Valeton GP-5

[Dirk Dimehead]

It can sound as good as NAM using a fraction of the DSP. Thats way more important.

That’s actually a huge factor.

ToneX runs on something like a ~$3 600 MHz microcontroller. Native NAM A1/Standard, on the other hand, requires at least something in the range of an ARM Cortex-A53 class processor. And that pulls in a whole chain of complexity: solid embedded Linux know-how, much more complex hardware, PMIC, LPDDR4 RAM, eMMC flash, etc.

Especially RAM and flash are a real issue right now. Due to the AI boom since late last year, prices for even the smallest chips (1 GB RAM / 4 GB flash) have almost increased tenfold. That makes building an affordable device with that kind of architecture significantly harder.


Valeton, for example, relies on an ultra low-cost Bluetooth SoC with integrated DSP and codec. That’s basically the only way to hit those price points. The entire hardware BOM can’t exceed ~20 € if you want to achieve that kind of 100€ish retail price.

But of course there are trade-offs: typically noisy 44.1 kHz audio, limited DSP headroom, and not enough processing power to properly handle converted NAM + IR at the same time.

 

[Sascha Franck]

It can sound as good as NAM using a fraction of the DSP. Thats way more important.

Oh right, didn't ever think about that too much.
Just did a test, comparing a full capture in each the NAM player and Tonex, there's indeed quite a significant difference (used a 400 epochs NAM file).
So yes, I do agree then, that surely adds to the hardware expense.

 

[Sascha Franck]

@Dirk Dimehead, thanks for the explanation (and yes, AI be damned, prices for all kinda stuff are completely through the roof and it'd possibly remain like that for a whole while to come - the external 2TB Sandisk SSDs I bought last year are twice as much now).

Anyhow, while we have you here - if you'd build a unit at either the Tonex Once or Valeton GP-50 form factor (doesn't need much bells and whistles, for my personal use case a single NAM capture plus MIDI support would do), I'd be *very* interested. I guess most of the Tonex One crowd would be.

 

[Dirk Dimehead]

@Dirk Dimehead, thanks for the explanation (and yes, AI be damned, prices for all kinda stuff are completely through the roof and it'd possibly remain like that for a whole while to come - the external 2TB Sandisk SSDs I bought last year are twice as much now).

Anyhow, while we have you here - if you'd build a unit at either the Tonex Once or Valeton GP-50 form factor (doesn't need much bells and whistles, for my personal use case a single NAM capture plus MIDI support would do), I'd be *very* interested. I guess most of the Tonex One crowd would be.

We get this request almost weekly through different channels

We’ve looked into this topic quite intensively. The problem is that squeezing the required native NAM processing power into something like the ToneX One form factor is close to impossible.

Even with a slightly larger enclosure, something Hammond 1590B-sized, the device wouldn’t end up significantly cheaper than our current NAM Player. The processor, RAM, power design, and overall platform requirements simply don’t scale down proportionally with enclosure size.

And from a business perspective, maintaining two support-intensive products as a very small company is also not trivial. Firmware, testing, updates, customer support... it all doubles.

 

[ragingplatypi]

I just saw an interview with Eric Klein where he compared captures to DAW plugins or video games, where people selling captures of particular amps are going to release them in multiple formats to work with all the major operating systems.

 

[Sedaxel]

All of this makes me notice that I wasn´t taking into account the change in the costs tendency because of the AI thing.

We were used to see costs decrease for the same horsepower hardware. Fast computers were much cheaper after a couple years. But yeah... now it seems to not be the case anymore.

So, even when full NAM hardware implementation is a matter of power, it maybe won´t be that cheaply available with time. And it won´t be easily integrated in multieffects, unless they are expensive.

 

[Sedaxel]

Sooooo... for NAM taking the whole capturing world anytime soon, it should either happen some of the following:

1 - Steve optimizes the software so it gets close to ToneX efficiency.

2 - Or prices of powerful hardware dips really down unexpectedly.

Anyway, this reinforces my feeling that NAM won´t be the standard in hardware units (at least, not full NAM). Not in its current state of resorces burner.

It actually makes me wonder if Fractal will be able to implement it without heavily cutting signal chain, flexibility, etc.

 

[Shask]

I can understand the advantages of the QC, but what are the advantages of the Tonex format?

I would say easy integration into their ecosystem. Same sounds as a plugins as cheap hardware, easily used inside Amplitube for a full package solution. All they really need now is a larger pedal that completely loads Amplitube sounds. Bonus if it can also load T-Racks effects.

 

[Sascha Franck]

The processor, RAM, power design, and overall platform requirements simply don’t scale down proportionally with enclosure size.

I'm kinda aware of that. Still, you can hope, can you?
Unfortunately, I could not really make use of your current model - but a smaller version (with, uhm, an autosave feature such as the TXO...) would really make me think again.

 

[Sascha Franck]

All they really need now is a larger pedal that completely loads Amplitube sounds.

That'd peak my interest. Unfortunately, I'd have little hopes for the UI quality, given it's IK and they love things the convoluted way.

 

[Shask]

That'd peak my interest. Unfortunately, I'd have little hopes for the UI quality, given it's IK and they love things the convoluted way.

I tend to learn the UI of whatever I am using if I like sounds. It's not a big deal.