Fractal VP4 Processors versus Axe-FX II, FM9, FM3 ... and implications for future products?

Why didn’t fractal use encoders on the four switches for the VP4? I mean it’s kind of a no brainer. Everybody’s going this route so the switches can become knobs at the same time. Just wondering.
 
Why didn’t fractal use encoders on the four switches for the VP4? I mean it’s kind of a no brainer. Everybody’s going this route so the switches can become knobs at the same time. Just wondering.
Who's "everybody"? So far the only products on the market that use this pattern are the QC and TMP.

I don't think it would make any difference in the design of the VP4. You either have 4 knobs right under the display, or 4 footswitch/knobs spread further apart.
 
Who's "everybody"? So far the only products on the market that use this pattern are the QC and TMP.

TC Electronic's G-System as well, IIRC. But... yeah.

Having owned a QC for some time, i have to say i didn't find the rotary footswitches being that big of a deal, UX-wise. The touchscreen definitely made a difference thou.
 
Who's "everybody"? So far the only products on the market that use this pattern are the QC and TMP.

I don't think it would make any difference in the design of the VP4. You either have 4 knobs right under the display, or 4 footswitch/knobs spread further apart.
well some effects may have more then 8 parameters to tweak? i think encoders on switches is ingenious
 
TC Electronic's G-System as well, IIRC. But... yeah.

Having owned a QC for some time, i have to say i didn't find the rotary footswitches being that big of a deal, UX-wise. The touchscreen definitely made a difference thou.
ok... well my point was that fractal likely isnt going to touchscreen in the future anyhow so yes...those encoders on the switches would be a great plus,,and those who say they will break easy should talk to P. thorne as he uses the qc on the road.
 
well some effects may have more then 8 parameters to tweak? i think encoders on switches is ingenious
It's got its own drawbacks. While it can make for a more compact unit, because they have to double as footswitches your controls are spaced much further apart, and further away from the display so mental mapping is more difficult to learn on the QC for example. TMP with its scribble strips is a bit better but it on the other hand does not intituitively follow the layout on screen like the QC does.

Then there's potential durability concerns.
 
Wondering how much power the VP4 Platfom has relative to it's siblings.

  • The VP-4 uses the Axe-Fx III DSP's little brother. It's a single core version with the same instruction set so porting stuff from the Axe-Fx III is almost trivial.
  • The VP4 uses a SoC. It has an ARM Cortex A15 core and a C66x core along with a bunch of peripherals. The DSP in the Axe-Fx III would be too power-hungry for this application.
  • The III uses (1) dual-core Texas Instruments DSP. The FM3 uses (1) dual-core Analog Devices DSP. The FM9 uses (2) dual-core Analog Devices DSPs. The TI DSPs are much more powerful than the Analog Devices DSPs per clock and run at around twice the clock speed as well. So one TI DSP core is about four times more powerful than one Analog Devices DSP core. If we normalize processing power to the III it would be:
    • Axe-Fx III: 100%
    • FM9: 50%
    • FM3: 25%
  • So why not use the TI DSPs in everything? Power. The TI DSPs use more power and generate more heat requiring active cooling. They are also more complicated to use requiring dedicated clock generation units, multiple power supplies with specific sequencing requirements, etc.
  • Even more here: https://wiki.fractalaudio.com/wiki/index.php?title=Axe-Fx_III,_FM9_and_FM3

Does this mean the VP4 has about 50% of the power of the Axe-FX III, about the same as an FM9, and double the FM3? And would that mean that the platform could run roughly the same patches as an FM9 ... assuming Fractal were to make such a device?

Apologies for "asking for more" and "what's next" the day after the launch of an amazing product. But it's interesting, eh? ;-)

If the VP4 is using the 66AK2G12, there's two variations, one that runs at 600Mhz and another at 1GHz.
There are differences in the memory subsystem vs the TMS320C6672

But if we just compare raw GFLOPs,

66AK2G12 600Mhz: 9.6 GFLOPS (Source: https://e2e.ti.com/support/processo.../processors-forum/650283/omap-l138-arm-vs-dsp)
66AK2G12 1GHz: 16 GFLOPS (Source: https://e2e.ti.com/support/processo.../processors-forum/650283/omap-l138-arm-vs-dsp)

Source: https://www.ti.com/lit/po/sprt577b/sprt577b.pdf
TMS320C6672 1.5GHz: 24GFLOPS/core
TMS320C6672 1.25GHz (Axe-Fx III Turbo) 20GFLOPS/core
TMS320C6672 1.0GHz (Axe-Fx III) 16GFLOPS/core

So assuming the amp modeling part is a fully dedicated core on the Axe-Fx III, then yeah roughly the VP4 would be roughly 60 or 100% compared to Axe-Fx III, or 50-80% of a Turbo.
 
Last edited:
I know s**t about processors…but is it far fetched to assume that any current/widely available processors viable for new hardware…will be overpowered for a device with a limited usecase?…so going forward..the software will set the boundaries of capability, and no longer the dsp/cost of dsp?
If you give users more power, they will always find a way to use it :)
 
I know s**t about processors…but is it far fetched to assume that any current/widely available processors viable for new hardware…will be overpowered for a device with a limited usecase?…so going forward..the software will set the boundaries of capability, and no longer the dsp/cost of dsp?
Nah, companies will simply choose a cheaper processor if it can do the thing they want to do.

Digital pedals are moving to generic ARM processors for example.

Even with current stuff, I'm sure Fractal could do for example 3-4 amp blocks on their hardware, but it just wouldn't be at the quality they want to do, or it would leave you no CPU horsepower for effects.
 
the-flintstones-listening-to-music.gif
 
U sure less powerfull still equals cheaper?
I could imagine whatever gets mass produced in silly quantities gets the edge in price…even when they are powerfull.
Whatever goes into iPhones/samsung phones for example..those production numbers must be of the scale..
Most DSP chips are sold in lots of 1000+ units if you want the best pricing. So less powerful tends to be directly just cheaper when the amounts are like that.
 
If TI is now offering DSPs that are more power efficient, cost effective and more straightforward to integrate, then future Fractal products could move to using only TI processors instead of the mix of TI and Analog Devices. Axe-Fx IV is likely to use the newer TI C77x cores, but "FM4" and "FM10" are still a mystery, if those products happen in the first place.
Yeah the new TI Soc offerings with the TI C71x core are the Jacinto 7 family, which is targeted for automotive applications.

The https://www.ti.com/product/DRA829J for example, which has 1 C71x core, 2 c66x cores and a plethora of peripherals (overkill for this application), including USB3 support, 3D graphics, display controllers, etc - basically an automotive version of OMAP5 which was canceled more than a decade ago.

That Soc is still "reasonably" priced at < $100 - I could see an Axe-Fx IV with that SoC :)

The 1 C671x at 1 GHz can do 80GFLOPS - so roughly like capable of running 2 axe-fx III's.

And that's not even including the two other c66x cores AND the 2 64-bit armv8 Cortex A72 cores that can run up to 2GHz.

It's a nice beast of an SoC.

Not sure what the power and cooling requirements are though.
 
Last edited:
If the VP4 is using the 66AK2G12, there's two variations, one that runs at 600Mhz and another at 1GHz.
There are differences in the memory subsystem vs the TMS320C6672

But if we just compare raw GFLOPs,

66AK2G12 600Mhz: 9.6 GFLOPS (Source: https://e2e.ti.com/support/processo.../processors-forum/650283/omap-l138-arm-vs-dsp)
66AK2G12 1GHz: 16 GFLOPS (Source: https://e2e.ti.com/support/processo.../processors-forum/650283/omap-l138-arm-vs-dsp)
TMS320C6672 1.5GHz (Axe-Fx III Turbo?): 24GFLOPS/core (Source: https://www.ti.com/lit/ds/symlink/tms320c6672.pdf?ts=1729275653668&ref_url=https%3A%2F%2Fwww.ti.com%2Fproduct%2FTMS320C6672)

TMS320C6672 1.25GHz (Axe-Fx III Normal?) 20GFLOPS/core

So assuming the amp modeling part is a fully dedicated core on the Axe-Fx III, then yeah roughly the VP4 would be rougly 40-60% compared to Axe-Fx III.
IIRC the axe fx III standard has two 1 GHz cores and the turbo 1.25 GHz.
 
That is not correct.

The FM9 has 1 core dedicated to Reverb, 1 core dedicated to Delay, 1 core dedicated to Amps and 1 for everything else.
Are you sure? Afaik it's 2 cores each dedicated to one amp block, 1 core for reverbs and delays and 1 for everything else.

Imho what you said just doesn't make sense, if that was the case then an fm3 would be perfectly able to run 2 amp blocks cuz it would have the same core dedicated to amps (yeah, I know that core on the fm3 runs delays too, but no way 2 delays = 1 amp in terms of cpu).

PS: iirc recently someone on the forum mentioned there's a page on the fm9 UI that shows the percentage for cpu2, should be easy to find out with that.

PPS: another combination that imho makes sense could be:
Core 1: Amp1 + Delay 1
Core 2: Amp2 + Delay 2
Core 3: Reverbs 1 and 2
Core 4: all other FX
 
Last edited:
Back
Top