Tonex Pedal....It's REAL

[mavrick102000]

You'd have to configure the settings/tone you like in Amplitube and then capture it. Now it's an Amp Tone Model with the pedal baked in at those settings.

 

[Ed DeGenaro]

I’m a big fan of BHG and Amalgam profiles. The other day I got hipped to XOG Music profiles…I was gonna pass on it because I haven’t really missed my xtc or lone stars…
Well I tried anyways, totally floored I got his xtc, line star and vibrolux packs… and haven’t used anything else since.

Plus I like that he does DI and complete in the pack.
Just fantastic captures

 

[skellig]

Hi folks! Has anyone experienced this? I am loading the same tone model ( Hi quality @amalgam Fender clean/edge one) into both the plugin and the hardware pedal, same settings,.. etc..
While the plugin sounds great, with the right amount of cleanish smooth crunchiness, the pedal on the other hand sounds thinner, much more clean with no crunch at all .. Am I missing something? Thanks so much in avance

 

[Warmart]

Hi folks! Has anyone experienced this? I am loading the same tone model ( Hi quality @amalgam Fender clean/edge one) into both the plugin and the hardware pedal, same settings,.. etc..
While the plugin sounds great, with the right amount of cleanish smooth crunchiness, the pedal on the other hand sounds thinner, much more clean with no crunch at all .. Am I missing something? Thanks so much in avance

You have trim levels to set in Global Settings on the pedal. If you are there, while playing, it will tell you "Low" or "Ok" or "High". Good chance that's it.

 

[Orvillain]

Okay, so this is kind of mental. I recently upgraded from a 3060ti to a 4090. An extremely more powerful graphics card.

Now here's what I said before:

Advanced training is about 14-15minutes on my machine. I've got a 3060ti GPU which gets hammered when doing the training, so I know it is defo using the CUDA cores.

But I did a capture of my Dual Recto using the 4090, advanced training, amp only.... so the exact same type of session as before... and it took..... 27minutes!!!

That is really really disappointing. Something is wrong with ToneX and how it interfaces with the CUDA cores on the 4090. It's definitely using some of them, because the card is at 99% utilisation. But there is something wrong here.

For CUDA rendering reference, a scene like this:


That would take me over 12 hours to render on the 3060ti. The 4090 does it in about 2 hours and 25 minutes. So I have categorical proof that the 4090 should absolutely trounce and murder the 3060ti in its sleep. But with ToneX, it simply isn't giving the results you would expect.

ToneX has some issues. @IK Multimedia you may want to get your developers to look into this.

 

[BenIfin]

Okay, so this is kind of mental. I recently upgraded from a 3060ti to a 4090. An extremely more powerful graphics card.

Now here's what I said before:


But I did a capture of my Dual Recto using the 4090, advanced training, amp only.... so the exact same type of session as before... and it took..... 27minutes!!!

That is really really disappointing. Something is wrong with ToneX and how it interfaces with the CUDA cores on the 4090. It's definitely using some of them, because the card is at 99% utilisation. But there is something wrong here.

For CUDA rendering reference, a scene like this:


That would take me over 12 hours to render on the 3060ti. The 4090 does it in about 2 hours and 25 minutes. So I have categorical proof that the 4090 should absolutely trounce and murder the 3060ti in its sleep. But with ToneX, it simply isn't giving the results you would expect.

ToneX has some issues. @IK Multimedia you may want to get your developers to look into this.

F.w.i.w .... Karlis .... who does the Amalgamaudio Captures upgraded to [I'm pretty certain] a 4090 also, and his Advanced Captures still take 14 - 15 minutes .... weird ?? ....I remember him also writing he was surprised at the distinct lack of time-reduction.

Maybe there's an A.I processing point of no return with the IK process ? or maybe the software needs decent optimizations ?

Ben

 

[Orvillain]

F.w.i.w .... Karlis .... who does the Amalgamaudio Captures upgraded to [I'm pretty certain] a 4090 also, and his Advanced Captures still take 14 - 15 minutes .... weird ?? ....I remember him also writing he was surprised at the distinct lack of time-reduction.

Maybe there's an A.I processing point of no return with the IK process ? or maybe the software needs decent optimizations ?

Ben

I'm thinking the underlying Tensorflow or Pytorch libraries that NAM and ToneX are using, has some fundamental bug in the way it allocates jobs to the CUDA cores. I'm not seeing a time boost with NAM either.

 

[EOengineer]

Okay, so this is kind of mental. I recently upgraded from a 3060ti to a 4090. An extremely more powerful graphics card.

Now here's what I said before:


But I did a capture of my Dual Recto using the 4090, advanced training, amp only.... so the exact same type of session as before... and it took..... 27minutes!!!

That is really really disappointing. Something is wrong with ToneX and how it interfaces with the CUDA cores on the 4090. It's definitely using some of them, because the card is at 99% utilisation. But there is something wrong here.

For CUDA rendering reference, a scene like this:


That would take me over 12 hours to render on the 3060ti. The 4090 does it in about 2 hours and 25 minutes. So I have categorical proof that the 4090 should absolutely trounce and murder the 3060ti in its sleep. But with ToneX, it simply isn't giving the results you would expect.

ToneX has some issues. @IK Multimedia you may want to get your developers to look into this.

Interesting.

I’m still getting my feet under me with my handful of captures so far, but everything has used the advanced mode and takes a little over an hour per capture on my m1 Mac mini.

Just curious - which version recto are you working with? Direct or full rig captures? I’ve had lost puppy syndrome over my old rev G - would love to check out what you’ve got.

I posted up a couple Mark III direct captures recently but they aren’t dialed in for the brutal stuff yet - more middle of the road high gain.

 

[MirrorProfiles]

Interesting.

I’m still getting my feet under me with my handful of captures so far, but everything has used the advanced mode and takes a little over an hour per capture on my m1 Mac mini.

Just curious - which version recto are you working with? Direct or full rig captures? I’ve had lost puppy syndrome over my old rev G - would love to check out what you’ve got.

I posted up a couple Mark III direct captures recently but they aren’t dialed in for the brutal stuff yet - more middle of the road high gain.

I’ve got a Dual Rectifier Rev G, any sounds in particular you’re missing?

 

[yeky83]

Okay, so this is kind of mental. I recently upgraded from a 3060ti to a 4090. An extremely more powerful graphics card.

Now here's what I said before:


But I did a capture of my Dual Recto using the 4090, advanced training, amp only.... so the exact same type of session as before... and it took..... 27minutes!!!

That is really really disappointing. Something is wrong with ToneX and how it interfaces with the CUDA cores on the 4090. It's definitely using some of them, because the card is at 99% utilisation. But there is something wrong here.

For CUDA rendering reference, a scene like this:


That would take me over 12 hours to render on the 3060ti. The 4090 does it in about 2 hours and 25 minutes. So I have categorical proof that the 4090 should absolutely trounce and murder the 3060ti in its sleep. But with ToneX, it simply isn't giving the results you would expect.

ToneX has some issues. @IK Multimedia you may want to get your developers to look into this.

Whoa why do you know how to do that stuff?

 

[Orvillain]

Whoa why do you know how to do that stuff?

What do you mean?

 

[eggpl@nt]

What do you mean?

What circumstances led you to learning animation (e.g. Blender)?

 

[Orvillain]

What circumstances led you to learning animation (e.g. Blender)?

That's a bit of a can of worms. But suffice to say, for the last 3 years I've been heavily getting into 3D, Blender, and particularly building my own custom addons using Python and the Blender API.

I've got a lot of interests across the board really. Music and 3D are just two of them.

Much like music, these days if anyone wants to do 3D, the tools are right there for the taking.

 

[EOengineer]

I’ve got a Dual Rectifier Rev G, any sounds in particular you’re missing?

Oh man there were a couple tones I really loved in the rev G.

- orange channel edge of breakup with tones around noon
- orange channel with gain ~ 1-2 o’clock tones around noon
- red modern with gain and tone like I just described above for orange

I think I probably used the tube rectification in all cases.

 

[FAKA!]

Okay, so this is kind of mental. I recently upgraded from a 3060ti to a 4090. An extremely more powerful graphics card.

Now here's what I said before:


But I did a capture of my Dual Recto using the 4090, advanced training, amp only.... so the exact same type of session as before... and it took..... 27minutes!!!

That is really really disappointing. Something is wrong with ToneX and how it interfaces with the CUDA cores on the 4090. It's definitely using some of them, because the card is at 99% utilisation. But there is something wrong here.

For CUDA rendering reference, a scene like this:


That would take me over 12 hours to render on the 3060ti. The 4090 does it in about 2 hours and 25 minutes. So I have categorical proof that the 4090 should absolutely trounce and murder the 3060ti in its sleep. But with ToneX, it simply isn't giving the results you would expect.

ToneX has some issues. @IK Multimedia you may want to get your developers to look into this.

Dude that is disappointing af. I just literally pulled my new rig out of a box today, i9 13900KF, 64gigs 6000mhz ram and a RTX 4090 founders edition and was gonna get into ripping my amp for Nam and Tonex. What a bummer.

 

[bgoodman]

I'm having an issue with the ToneX pedal involving midi communication. When I send a midi signal to the ToneX pedal, the display turns off and on, along with the guitar signal turning off and on, then makes a hissing noise just like when it turns on with the orange light inside the pedal remaining on all along.

This occurs when sending a single midi preset (from morningstar MC6 mkii) with 8 total messages (mostly control commands) to other pedals. When I disable 2 of those other messages (control commands to other pedals) within the MC6 preset, the ToneX responds seamlessly.

It seems the ToneX is having problems processing midi signals that contain multple messages to other pedal(s)? No other midi messages are sending signals to ToneX (Midi Channel 7). Moving the ToneX to other channels did not help. Any1 else having these issues or could test to see if you can reproduce the error?

@IK Multimedia

Physical Midi Cable Order:
MC6 (sending midi)
ML5
POG2
Stymon Timeline
Strymon Bigsky
Disaster Area Midi Box 4
- ToneX (TRS midi from midi box to 5pin midi cable)

Thanks!
Bob

 

[Orvillain]

Just curious - which version recto are you working with? Direct or full rig captures? I’ve had lost puppy syndrome over my old rev G - would love to check out what you’ve got.

Sorry, missed this the first time around. I've got a 2018 multi-watt Dual Recto. And honestly?? It is one of the best amps I've ever played. It is right up there with my Satch JVM and my Diezel VH4 as a proper "lifer" amp.

 

[Orvillain]

Dude that is disappointing af. I just literally pulled my new rig out of a box today, i9 13900KF, 64gigs 6000mhz ram and a RTX 4090 founders edition and was gonna get into ripping my amp for Nam and Tonex. What a bummer.

Let me know your results. Always good to confirm with other people. Also try turning off hardware scheduling for your graphics, @Deadpan advised me to do that, and it did seem to speed up capturing with NAM. But I've not tried ToneX since. Even so... I only saved around 7-8minutes with NAM, which is decent... but given how powerful this card is, is still a bit disappointing overall.

At least the benefit with NAM is, you can set it up to train a bunch of stuff in one run - so do all your capturing in one fell swoop, and then train everything overnight.

 

[laxu]

Let me know your results. Always good to confirm with other people. Also try turning off hardware scheduling for your graphics, @Deadpan advised me to do that, and it did seem to speed up capturing with NAM. But I've not tried ToneX since. Even so... I only saved around 7-8minutes with NAM, which is decent... but given how powerful this card is, is still a bit disappointing overall.

At least the benefit with NAM is, you can set it up to train a bunch of stuff in one run - so do all your capturing in one fell swoop, and then train everything overnight.

With Tonex relying on PyTorch, I expect that its developers would have to update it to support the latest version to get better gains on Nvidia 40 series GPUs. It seems PyTorch 2.0+ supports newer CUDA libraries that have improvements for the 40 series architecture.

Might be a similar situation for NAM.

 

[Orvillain]

With Tonex relying on PyTorch, I expect that its developers would have to update it to support the latest version to get better gains on Nvidia 40 series GPUs. It seems PyTorch 2.0+ supports newer CUDA libraries that have improvements for the 40 series architecture.

Might be a similar situation for NAM.

I suspect you are correct./