best strategy, since it does make sharing presets, including the "shared" factory ones, less consistent than if users adjusted the effective input gain to match their guitars.

EDIT: I see others posted the same thing while I was typing.

It's not "more like real world amps". It's EXACTLY like real world amps.

 

[IronSean]

The current Fractal paradigm is more like real world amps. Hotter pickups hit the amp harder, that's part of how they sound. Input level adjustments only optimise s/n in the input convertor.

It's not clear that that's the best strategy, since it does make sharing presets, including the "shared" factory ones less consistent than if users adjusted the effective input gain to match their guitars.

Yeah, currently it means you might need to use the input trim to adjust your guitar to a preset if it doesn't match up. This suggestion would mean users would need to use the trim to adjust their strat to sound like a strat on every preset they use.

Currently if you switch guitars you adjust the input sensitivity to get the right SNR. With this suggestion, if you switch guitars you need to adjust all your presets input trims to act like a lower output guitar was plugged in.

Having a universal signal level would make it easier to change guitars and keep the same tone. It would make it harder to change guitars and have the amps react like you changed guitars.

 

[GlennO]

You want your presets to work independant of the guitar, but many people want their guitar to work like it does on real amps, and in real amps the guitar level varies wildly.

I think that's the bottom line here. The issue is: do you want your AxeFX to work like the real amp, down to getting the same tone when the knobs are at the same positions? Or are you more interested in achieving the benefits of digital modeling over tube amps? There's no right or wrong here, just a preference.

Personally, the former holds very little interest for me. I've never owned 99% of the amps modeled in the AxeFX, so why would I care if the knob positions match the real amp? OTOH, the advantages of digital modeling over tube amps is the very reason I bought an AxeFX, so the latter is important to me. I wonder...is that an unusual viewpoint?

 

[IronSean]

I think there's arguments for both. But there's definitely a lot of people who would want their strat to behave like a strat when they plug it into an amp, rather than some kind of really high output uberstrat. Or that want their really high output pickups to overdrive the crap out of any amp they plug into, because that's why they bought them.

There is no right answer, but this behaviour is definitely one of the differentiating factors that makes so many people impressed with the Fractal stuff. I can't count the number of times I've heard someone saying this though: "It's amazing, using different guitars actually makes a difference, just like my real amps. The Fractal stuff is so good. On X platform if doesn't seem to matter what you plug in, it always kind of sounds the same."

Edit: Personally, I don't use single coils ever, so I'm not one of those people. But I know they exist.

 

[jefferski]

Using and sharing presets is not going to be completely consistent, in so many different aspects that have been covered here ad nauseum. Different speaker setups, different guitars, different fingers, listening at different volumes... I'm not calling out the OP, because your request does make sense. You want consistency, in a way that makes sense to you.

My experience through many iterations of new factory presets, along with tons of user presets, is some are killer right away, and others I need to tweak. Or I just leave them alone and move on... I don't need or expect every preset to be perfect for me. That isn't how it works, any more than walking into a store, plugging in to whatever amp, and then complaining that the last guy had it set with too much bass and not enough gain, and the guitar I pulled off the wall had the action too high. Wouldn't you just adjust the amp and grab a different guitar?

 

[DLC86]

The issue is: do you want your AxeFX to work like the real amp, down to getting the same tone when the knobs are at the same positions? Or are you more interested in achieving the benefits of digital modeling over tube amps? There's no right or wrong here, just a preference.

That's not necessarily a choice cuz we have both in the axe fx, the newly added global input gain serves exactly that purpose. Leave it at 0 if you want authentic amp/fx behaviour, turn it up or down if you want to adjust for differences between your guitars.

I don't see the issue here

 

[GlennO]

My experience through many iterations of new factory presets, along with tons of user presets, is some are killer right away, and others I need to tweak. Or I just leave them alone and move on... I don't need or expect every preset to be perfect for me.

And that's my point. I think a leading reason people dismiss a shared or factory preset is because of a misunderstanding about the expected input signal strength. There is simply no way to know what the preset author intended. I think the comments about a factory presets video Cooper did recently bear this out. The concept of a nominal input level would help with this. But of course there are other factors that might determine whether a shared preset works for you and your guitar. And has been mentioned above, there are pros and cons to this. That's why I didn't post this as a Wish...it's just an idea for discussion.

 

[yek]

Currently if you switch guitars you adjust the input sensitivity to get the right SNR. With this suggestion, if you switch guitars you need to adjust all your presets input trims to act like a lower output guitar was plugged in.

Not necessary at all. Set the parameter using your loudest guitar, and then stick to that. You'll be fine.

 

[Gilmourizing]

And that's my point. I think a leading reason people dismiss a shared or factory preset is because of a misunderstanding about the expected input signal strength. There is simply no way to know what the preset author intended. I think the comments about a factory presets video Cooper did recently bear this out. The concept of a nominal input level would help with this. But of course there are other factors that might determine whether a shared preset works for you and your guitar. And has been mentioned above, there are pros and cons to this. That's why I didn't post this as a Wish...it's just an idea for discussion.

So, would you want a kind of "Gain Match" feature where you (and others) could match gain of an actual guitar based on a reference?
Personally, I don't know if I would use, but sounds like a reasonable wish.

EDIT: When I said that I was thinking something similar to ToneMatch block for gain, but maybe you would need it to work as a Global Block to get what you want.

 

[rayza]

The way i understand it, the 'compensation / normalization' that's happening is there to give a more true representation of the pickups / guitar used and its interaction the amps etc. If you need to increase or decrease input level / gain to get the best signal to noise ratio at the input, that addition or reduction of gain is not carried through to the actual signal chain and affecting the way things sounds and react?

 

[Admin M@]

I think the Input controls are perfect as they are. Also, no one has mentioned the recently added INPUT 1 GAIN control in SETUP: I/O: Input, which is an important part of this conversation. By the way, I know a guy who's been on the cover of every guitar magazine ever who sometimes likes to overdrive the A/D input of the Axe-Fx because he likes how it sounds.