What would the new-gen even have that the current gen doesn’t?

Peavey Herman said:
I don't think that's what they mean. I read it like "Cliff can tell AI to code a better firmware".
Click to expand...

For niche and novel work like DSP programming ? Not a chance. LLMs don't come up with breakthrough.

Also 0 chance to have AI to create patches inside a unit in the near future. I guess FA could train models to create patches based on user inputs and let it access axe edit, so you'd go on the online UI and plus your unit on your computer. I doubt that would be any better than experienced users though.
 
alan.thomas said:
For niche and novel work like DSP programming ? Not a chance. LLMs don't come up with breakthrough.
Click to expand...
Yes they are limited in this regard, they need a body of knowledge to draw from. They can be very helpful to suggest things about known APIs, libraries, save time on repetitive tedious tasks, but not the type of things fractal are doing.
 
I wonder if AI could create a model from a scanned schematic more quickly than doing that by hand. You'd definitely want a human to check and refine it, but it might be a useful starting point. Maybe.
 
Dave Merrill said:
I wonder if AI could create a model from a scanned schematic more quickly than doing that by hand. You'd definitely want a human to check and refine it, but it might be a useful starting point. Maybe.
Click to expand...

I don't know what FAS day to day is like, but my guess is they have this part down pretty well. I assume the hard part would be coming up with new/improving algorithms, performance etc.
 
alan.thomas said:
I don't know what FAS day to day is like, but my guess is they have this part down pretty well. I assume the hard part would be coming up with new/improving algorithms, performance etc.
Click to expand...
Not clear.
I'm sure they have it down in the sense of having a clear process, but if it took less developer time it might be possible to crank out more of the amps folks request.

Just to be clear, there are a ton of amps in there already. I personally don't consider needing more of them or particular missing ones a serious problem, at all.
 
Dave Merrill said:
Not clear.
I'm sure they have it down in the sense of having a clear process, but if it took less developer time it might be possible to crank out more of the amps folks request.

Just to be clear, there are a ton of amps in there already. I personally don't consider needing more of them or particular missing ones a serious problem, at all.
Click to expand...

Speaking of a ton of amps, I would like to see single amp models with all the available modes, switches, inputs, and channels condensed into one single model that you then select what to engage/disengage from the “front” panel.

I imagine a lot of amp models go untouched due to there just being so many options to load and then start from scratch on each one.
 
Here’s what I’d like to see:

1. More digital io, especially a Dante port. Possibly even make additional digital io an optional module, similar to modules on high end converters like the Lynx Aurora. That would give users the option of only purchasing the digital io they need, keeping costs down for all.

2. A plugin that allows me to control the Fractal HW directly in Pro Tools. It would also speed up workflow by eliminating the need to print tracks before they are 100% done
 
Last edited:
charley said:
A plugin that allows me to control the Fractal HW directly in Pro Tools. It would also speed up workflow by eliminating the need to print tracks before they are 100% done
Click to expand...
It's a shame what happened to Glenn. AxeCapture is super legit. Granted, it's not a full control plugin, it just stores your current patch in the plugin instance so you can restore it to your Axe with one click.

Something like that but 1st party as an AAX or VST would be pretty sweet though.
 
vangrieg said:
I’m pretty sure that won’t happen. Dante is a proprietary protocol and I believe you have to pay licensing fees for using it.
Click to expand...
Interesting. I thought it was just a matter of buying and installing a Bkln card and an Ethernet port?

As someone who uses Dante in my own studio, I pay a subscription annually to use the Dante protocol, but not on a per item basis. The subscription fee is the cost of running the AoIP network in my studio and using the Dante Controller software to manage it. It’s about $75 annually. I am
not aware of any sort of fee a manufacturer pays to install Dante cards and Ethernet ports into their gear. The only cost is the cost of parts. I think the BKLN3 card is like $2-300, and the Ethernet port is probably a couple bucks.
 
Zedhed said:
What happened to Glenn?
Click to expand...
Apparently he had a conflict with Fractal and left, also taking down the download links for his software in the process.


rossipedia said:
Something like that but 1st party as an AAX or VST would be pretty sweet though.
Click to expand...
Edit apps could be compiled and built as plugins, exposing controls to a host as parameters. Or they could be standalone but with midi assignable controls.

There doesn’t seem to be an interest from Fractal in any of this though, this has been a recurring request since times immemorial. Probably they don’t see the benefit high enough to justify the support costs if they do it.
 
charley said:
I am
not aware of any sort of fee a manufacturer pays
Click to expand...
I don’t claim to know all the details but as far as I know manufacturers pay license fees to Audinate, maybe just through purchasing a chip though. What you’re saying about yourself paying is their new licensing scheme where the cost is shared between the manufacturer and the user, but it’s still not free, I think.

There are open source alternatives to Dante, but they aren’t as widely supported.
 
LucasLeCompte said:
I really do not want a touch screen at all. It leaves too much that can go wrong.
Click to expand...
biggness said:
What can go wrong?
Click to expand...

I had to reply to this...

I have a 10 year old $30k SUV with a touchscreen on the entertainment system. It controlled the radio, Bluetooth, and even mechanical/tire pressure/mpg monitoring and maintenance schedules for the car.
Slowly, that touchscreen would have spots that didn't work. Today, the entire touchscreen is unresponsive after years of sun. My only solution is to replace the radio with something 3rd party that will cause me to lose the vehicle monitoring and MPG detail functionality.

I also have a 1 year old $3k laptop with 4K monitor and touchscreen. Already, I've sent this laptop in twice because the touchscreen has stopped working.

Please, understand that the tiny displays on phones are very expensive and the ones that aren't do not last. Eventually they ALL fail, and it's not like just soldering in a new POT. There is too much risk and too much expense and support nightmares for too little benefit. I seriously doubt any touchscreen will last more than 10 years of gigging with the abuse musicians put them through. Do you really want to be forced to buy new $$$$ hardware every time the touchscreen goes? Google "neural dsp touchscreen failure" and see what to expect.
 
Budda said:
If the human has to check it, might as well do it right the first time.
Click to expand...
I've been a programmer for decades.
I needed to do a somewhat tricky database update, which i didn't know exactly how to do without trying it, but i knew i could figure it out, since i understood what needed to happen, knew the different pieces involved, and conceptually how they'd need to fit together.

I asked chatgpt a carefully phrased question, and it spat out a fully working chunk of sql in less time than it would have taken me to type it, if I'd known exactly how to do it off the top of my head, which i didn't.

Of course that's totally different than analyzing schematics or writing amp sim code, but there might be some value there.
 
Dave Merrill said:
I've been a programmer for decades.
I needed to do a somewhat tricky database update, which i didn't know exactly how to do without trying it, but i knew i could figure it out, since i understood what needed to happen, knew the different pieces involved, and conceptually how they'd need to fit together.

I asked chatgpt a carefully phrased question, and it spat out a fully working chunk of sql in less time than it would have taken me to type it, if I'd known exactly how to do it off the top of my head, which i didn't.

Of course that's totally different than analyzing schematics or writing amp sim code, but there might be some value there.
Click to expand...
How long did it take to verify 100% it hadnt made mistakes
 
Budda said:
How long did it take to verify 100% it hadnt made mistakes
Click to expand...
Not long actually.
There wasn't that much data that needed to be updated in this way, and reviewing those records by eye was quick, clear, and sufficient.

When i said this update was "tricky" i may have overstated it.
The details don't matter here, but I'd say it was non-trivial, for me, but well below rocket surgery.
I could have written it myself, but this was clearly quicker.
 
Budda said:
How long did it take to verify 100% it hadnt made mistakes
Click to expand...
💯 So true. It spits out code and I’m like, nope, that’s not gonna work. It’s really good at making UIs. Just tell it your language and the frameworks you’re using and you have a good starting point. It’s essentially a junior dev where you as the senior dev need to check their work. Even though the output isn’t always on point it does save hours of work.
 
You must log in or register to reply here.
Back
Top Bottom