Can the Extra 10% DSP Improve Virtual Capo?

22 frets

22 frets

Experienced
I have read here that to insure the best performance and sound for the Pitch Virtual Capo one should build your preset as DSP efficient as possible by eliminating unnecessary blocks, shunts, etc. I don't want to spark a controversy here, but I was wondering if the difference in DSP power between the non-Turbo and the Turbo model with the additional 10% DSP would make any audible difference? Not that the sound of the Virtual Capo is bad on the non-Turbo model, but in the interest of seeking the best possible tone with the least possible artifacts, wavering, etc., has anyone done any informal testing? I realize the audience with the luxury of having both units to do an A-B test is somewhat limited. Just curious.
 
Both the turbo and non-turbo run on the same firmware so I don’t see how the 11% DSP power difference could make any difference in VC performance. The 10% just allows form making slightly more complex presets or running reverb / cabs at a higher quality.
 
Jeronimo said:
Both the turbo and non-turbo run on the same firmware so I don’t see how the 11% DSP power difference could make any difference in VC performance. The 10% just allows form making slightly more complex presets or running reverb / cabs at a higher quality.
Click to expand...
Doesn’t the reverb on its own dedicated DSP?
 
Imo no: More processing power allows more blocks / block features to run below the processors' redlines, but does not improve the quality of
what's running if compared between Ax and Ax-T (or inversely, less processing power does not dimimish the quality of what's running with equal settings).
 
Last edited:
22 frets said:
I have read here that to insure the best performance and sound for the Pitch Virtual Capo one should build your preset as DSP efficient as possible by eliminating unnecessary blocks, shunts, etc. I don't want to spark a controversy here, but I was wondering if the difference in DSP power between the non-Turbo and the Turbo model with the additional 10% DSP would make any audible difference? Not that the sound of the Virtual Capo is bad on the non-Turbo model, but in the interest of seeking the best possible tone with the least possible artifacts, wavering, etc., has anyone done any informal testing? I realize the audience with the luxury of having both units to do an A-B test is somewhat limited. Just curious.
Click to expand...

Yea, in addition to that it would be nice to have like a best practice guide for use.
Such as placement on the grid for how you want to use it, and other major parameter settings for the desired affect.
 
sprint said:
Imo no: More processing power allows more blocks / block features to run below the processors' redlines, but does not improve the quality of
what's running if compared between Ax and Ax-T (or inversely, less processing power does not dimimish the quality of what's running with equal settings).
Click to expand...
I'm no DSP coder but afaik more horsepower COULD be used to implement a more precise pitch shift algorithm. One could use a bigger sample window to achieve the same but that would increase latency. Shifting the higher notes is probably less of a problem than pitch shifting lower frequencies.
 
ChristThePhone said:
I'm no DSP coder but afaik more horsepower COULD be used to implement a more precise pitch shift algorithm. One could use a bigger sample window to achieve the same but that would increase latency. Shifting the higher notes is probably less of a problem than pitch shifting lower frequencies.
Click to expand...

k, but that just suggests a wish to develop a new and improved effect, which does not necessarily require a whole different box with more overall DSP - just an improved effect that may take more DSP (ie 5% vs 4%) than the previous version (like many other improvements we've seen along the way within the lifespan of any given Axfx generation).
 
Last edited:
sprint said:
k, but that just suggests a wish to develop a new and improved effect, which does not necessarily require a whole different box with more overall DSP - just an improved effect that takes more DSP (ie 5% vs 4%) than the previous version (like many other improvements we've seen along the way within the lifespan of any given Axfx generation).
Click to expand...

Again, I'm no expert and hopefully FAS developers will not stone me for all the rubbish I'm writing here. I just don't know how 10% more CPU power would translate into a noticably better result. In the end it's a decision of what the most bang for the buck would be.
 
ChristThePhone said:
Again, I'm no expert and hopefully FAS developers will not stone me for all the rubbish I'm writing here. I just don't know how 10% more CPU power would translate into a noticably better result. In the end it's a decision of what the most bang for the buck would be.
Click to expand...
its a valid wish - but ya, an overall increase in CPU of the entire unit is not necessarily required to improve a given effect that uses a small slice of CPU on its own. Increasing overall horsepower does not, in itself necessarily improve individual fx performance in any case. As we've seen many inprovements can be made without an overall increase in CPU - there could be rebalancing of resources internally, making options available to users for running an effect in various modes, or, just changing an effect to makes it more cpu expensive to use.
 
Last edited:
22 frets said:
I have read here that to insure the best performance and sound for the Pitch Virtual Capo one should build your preset as DSP efficient as possible by eliminating unnecessary blocks, shunts, etc. I don't want to spark a controversy here, but I was wondering if the difference in DSP power between the non-Turbo and the Turbo model with the additional 10% DSP would make any audible difference? Not that the sound of the Virtual Capo is bad on the non-Turbo model, but in the interest of seeking the best possible tone with the least possible artifacts, wavering, etc., has anyone done any informal testing? I realize the audience with the luxury of having both units to do an A-B test is somewhat limited. Just curious.
Click to expand...
Yes, the pitch detection function runs in its own thread. At high CPU utilization that thread can be postponed. The extra power of the turbo enables that thread to run more often improving pitch detection latency.
 
scottp said:
Maybe @unix-guy could shed some light here?
Click to expand...
I started to reply earlier and then didn't... :)

Actually, since the processor is faster than the non-Turbo it could allow for the algorithm to work better IF faster clock speed helps the specific algorithm(s) being used... However, I believe that due to sharing the same code on both FM9 versions, it would only work by nature of greater clock speed.

As far as my understanding from what Cliff has posted about pitch detection, it can only be so fast because of the physics of how pitch is detected. I'm not sure that clock speed will help or not, but I don't think it will make the pitch detection better... And again, it would need to work the same on the non-Turbo.

Edit:

Cliff posted while I was typing :)
 
FractalAudio said:
Yes, the pitch detection function runs in its own thread. At high CPU utilization that thread can be postponed. The extra power of the turbo enables that thread to run more often improving pitch detection latency.
Click to expand...
I never run my patches above 75% on Ax3 mk1 - so would turbo really improve my experience with the pitch block (running below redline)? - if yes, I'm surprised since I thought above, both would be the same with idle reaources available.
 
sprint said:
I never run my patches above 75% on Ax3 mk1 - so would turbo really improve my experience with the pitch block (running below redline)? - if yes, I'm surprised since I thought above, both would be the same with idle reaources available.
Click to expand...
At that CPU you wouldn't notice a difference. As you exceed 85-90% resources start to become critical. That 10-15% has to get divided up among all the lower priority threads: USB, display, controllers, foot controller(s), LED, etc., etc. as well as pitch detection. Pitch detection is a higher priority thread but you can't get blood from a stone.
 
FractalAudio said:
At that CPU you wouldn't notice a difference. As you exceed 85-90% resources start to become critical. That 10-15% has to get divided up among all the lower priority threads: USB, display, controllers, foot controller(s), LED, etc., etc. as well as pitch detection. Pitch detection is a higher priority thread but you can't get blood from a stone.
Click to expand...
Thanks

I can never quite get used to how awesome it is that the Fractal owner + experts frequently chime
in on forum discussions to help clarify.
 
FractalAudio said:
Yes, the pitch detection function runs in its own thread. At high CPU utilization that thread can be postponed. The extra power of the turbo enables that thread to run more often improving pitch detection latency.
Click to expand...
@FractalAudio

Thanks for clarifying that. It is helpful to my understanding and is consistent with how I thought it should work!
 
You must log in or register to reply here.
Back
Top Bottom