Helix Amp Models Producing DC

[MirrorProfiles]

Ahhh, gotcha. Still, that still looks like an artifact of how the FFT for the spectral display is computed to me.

Shifting the bias controls around should not behave like that thou, so there's likely some other issue with those models.

That’s absolutely what an FFT will do when there is DC present though. It can’t show 0Hz but these behaviours are typical for when DC is present. It’s a strong indicator for it, and is easy to verify. Jay’s issue is it doesn’t conclusively show that it IS DC, but something like that would be externally unlikely to be anything else in the circumstances. As confirmed by the other graphs, which honestly shouldn’t come as a shock to anyone.

The way I see it, MP made two initial assertions, one seems correct, the other was false

I made one initial assertion which was true then, and true now. I checked analog outputs, and got a result that I now can’t receate. It’s likely that I made a mistake on that one. Thankfully others have gone to the trouble of checking, which is a helpful thing to do as far as troubleshooting goes.

The analog outs has no real bearing to this as far as I’m concerned though, because the problem itself lies within the amp modelling, not the D/A or analog path. Whether or not DC is detected after the analog outputs does not change the fact that some models are producing it when they (probably) shouldn’t be. This is the part that should be fixed, and what the OP is laying out (I didn’t mention the analog outs at all there because I’d already isolated the problem to elsewhere).

 

[Chocol8]

The difference is that DC in the digital domain is a software issue and can be fixed with a firmware update. DC on the analog output is a hardware issue that cannot be fixed easily. I think that is a very significant difference. YMMV.

Either way, I am interested in what Line 6 finds out.

 

[MirrorProfiles]

The difference is that DC in the digital domain is a software issue and can be fixed with a firmware update. DC on the analog output is a hardware issue that cannot be fixed easily. I think that is a very significant difference. YMMV.

Either way, I am interested in what Line 6 finds out.

Yep. But the issue has been outlined from the start as being caused in the amp modelling, not elsewhere. It was whether the issue in the digital domain was reaching the analog outputs, or whether it was getting filtered out (rather than being an analog problem OR digital problem).

Filtering after the fact (be it analog or digitally) doesn’t help if the DC is happening before any non linear processes in the modelling. So even if the DC is filtered out, if it’s causing problems in the modelling it would need to be fixed. It’s the presence of it in critical stages of the modelling that is the bigger problem, removing DC is easy.

 

[Lysander]

That’s absolutely what an FFT will do when there is DC present though. It can’t show 0Hz but these behaviours are typical for when DC is present.

Sorry, but no, not necessarily: it depends on the implementation. DC offsets need to be properly accounted for if you want an accurate FFT transform, and i've seen plenty cases where this is not even considered, simply because inputs converted into digital signals are always DC-decoupled.

Anyway, this is all guesswork from our end. I do agree with you something is off with the way bias controls behave in those models, and it's good you brought it up.

 

[MirrorProfiles]

Sorry, but no, not necessarily: it depends on the implementation. DC offsets need to be properly accounted for if you want an accurate FFT transform, and i've seen plenty cases where this is not even considered, simply because inputs converted into digital signals are always DC-decoupled.

I agree with not necessarily, but 100% of the time I’ve seen this behaviour in a FFT, it’s made me check for DC offset, and removing DC offset always fixes it. I’m not saying it’s conclusive, but it can be a strong indicator to investigate further. I say that with no absolutes or anything - just that it’s very common to see something like that when DC is present (and especially when purely in the digital domain).

 

[Digital Igloo]

The DC on outputs discussion is a bit of a distraction now though, because the issue is to do with DC being produced in the amp modelling. If it gets filtered at the analog outputs, it doesn’t fix any potential issues it may be causing to the modelling, or any digital processes happening after the amp modelling.

Yep, DC (or DC offset) can totally hork with your signal's headroom.

I bought a Digidesign 192 ProTools interface years ago with a second 8-channel D/A card. It had terrible DC offset and it didn't take long to figure out what was the culprit. Got it fixed, but by then my whole studio was put back together and I never got around to reinstalling it nor using all 16 channels of my analog mix bus. So stupid and lazy.

 

[MirrorProfiles]

I bought a Digidesign 192 ProTools interface years ago with a second 8-channel D/A card. It had terrible DC offset and it didn't take long to figure out what was the culprit. Got it fixed, but by then my whole studio was put back together and I never got around to reinstalling it nor using all 16 channels of my analog mix bus. So stupid and lazy.

Ha those 192's got very old VERY suddenly. Kind of amazing that for years they were basically an industry standard, and now you can't give them away.

 

[Digital Igloo]

Ha those 192's got very old VERY suddenly. Kind of amazing that for years they were basically an industry standard, and now you can't give them away.

Oh, I definitely gave mine away haha.

 

[Elf]

Very interesting discussion, If I remember correctly there is a range of DC voltage present at the outputs of audio power amps is ok, is there such a thing in units like Helix, FAS units etc?

 

[Devin]

Very nice research kimg

 

[Lysander]

Very interesting discussion, If I remember correctly there is a range of DC voltage present at the outputs of audio power amps is ok, is there such a thing in units like Helix, FAS units etc?

tl;dr it's not at the outputs. All of these devices have some type of decoupling in place, so DC voltage cannot get through even if something else breaks.

What @MirrorProfiles found out is that the digital signal might have a DC offset - i.e. it's shifted from zero. This can cause troubles of its own, but always in the digital domain.

 

[Frodebro]

You know a thread is going weird when @Orvillain is the voice of reason.

 

[Lysander]

You know a thread is going weird when @Orvillain is the voice of reason.

 

[Chocol8]

You know a thread is going weird when @Orvillain is the voice of reason.

Yeah, but at least we can still call him a cunt.

 

[GTR37]

Thread is about the Helix amp models producing DC. And do we agree that’s the case?

Line 6 saw enough from the post to warrant looking at it, thread has achieved what it set out.

Exactly, I think you gave them enough the Ben is on it so it was valid enough for them to pursue

 

[Orvillain]

Yeah, but at least we can still call him a cunt.

Yeah but I'm your cunt. xxx

 

[BenIfin]

Great work, Jay! As always THANK YOU for your contributions. We are so lucky to have such a great resource here at TGF.

* part of this is sarcasm. Part of it is moderate hyperbole laced, as all hyperbole is, with truth. If you plug enough of my posts into ChatGPT it maaaaay be able to figure out which part is which.

Hey ... multi-colored, bolding and bold-underlining is my gig .... ok

And while you're at it, don't even think of going the bold underlined italics :)

 

[AlbertA]

 

[deadpool_25]

Just so I'm clear on the bottom line here:

MP reported what seems like a bug and the Line6 crew are looking into it.

Is that correct?

 

[Orvillain]