I wonder if there will be a more powerful Axe Fx...

chucma

Fractal Fanatic
I was going to post this in the Wish list forum, but it isn't really a 'wish', more a ponderment (new word time folks). :encouragement:

The new Reverb has impressed me so much that I think it's about time to get the XL. There is just one thing though, with all the expected experimenting that is going on there are a lot of users who are getting high CPU usage. I totally understand this is normal, getting such silky, smooth reverb takes a super computer to perfect; however it raises 2 questions for me:

1) Is it worth waiting to see if Cliff will release a higher powered Axe unit with a CPU that has more headroom for the advanced Reverb algorithms? i.e. don't buy the XL yet, wait to see if a more powerful Axe Fx comes out of Fractal's magic bag in the near future? Tbh, this question is more hypothetical as only Cliff and a select few others know the answer to that.

2) Is the CPU high usage really an issue? Or is it the result of something new with FW16 that everyone is still learning how to use properly (and efficiently)
 
To me I don't think the cpu is such a big issue, I just think you have to be a little more inventive with your presets if you want to go hi res everything. I used to use one preset that had every thing in it, but every thing was lo res, it still sounds great. My new philosophy is be more specific, with what I need for that particular song, it's Just a different way of thinking. We have a ton of preset memory locations now in the XL, so I'm good.

I hope he keeps make things even more hi res, I'll keep less in a preset and will use every effect effectively.

its amazing to me, we just got 16.04 and he's already talking about 16.05, how many people do you know that thinks like this.


ps. Of course I wish he made a four processor box, who wouldn't.
 
Aah, The Lounge, that was probably the right forum! :D

Totally agree, as more things get more hi res, we are going to have to get smarter with our presets to get the best out of them or cut down on the effects a little. Luckily I'm not much of an effects guy, I like a good amp tone with a little delay and reverb which pretty much sums up my requirement; so I might not have the CPU issue anyway.

I do love a product that is constantly moving forward, I cannot think of any other product I have that does it the same way as Fractal does.
 
The way fractal keeps updating they will have to make a bigger unit pretty soon. All this magic cant fit in that little box. May have to build a ax caldron, rack mountable of course
 
I would bet money on it! The wish list is long and full of great ideas that could be implemented. Algos are not the only thing that are driving the next generation of Axe FX, physical interface is a large part of it as well.
 
I believe processing power will be limited to currently available DSP's. AFAIK the TigerSHARC DSP's used in the AxeFX II & XL are still the fastest available. Currently the two DSP's are split - one is dedicated to the two Amp blocks, the other handles the Cab blocks and all the effects blocks. IIRC pairing two DSP's to handle the same effects isn't a viable option so I would assume that the only way to get more CPU availability per preset would be to further subdivide the effects blocks out to different DSP's. This solution would likely add a small amount of latency though. Whether or not this is even possible and if the resulting latency would make for a noticeable difference would be a question for Cliff.

You'll notice the qualifiers (IIRC, AFAIK, etc.). I haven't been following progress in DSP development so there may well be a faster and/or more powerful DSP available that would work for a next gen AxeFX. Again, a question for Cliff.

If I had to guess I'd say we're still a few years out from a next gen unit. The current platform still seems to have the necessary resources for continued development.
 
Last edited:
I say go for it and get the XL. I just sold my Axe FX Mark II on Ebay earlier this week. It sold within hours and I got around $200 less than I paid for it over a year ago. I had 33 watchers within an hour. Needless to say these things are very popular and hold their value - especially if Fractal is out of stock. I am ordering an XL soon.
 
I wasn't around for the standard/ultra days, but how limiting did the processing power get before the II came out?

I'm pretty simple with my patches... Drive-amp-cab(s), and I rarely go over like 55%. But I have downloaded patches with a lot of effects that push 90%. Even some of the factory presets are like that.

How much do the processors cost? I imagine the III might be a 5k unit with 8 processors!
 
How much do the processors cost? I imagine the III might be a 5k unit with 8 processors!

I think Cliff mentioned somewhere in a post that the TigerSharcs are in the region of over 300 quid each, whereas competitors units use CPUs that are less than 100. I can't find the original thread that he posted that on, but funnily I came across this one that covers the same subject which I had missed earlier: :pirate:
http://forum.fractalaudio.com/lounge/86972-question-about-axe-fx-iii-beyond.html

The general consensus seems to be that the average preset uses around 50-60% and the ones with more effects than normal hit around the 80-90% mark; so this tells me there is lots of headroom available still in the XL. If you take an average of 4-10% CPU usage per effect then there is lots of room to play.
 
I think Cliff mentioned somewhere in a post that the TigerSharcs are in the region of over 300 quid each, whereas competitors units use CPUs that are less than 100. I can't find the original thread that he posted that on, but funnily I came across this one that covers the same subject which I had missed earlier: :pirate:
http://forum.fractalaudio.com/lounge/86972-question-about-axe-fx-iii-beyond.html

The general consensus seems to be that the average preset uses around 50-60% and the ones with more effects than normal hit around the 80-90% mark; so this tells me there is lots of headroom available still in the XL. If you take an average of 4-10% CPU usage per effect then there is lots of room to play.

The XL does not have more CPU power than the regular II.
 
Under the current paradigm, there's plenty of processing power still. But if Cliff starts to explore emulation of studio gear (mic pres, channel strips, etc.) then there would definitely be a limitation in processing power. UAD has shown the ability to emulate this gear with one processor, but the number of separate effects is limited. They have units with 8 processors that can run a ridiculous number of separate instances, and you have to pay serious money for those units. I can definitely see room for one or two more processors in the next AxeFx with a dedicated processor for amps, cabs, effects, and studio gear (as distinct from effects although there is a lot of overlap there).

UAD sells the Apollo Quad with four processors for the same price as the AxeFx II XL, although I'm guessing they sell more units? Not really sure. I definitely think they have more market capital and the ability to defray costs by the number of units they're selling. I have the Apollo Quad and it's a fantastic interface and a very powerful plug in processor. In fact, I'd love to see FAS adopt UADs approach to graphical representation of the analog units they model.
 
UAD sells the Apollo Quad with four processors for the same price as the AxeFx II XL, although I'm guessing they sell more units? Not really sure. I definitely think they have more market capital and the ability to defray costs by the number of units they're selling.
A little perspective:

The Apollo Quad uses SHARC processors. The fastest SHARC sells for $31 if you buy in bulk.

The Axe-Fx uses the fastest TigerSHARC processors available, which cost $252 each in bulk.
 
UAD sells the Apollo Quad with four processors for the same price as the AxeFx II XL, although I'm guessing they sell more units? Not really sure. I definitely think they have more market capital and the ability to defray costs by the number of units they're selling.

And some more perspective; when you buy the unit you get a small compliment of effects. Then you buy the rest at $100, $200, $400 a pop. I am only guessing, but I think the plug ins might be the lion's share of their revenue stream. But yes, they make nice gear.
 
UAD sells the Apollo Quad with four processors for the same price as the AxeFx II XL, although I'm guessing they sell more units? Not really sure. I definitely think they have more market capital and the ability to defray costs by the number of units they're selling. I have the Apollo Quad and it's a fantastic interface and a very powerful plug in processor. In fact, I'd love to see FAS adopt UADs approach to graphical representation of the analog units they model.

Reading up on SHARC vs TigerSHARC, the TigerSHARC seems to be way ahead in terms of multiprocessing when compared to the SHARC. So I guess if you had a single dedicated purpose, then the SHARC would be ideal, however with the Axe-Fx's unlimited flexibility and high quality I can see the reason for going to TigerSHARC. I suppose if Cliff went the SHARC route he would have had to put a lot more than 4 processors in to achieve the same results - and that would probably ramp the cost up even higher.

Still, it's a pity the consumer CPUs don't have the same drive for performance upgrades as Intel and co.
 
A little perspective:

The Apollo Quad uses SHARC processors. The fastest SHARC sells for $31 if you buy in bulk.

The Axe-Fx uses the fastest TigerSHARC processors available, which cost $252 each in bulk.

That's a really good point. Do you know the spec comparison between what UAD is using compared to the Tiger Sharc? I've tried to find details with UAD and all I find is just "SHARC" processor with no description of which model, etc. I know the quad can run a huge number of simultaneous plugins, but it's probably apples to oranges with how the AxeFx is working.
 
And some more perspective; when you buy the unit you get a small compliment of effects. Then you buy the rest at $100, $200, $400 a pop. I am only guessing, but I think the plug ins might be the lion's share of their revenue stream. But yes, they make nice gear.

Another great point. They probably do make a lot more money selling all the additional plugins over time, not to mention how much analog gear they sell.
 
Back
Top Bottom