Helix Amp Models Producing DC

[EOengineer]

Just stating the obvious - its amazing that we have L6 and Fractal folks here who are open enough to respond to these types of inquiries and let us know. They could tell us to go kick rocks but instead they...thank us. What kinda clown world is this?

 

[molul]

 

[TheTrueZoltan!]

@MirrorProfiles about to fix fractal and line 6 modeling in 2 months time

And don‘t forget his heroic efforts for plugin users.

Legend!

 

[Frodebro]

Just stating the obvious - its amazing that we have L6 and Fractal folks here who are open enough to respond to these types of inquiries and let us know. They could tell us to go kick rocks but instead they...thank us. What kinda clown world is this?

That’s the difference between somebody who truly cares about providing the best product they can, and somebody who serves only as a buffer between the customer and the bean counters.

 

[KingsXJJ]

Nice find! I hope y’all get this solved.

 

[BenIfin]

This issue isn't present in *all* the current HX Amps ? Is it ?

 

[GTR37]

This issue isn't present in *all* the current HX Amps ? Is it ?

No , only the ones he mentioned

 

[BenIfin]

No , only the ones he mentioned

Good to know. That means its not an unfixable hardware problem so should be rectified for Augora - and possibly even HX FW 3.9 whenever it arrives :)

Awesome find and demo by MP and already received and escalated by D.I.

Doesn't get much better than that :)

 

[DieSchmalle]

Igor’s said he’s got someone looking into it.


Haven’t tried via USB or SPDIF but I did test on HX stomps analog outputs and it’s present there, so I’m pretty certain it’s across the entire line.

I've attached a multimeter to the HX Stomp's output in DC voltage mode and couldn't confirm your findings with said amp models. At most that would be less than 1 mV DC which is my measurement threshold.

 

[MirrorProfiles]

I've attached a multimeter to the HX Stomp's output in DC voltage mode and couldn't confirm your findings with said amp models. At most that would be less than 1 mV DC which is my measurement threshold.

Interesting - I can't seem to re-create it either on the HW analog outs now, aside from when adjusting parameters so its likely I goofed up when testing that. It does seem to be present when using HX Stomp as a USB interface which must mean it's still being created by the model and the analog outputs are filtering the DC out.

 

[GTR37]

Interesting - I can't seem to re-create it either on the HW analog outs now, aside from when adjusting parameters so its likely I goofed up when testing that. It does seem to be present when using HX Stomp as a USB interface which must mean it's still being created by the model and the analog outputs are filtering the DC out.

View attachment 46981

That’s good news then if it just created in the code and much much easier to fix

 

[James Freeman]

I confirm, there is plenty of DC at the analog outputs of the Stomp going into the analog inputs of my audio interface.
The DC decays pretty fast because of analog ins/outs but there is indeed a big DC spike.
*Of course I still have a Stomp, what do you think I am, a fool? NEVER sell a Stomp.

A few dirty power treble peaks from my room air conditioner inverter, ignore that.

 

[jay mitchell]

I confirm, there is plenty of DC at the analog outputs of the Stomp going into the analog inputs of my audio interface.
The DC decays pretty fast because of analog ins/outs but there is indeed a big DC spike.
*Of course I still have a Stomp, what do you think I am, a fool? NEVER sell a Stomp.

This goes to show that having the means to do spectral analysis is insufficient. One must also understand what one is looking at.

The spectral display you show does not show DC. I had not bothered to watch the video in the OP, which was a mistake on my part. It also does not show DC. There are a multitude of mistakes and misunderstandings that can lead to this, but you and MP need to stop before you announce "discoveries" like this. You've just made it clear that you lack understanding of the tools you're using. If it's any comfort, you've got lots of company.

If you want to check for the presence of DC in an analog output, use a DMM set to "DC Voltage." If you want to detect it in a digital stream, record the device in question with no signal and look closely at the signal trace. Do not - I repeat, do not - attempt to use spectral analysis to detect DC. It's possible to do so, but you really have to know what you're doing.

This whole thread is useless and should be deleted.

 

[laxu]

@MirrorProfiles have you tested with Helix Native, just to take the Helix hw out of the equation?

 

[DieSchmalle]

No. The fact that DC can ever make it to an analog output is, by definition, a hardware problem. It is never permissible for a signal-processing device that produces a line-level or lower output to have the ability to pass DC.

But we haven't confirmed that Helix is doing this with prober measuring tools. An audio interface isn't the right equipment - it blocks DC.
The before mentioned "DC spikes" when changing the Bias paramenter don't count - it's a contradictory term in itself: a signal with a spike isn't constant.

So we need a low frequency threshold where we start to complain. 20 Hz? 10 Hz?

 

[jay mitchell]

But we haven't confirmed that Helix is doing this with prober measuring tools. An audio interface isn't the right equipment - it blocks DC.
The before mentioned "DC spikes" when changing the Bias paramenter don't count - it's a contradictory term in itself: a signal with a spike isn't constant.

So we need a low frequency threshold where we start to complain. 20 Hz? 10 Hz?

See my post above.

 

[KingsXJJ]

The thot plickens.

 

[Boudoir Guitar]

This goes to show that having the means to do spectral analysis is insufficient. One must also understand what one is looking at.

The spectral display you show does not show DC. I had not bothered to watch the video in the OP, which was a mistake on my part. It also does not show DC. There are a multitude of mistakes and misunderstandings that can lead to this, but you and MP need to stop before you announce "discoveries" like this. You've just made it clear that you lack understanding of the tools you're using. If it's any comfort, you've got lots of company.

If you want to check for the presence of DC in an analog output, use a DMM set to "DC Voltage." If you want to detect it in a digital stream, record the device in question with no signal and look closely at the signal trace. Do not - I repeat, do not - attempt to use spectral analysis to detect DC. It's possible to do so, but you really have to know what you're doing.

This whole thread is useless and should be deleted.

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.

 

[MirrorProfiles]

If you want to detect it in a digital stream, record the device in question with no signal and look closely at the signal trace. Do not - I repeat, do not - attempt to use spectral analysis to detect DC. It's possible to do so, but you really have to know what you're doing.

You can look at the waveform and see significant DC offset. And DC detection algorithms show DC offset.

Here is a print of the signal with no audio at input, and me zooming in on the waveform. Is there offset or is it symmetrical?

Here is a DC Offset meter from Izotope on the above for the avoidance of doubt.

Jay's welcome to flex his ego as much as he likes so long as its confirmed as either a bug or intentional. I think there is reasonable evidence to suggest there is DC unintentionally being created from the amp modelling.

We can have the electronics class later if you like.

 

[JiveTurkey]

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, at least you know how to colored font where those of us in dark mode can still read the post