IK Multimedia TONEX

[Chocol8]

I just think NAM is still too nerdy for the masses. When they make cheaper, universal hardware, and an easy step-by-step capture process, then I think it will have more of a chance.

Yep, NAM is incredibly capable, but also has a big learning curve and a lot of quirks. Not that Tonex doesn't as well, but Tonex is more integrated and the software has been improving.

I think mass adoption of NAM is going to require some major user experience upgrades either from the community or from vendors who support it in their ecosystems. There is no reason to think that won't happen, but I don't think we are there yet.

 

[bucephalus]

Plus the Tonex pedals are super cheap and sound the same as the plug in. Nothing out there yet that does the same for Nam at full res to my knowledge

 

[Jarick]

Trying to process a Tonex capture on my new computer, it's the same length of time as before. I thought Tonex captures were supposed to be fast in the cloud but they seem to take 45-60 minutes. The new M5 is significantly faster specifically in this type of work. Maybe it is cloud based but really bottle necked?

 

[MirrorProfiles]

Trying to process a Tonex capture on my new computer, it's the same length of time as before. I thought Tonex captures were supposed to be fast in the cloud but they seem to take 45-60 minutes. The new M5 is significantly faster specifically in this type of work. Maybe it is cloud based but really bottle necked?

ToneX only has local training, there’s no cloud training option yet. And it’s always been slow on mac unfortunately.

 

[Jarick]

ToneX only has local training, there’s no cloud training option yet. And it’s always been slow on mac unfortunately.

Hmm, that's really weird that there's literally no difference between computers four generations apart.

Anyways I'm messing around with trying both Apollo Twin and MOTU M4 for captures. Despite being a fraction of the price, the MOTU was way easier to setup and route audio. I think there's some weird preamp cross talk between the Apollo Twin channels...raising the gain of one impacts the other somehow, and that may be causing the noise issues I've found. Processing both captures now to see if there's an audible difference.

 

[Sascha Franck]

Hmm, that's really weird that there's literally no difference between computers four generations apart.

It's likely the GPU which is the culprit here. Very little gains there between mobile M-generations.

 

[Chocol8]

If the code is optimized for an NVidea GPU, which I believe is the case, everything else will be slow. That said, they take about 25 minutes on my M3 MacBook Pro, so something doesn't seem right if it is taking nearly twice as long on an M5.

 

[Jarick]

It's likely the GPU which is the culprit here. Very little gains there between mobile M-generations.

I think you’re right it’s making YouTube choppy but not maxing cpu or memory. Lots of room for optimization if IK wanted.

 

[Shask]

Trying to process a Tonex capture on my new computer, it's the same length of time as before. I thought Tonex captures were supposed to be fast in the cloud but they seem to take 45-60 minutes. The new M5 is significantly faster specifically in this type of work. Maybe it is cloud based but really bottle necked?

No cloud, you need a Nvidia graphics card. They take 8 minutes for extended training on my computer.

 

[Deadpan]

No cloud, you need a Nvidia graphics card. They take 8 minutes for extended training on my computer.

Which card do you have? Mine is. 5080 and takes 8 minutes training.

 

[Shask]

Which card do you have? Mine is. 5080 and takes 8 minutes training.

5070

 

[Jarick]

Uploaded four captures from the Friedman IRX...these sound pretty good too.

 

[Sascha Franck]

Uploaded four captures from the Friedman IRX...these sound pretty good too.

Link?

 

[Sascha Franck]

I think you’re right it’s making YouTube choppy but not maxing cpu or memory. Lots of room for optimization if IK wanted.

Not sure about that. Pretty much all neural network processing, crypto mining and AI stuff relies a lot on GPU juice (for whatever reasons, being a mere user I don't understand any of that myself...), which is also why GPUs got so ridiculously expensive during the last years.

Which is why I really prefer any kinda cloud based processing. The way Tone3000 does is it is tough to beat IMO.

 

[Jarick]

Link?

If you go on ToneNET and search "Jarick" it's got all my stuff.

The three Friedman JJ Junior captures are all made with the tube amp into the Suhr reactive load (no cab). There's four Friedman IRX captures made just with that and the first stock cab. I may shoot some more without the cab but I need to get some kind of direct box to adjust the signal first.