• We would like to remind our members that this is a privately owned, run and supported forum. You are here at the invitation and discretion of the owners. As such, rules and standards of conduct will be applied that help keep this forum functioning as the owners desire. These include, but are not limited to, removing content and even access to the forum.

    Please give yourself a refresher on the forum rules you agreed to follow when you signed up.

EQ after a volume pedal? Does resolution decrease???

karmar

Member
Imagine an EQ in a chain AFTER a volume pedal. When the pedal is backed off, the signal is reduced into the filter and - since the signal voltage is divided up into discrete steps at digitization, the filter does not have as much data/information to process - a resolution problem. Even if the various numeric values are simply reduced digitally, you end up with less resolution in your signal after the volume pedal because the signal is limited by its bit depth - when your volume pedal gets REALLY small you are down to 1 bit of resolution you are trying to process and extract frequency information! So I am wondering if, in the AxeFx internal processes, does the signal actually get reduced in size when you put a volume pedal in a chain - or does the Axe retain all of the information in the full sized signal, and apply an analog volume reduction at the point of digital to analog conversion. My bet is that they retain the full information for processing precision - but I am curious how they actually do it.... If they do it like I think, then it doesnt matter whether the pedal is before or after the EQ - but if the numbers actually get reduced going into the EQ processor as the chain diagram suggests, then I need to keep my volume pedal as the last thing in my chain! (otherwise the quality of the EQ - or reverb or whatever - suffers from resolution deficiency downstream in the chain when you have a low level from your volume pedal)
 

karmar

Member
OK... Thanks! So that takes care of the precision for small numbers - so the mathematical transformations in the blocks just continue to operate on the minuscule values to the same precision using floating point - that makes sense... But it raises a new question - since all the numbers are floating point, why do blocks clip when numbers get too big? It seems like the processor should just deal with the large numbers using floating point in the same way it deals with small numbers... Is the red-line clipping an artifact that is programmed in to keep users from having unrealistic levels at the end of their chains when the digital needs to be converted to analog?
 

yeky83

Power User
OK... Thanks! So that takes care of the precision for small numbers - so the mathematical transformations in the blocks just continue to operate on the minuscule values to the same precision using floating point - that makes sense... But it raises a new question - since all the numbers are floating point, why do blocks clip when numbers get too big? It seems like the processor should just deal with the large numbers using floating point in the same way it deals with small numbers... Is the red-line clipping an artifact that is programmed in to keep users from having unrealistic levels at the end of their chains when the digital needs to be converted to analog?
Blocks don't clip, virtually impossible:
https://forum.fractalaudio.com/threads/can-internal-clipping-occur.47224/#post-615288
 

karmar

Member
Thanks for the reference! When I notice clipping, I am pretty sure it is always in the delay as has been discussed in the AxeII thread you sent (I still want to confirm when I get time...) So the red-lining we see with the level meters in the layout view is just as a guide for us to help us end up with the proper level at the end of our chain? (and to avoid clipping in the delay block as was discussed)
 

FractalAudio

Administrator
Fractal Audio Systems
Moderator
OK... Thanks! So that takes care of the precision for small numbers - so the mathematical transformations in the blocks just continue to operate on the minuscule values to the same precision using floating point - that makes sense... But it raises a new question - since all the numbers are floating point, why do blocks clip when numbers get too big? It seems like the processor should just deal with the large numbers using floating point in the same way it deals with small numbers... Is the red-line clipping an artifact that is programmed in to keep users from having unrealistic levels at the end of their chains when the digital needs to be converted to analog?
Some blocks have clipping simulation, the delay blocks among others. This simulates the input of the delay line being overdriven (like a real analog delay). It also prevents runaway if the feedback is >= 100%.
 
Top Bottom