!Apparently that's already happened at lease once, thanks to Fractal. See this thread for comments on development of the Carol Ann Triptik 2:
Hi there! Aaron Lanterman, said Georgia Tech professor and poster of the video in question chiming in here.
I'm pleased that my musings have generated so much discussion!
To clarify, of course the results of such an effort wouldn't equate to what's in one of Fractal Audio's boxes or a Quad Cortex. And of course most of the cost of units like these is in recouping the massive cost of the many hours that were spent developing the software inside of them.
I received my FM3 last week and love it. (And I plan to pick up a Quad Cortex if it ever claws its way out of backorder hell). So, no, the intent is not "let's do this and some magic will happen and we will save money on buying an Axe FX-III or Quad Cortex."
Why would someone want to engage in a DIY DSP amp modeling effort? Because it can be (a) fun, (b) educational, and (c) allow someone to implement weird new things Fractal Audio and Neural DSP might not think of. The idea for the video started with me simply being curious about how much horsepower was in those boxes, to get a target to shoot for in terms of available horsepower.
Why do people build studio gear from kits, like the Hairball Audio 1176, when they can buy a ready made 1176 from Universal Audio or a clone from someone like Warm Audio? Why do people build amps from kits provided by people like CeriaTone -- or build from scratch -- when they could buy from Two-Rock or Red Plate? When you factor in the cost of your time, you're not really saving any money, but it can be fun, educational, and be a gateway into designing your own hardware.
I'm intrigued by how deep the programming on the FM3 goes -- I see things like "Cathode Follower Grid Clipping" and "Power Tube Bias Excursion" and I cannot wait to dig into it more. I find myself wondering what's happening under the hood, and how these parameters map to the various models of the various amps.
And then I imagine other possibilities, like, what about a piece of software that lets the user actually construct the amp at the component level by dragging circuit elements onto a schematic? Would that sell? No. Would lots of people use it? Probably not. Would it make any sense for a company like Fractal Audio to invest time in such a thing? Obviously not. But I think it would be fun and a great tool for learning about electronics.
So much focus has been on modeling existing amps, but devices like the Axe FX-III go beyond that, allowing you to create amps that have never existed before. That's fascinating to me. It's like you could prototype a new amp and tweak it in software before building the actual new amp.
And think about the Kemper and the Quad Cortex -- they are amazing, but there are intriguing possibilities yet unexplored. These devices collect data and then build models that exist in some abstract parameter space. How about morphing between these parameter sets to morph between different kinds of amps? Are there profiling technologies that could let the users sculpt the underlying abstract parameters in an intuitive way to create new tones? Perhaps the underlying tech of the Kemper and the Quad Cortex would allow that, and they just haven't tapped into it that way yet. Or perhaps new tech needs to be invented to explore such possibilities. (Acustica Audio apparently has some tech based on Volterra series that looks like it might be interesting.)
Every time I look at any of these devices -- or pedals like Strymon -- I think "wow, it would be REALLY cool if users could incorporate their own algorithms." Like, imagine a SDK where a user could write their own code for a Fractal Audio block and mix and match it with Fractal's already extensive library. Something somewhat analogous to the Net Yaroze SDK that Sony put out for the original Playstation. Now, would it make any business sense for Fractal to go through the effort to make an SDK that integrates end-user code with their existing proprietary closed source blocks? Probably not, but it's fun to think about. My "DIY?" video was born out of these kinds of musings.
On another note, I wanted to thank many of you who said kind things about some of the other videos on my channel. Much of that is part of my "Guitar Amplification and Effects" class. Over the next couple of weeks I'll be covering effects pedal circuits, and then will return to tubes for a discussion of pentodes and power amps. If you are interested in seeing the full set in sequence, the playlist is here:
It starts with the equations of vibrating strings, then talks about pickups & pickup wiring, and then launches into triodes, preamps, etc.
If you are interested in synthesizers, I also have a playlist up for my "Analog Circuits for Music Synthesis" class:
If/when I find time, I was thinking of doing some videos on some of the more obscure parameters in the Fractal Audio amp block, pulling up schematics of real amps and talking about how those parameters map to what's on the schematics (or at least my best guess).
Also, I should be clear that I never said I was going to do this. I'm not even sure what this is. It was just food for thought. But now that I've seen so much interest, I'm inclined to pursue it more seriously.
Or, "Watch this!"
C'mon guys... "Are you silly? I'm still gonna send it"
Hi there! Aaron Lanterman, said Georgia Tech professor and poster of the video in question chiming in here.
I'm pleased that my musings have generated so much discussion!
To clarify, of course the results of such an effort wouldn't equate to what's in one of Fractal Audio's boxes or a Quad Cortex. And of course most of the cost of units like these is in recouping the massive cost of the many hours that were spent developing the software inside of them.
I received my FM3 last week and love it. (And I plan to pick up a Quad Cortex if it ever claws its way out of backorder hell). So, no, the intent is not "let's do this and some magic will happen and we will save money on buying an Axe FX-III or Quad Cortex."
Why would someone want to engage in a DIY DSP amp modeling effort? Because it can be (a) fun, (b) educational, and (c) allow someone to implement weird new things Fractal Audio and Neural DSP might not think of. The idea for the video started with me simply being curious about how much horsepower was in those boxes, to get a target to shoot for in terms of available horsepower.
Why do people build studio gear from kits, like the Hairball Audio 1176, when they can buy a ready made 1176 from Universal Audio or a clone from someone like Warm Audio? Why do people build amps from kits provided by people like CeriaTone -- or build from scratch -- when they could buy from Two-Rock or Red Plate? When you factor in the cost of your time, you're not really saving any money, but it can be fun, educational, and be a gateway into designing your own hardware.
I'm intrigued by how deep the programming on the FM3 goes -- I see things like "Cathode Follower Grid Clipping" and "Power Tube Bias Excursion" and I cannot wait to dig into it more. I find myself wondering what's happening under the hood, and how these parameters map to the various models of the various amps.
And then I imagine other possibilities, like, what about a piece of software that lets the user actually construct the amp at the component level by dragging circuit elements onto a schematic? Would that sell? No. Would lots of people use it? Probably not. Would it make any sense for a company like Fractal Audio to invest time in such a thing? Obviously not. But I think it would be fun and a great tool for learning about electronics.
So much focus has been on modeling existing amps, but devices like the Axe FX-III go beyond that, allowing you to create amps that have never existed before. That's fascinating to me. It's like you could prototype a new amp and tweak it in software before building the actual new amp.
And think about the Kemper and the Quad Cortex -- they are amazing, but there are intriguing possibilities yet unexplored. These devices collect data and then build models that exist in some abstract parameter space. How about morphing between these parameter sets to morph between different kinds of amps? Are there profiling technologies that could let the users sculpt the underlying abstract parameters in an intuitive way to create new tones? Perhaps the underlying tech of the Kemper and the Quad Cortex would allow that, and they just haven't tapped into it that way yet. Or perhaps new tech needs to be invented to explore such possibilities. (Acustica Audio apparently has some tech based on Volterra series that looks like it might be interesting.)
Every time I look at any of these devices -- or pedals like Strymon -- I think "wow, it would be REALLY cool if users could incorporate their own algorithms." Like, imagine a SDK where a user could write their own code for a Fractal Audio block and mix and match it with Fractal's already extensive library. Something somewhat analogous to the Net Yaroze SDK that Sony put out for the original Playstation. Now, would it make any business sense for Fractal to go through the effort to make an SDK that integrates end-user code with their existing proprietary closed source blocks? Probably not, but it's fun to think about. My "DIY?" video was born out of these kinds of musings.
On another note, I wanted to thank many of you who said kind things about some of the other videos on my channel. Much of that is part of my "Guitar Amplification and Effects" class. Over the next couple of weeks I'll be covering effects pedal circuits, and then will return to tubes for a discussion of pentodes and power amps. If you are interested in seeing the full set in sequence, the playlist is here:
It starts with the equations of vibrating strings, then talks about pickups & pickup wiring, and then launches into triodes, preamps, etc.
If you are interested in synthesizers, I also have a playlist up for my "Analog Circuits for Music Synthesis" class:
If/when I find time, I was thinking of doing some videos on some of the more obscure parameters in the Fractal Audio amp block, pulling up schematics of real amps and talking about how those parameters map to what's on the schematics (or at least my best guess).
Also, I should be clear that I never said I was going to do this. I'm not even sure what this is. It was just food for thought. But now that I've seen so much interest, I'm inclined to pursue it more seriously.
Also, I should be clear that I never said I was going to do this. I'm not even sure what this is. It was just food for thought. But now that I've seen so much interest, I'm inclined to pursue it more seriously.
Kemper already has morphing.
According to the manual, Morphing only works on parameters that have continuous values (delay, drive, modulation, etc). It's not clear to me if the Profiles themselves can be Morphed.
