[Implemented] Get rid of switching Audio GAPs p6/6 The "GAP FILLER..." (effective for all BLOCKS) - TIP & TRICKS

GlennO said:
Paradoxically, reducing the gap to zero could even make things worse because the abrupt timbre change would in many cases be worse than the fade-in/fade-out that the current gap permits.
Click to expand...

In my opinion I'd prefer a gapless sudden sound switch than an audible lack of sound while I'm playing live.

What I proposed is to fill the gap and all the same cross-fade the audio.
We could even make that cross-fading user adjustable fitting users' preferences and needs.

On an engineering point of view (I'm an engineer) the proposed trick (and possible development by FAS) isn't a perfect solution, but it's effective, I heard it and you can hear the difference.
FAS can do very much better than my "ducked delay Proof Of Concept" for several reasons. ;)

IMHO 50% CPU used just for gapless sound and cross-fading/morphing switching isn't an option.
(I'm aleady around 60-70% CPU with my AF3 presets...)
 
Last edited:
GlennO said:
First of all, there is no catching up to be done. Competitor amp modelers offer gapless switching only at the cost sacrificing half of their cpu capacity, not because they have some magic that FAS lacks.

Second, the algorithm proposed here would not be very effective. The more common algorithm to employ in situations like this is ping pong time stretching. That would extrapolate the audio sufficiently that it would not be noticeable over a time range of a few dozen milliseconds. It's not rocket science, but it's not clear if that would be a worthwhile thing to do since the gap is not the issue. When you get the gap down to the small number of milliseconds where it is today on the Axe-FX, it's the timbre change that makes the change sound abrupt, not the gap.

The solution is to be found in cross-fading, not gap elimination.
Click to expand...
This.
 
Would ping pong extrapolation combined with cross-fading be a handy feature? The answer might be "yes". It depends on the quality of the extrapolation and the details of exactly what is changing in the preset during the switch. It would certainly be an interesting thing to explore. Maybe Cliff has already looked into it.

However: Would extrapolation/cross fading give amp channel switching results as smooth as you can currently achieve with 2 cross-faded amp blocks? The answer to that is definitely "no". You'd almost certainly be more satisfied using two amp blocks with the current firmware. Adding a second amp block to your preset on the Axe-FX III has no effect on the cpu usage of the preset.
 
unix-guy said:
Same for the FM9...
Click to expand...
That's simply untrue. When I add a 2nd amp to my fm9t, it bumps up 1%. The difficult block is the 8% increase for the second cab block if you are trying do a full mixer/mplex setup. That's 8% i don't have to spare. For true gapless you can't be switching cab channels.
 
dean701 said:
That's simply untrue. When I add a 2nd amp to my fm9t, it bumps up 1%.
Click to expand...
That could be down to the shunt. The point is, the AMP blocks aren't significant in CPU use calculations so go ahead and use 2 if 2 gets you where you need to be.
 
iaresee said:
That could be down to the shunt. The point is, the AMP blocks aren't significant in CPU use calculations so go ahead and use 2 if 2 gets you where you need to be.
Click to expand...
That's the point, it doesn't. I use dual amps. The second cab block is the issue. Even then it's still not gapless. The ideas in the thread presented by bob4u are the key topic.
 
iaresee said:
I'm not sure what you're trying to say. The point by @chris and @unix-guy was a second AMP block uses no CPU on either of those platforms because there are dedicated resources in the DSP for the AMP block. No one is making that claim for the CAB block.
Click to expand...
..and I clearly stated that it does add 1%. It may be insignificant but for those of us running a preset at 83%,that 1% does matter.
 
Last edited by a moderator:
dean701 said:
..and the legendary idiocy among the moderators rears its head
Click to expand...
Calling out a moderator on something completely unrelated to moderation of the forum?

That's not only uncalled for, it is entirely antagonistic.

It's also against forum rules as far as I remember.

The moderators are also users, many with extensive knowledge and experience with the products.

Their opinions are also allowed.
 
Interesting discussion and I admire your persistence! Seriously!

I wonder if you had a time-based effect (delay, reverb) later in the chain if you would end up getting a similar result. Sure, there is an upstream gap but the delay/reverb will "spill over" during it and probably (likely) mask it.

So, what do the waveforms look like at the listener end, say, after a room reverb?
 
I have rescinded my previous comment. The eye roll was not the response I was expecting from a moderator for simply pointing out that there is a small cpu hit for adding a second amp. Regardless, my comment was out of line.
 
dean701 said:
I have rescinded my previous comment. The eye roll was not the response I was expecting from a moderator for simply pointing out that there is a small cpu hit for adding a second amp. Regardless, my comment was out of line.
Click to expand...
I'm sorry I antagonized you with the eye roll. I could have just explained you can make up the 1% elsewhere.
 
I'm happy that my works is stimulating a discussion even from experienced people (more than me) ;).

Anyway the gap appears not only with AMPs channel switching, also with CABs, CHORUSes, COMPRESSORs, DRIVEs and others,
whilst other blocks don't activate any gap at all (I mean absolutely ZERO gap switching their channels).

I have listed both in first part of my tests: Part 1, hope it's useful.

I have also found that - when the gap is already "excited" - that there are a series of blocks that increase the gap (listed in Part 3 ).

All above even if all those blocks are kept inactive. This already suggests first precautions to reduce/eliminate the gap.

The gap then could be 12ms to 120ms (in the craziest case). AMPs are anyway the major contributors with IMHO a too large gap (min 40ms in AF3 Turbo).

It's so typical and basic to switch amps channel so that forcing to use 2 AMPS just to avoid the gap appear to me really the last chance (that FM3 users haven't).

In my "GAP FILLER" Proof Of Concept I hope I enough demonstrated with an audible improvement, that there is a relatively easy way to improve a lot the quality of the sounds switching, if done by FAS, that could possibly be applied even for Presets Switching.
Not perfect, but very much better than a silence and open to an user adjustable cross-fading.
 
You must log in or register to reply here.
Back
Top Bottom