Axe-Fx III Firmware 32.02 Public Beta #2 (Beta 3)

Status
Not open for further replies.
Which one is more accurate to a real world pedal, the previous version that introduced aliasing, or the new oversampled option that doesn’t?

Whichever one that is should be labeled “Normal” and the other “Economy.”

If going by the spirit of Fractal Audio wanting to have the truest recreation to real world equipment still standing.
 
It's Normal mode and Cork Sniffer mode ;-)

It's one thing to think we hear a difference/improvement when we're the one who just toggled the setting. I wonder how many people could tell in a blind test, and even if they hear a difference, could they tell which was which?
 
Which one is more accurate to a real world pedal, the previous version that introduced aliasing, or the new oversampled option that doesn’t?

Whichever one that is should be labeled “Normal” and the other “Economy.”

If going by the spirit of Fractal Audio wanting to have the truest recreation to real world equipment still standing.

Aliasing is a digital signal issue. Analog pedals do not have it at all. Less aliasing would be closer to the real pedal but at some point the extra oversampling required to remove more of it is not cost effective in terms of CPU load.
 
Aliasing is a digital signal issue. Analog pedals do not have it at all. Less aliasing would be closer to the real pedal but at some point the extra oversampling required to remove more of it is not cost effective in terms of CPU load.
Which means you have to lower the sampling to reduce CPU load, which is also called…

drumroll

Economy Mode 😅
 
Regarding the oversampling, I can hear a difference with the 5150 and 808 in the factory preset when switching between the two modes. Much prefer the higher mode. It seems to add up to 5% cpu, it fluctuates a bit though so hard to be exact.
Interesting, I try to be careful when I put an 808 into a high gain amp on my FM9 because it seems like it’s easy to get a mechanical-like texture to the distortion. Can’t tell if aliasing or what happens when you push the earlier stages of the amp super hard.
 
unix-guy he is not literally saying you save CPU over the current default. he is saying as a matter of general user language access that normal is a CPU savings over high and that "Eco" as a descriptor would convey that accessibly rather than keeping "Oversampling" as a category which is less accessible
Ok, but he literally said "saves CPU" multiple times.

This is going to be my last reply on the topic as Cliff will do what he thinks it's best and I'll live with it.

I just don't agree with labeling what has been the normal mode as "economy" just to make the label more palatable, because that's not what it is.

For example, when Full Res IRs came out, we didn't change the names of Standard Res and Ultra Res...
 
Added “Oversampling” parameter to Drive block. When set to HIGH the oversampling rate is doubled which results in less aliasing. NOTE: This will increase CPU usage considerably.
So what good is CPU if you don’t stress-test it every now and then?!? Love me some oversampling, I do.
 
Placebo effect definitely sounds better with the improved specifications.
Not placebo but definitely pretty subtle to my ear. I could not hear any difference between Normal / High on a stock SuperOD drive into the Archean Clean amp model so I set up the same Super OD left and right except oversampling = "normal" left and = "high" right + with flipped phase right and then combining L/R to see (hear) how much of a difference there actually is. Barely any sound gets thru at all confirming its very subtle for me - so on it's own, seems not worth the CPU in context of my mindless basement hacker's noodling on kitchen sink presets - but, for recordings with non kitchen sink presets, I'd turn it on just so it can be the best it can be.
 

Attachments

No, but it's become a specmanship thing so we have no choice but to address it.

It's like latency. People compare latency of products and because they don't understand the subtleties simply assume that lower = better.

I've explained why lower latency isn't necessarily better and usually means worse but it's a losing battle.

Another one is "32-bit converters". 32 is more than 24 so it must be better, right? It doesn't matter that the actual dynamic range of the 32-bit converter is less than the supposedly inferior 24-bit converter, people only know that 32 > 24.
Got any links to those discussions / explanations? I'd be interested in learning more.
 
Status
Not open for further replies.
Back
Top Bottom