NAM support announced by Fractal Audio

[Orvillain]

Cliff knows better than I, but I would guess that they got it running in realtime at low-shit-balls-not-on-fire levels of quality, but from there it was hard to hit the levels required for the feature to be worth bothering.

Running NAM on my DaisySeed is a project I have in the pipeline, just haven't gotten around to it yet!

 

[timbuck3]

This is interesting and I don't quite understand why they would allow NAM to run on their platform.
Seems like there will be TONS of "cheaper" hardware units coming out that will run NAM and presumably lots of them will have a better UI.
I'm thinking that it's about to become a very competitive space.
Fractal stuff is normally expensive, and it just seems like a weird choice to even attempt to get into that market at all.

Seems at this time going forward that having a profile feature(playing or capturing) in a flagship is a bar set by qc in a modelers. As will be touchscreens likely. I mean why not just add the ability to play captures?

 

[Antipodes]

You probably want something like 8 or16 GB of storage on a machine which is going to be working with lots of NAM captures and collections of IRs. How optimised DSP chips are for running NAM is another question. As others mentioned, ARM multicores are going to be the cheap and efficient option for that sort of stuff. Why NAM on FAS? Well you have the I/O, the effects library and the control interface options with the AXE and FM series. Grafting on another feature that the punters want is like adding support for IRs.

 

[paisleywookiee]

Not going back and forth. I did say there will be cheaper hardware units running NAM that will likely come out in the future. I believe that. Pick apart posts however you feel like

I think what he's getting at is that there's nothing outside of the Dimehead and maybe a couple others that actually run NAM profiles without converting them to "crappier" quality. Bringing that in cheaper will be a little challenging, since it requires a decently powerful chip.

 

[KHAAAAAAANNNNNNN!!!!!!!!]

Most units will support some kind of profiling. Fractal sitting that one out entirely seems like an unwise move. Many people want captures, and some enjoy both modelling and captures.

Cliff said they couldn't make it run in real time on the Axe-FX III. He didn't sit this one out.

He doesn't care for captures. He's stated that multiple times. But he supports his customers, and tried to implement it on the current hardware that wasn't designed to run NAM. Apparently he's been able to make that work on the next gen hardware in the works.

When Kemper was an anomaly, it was easy to dismiss it. But hybrid modelling/capture hardware is becoming the norm now. NAM is the perfect way to add it without having to create an entire ecosystem. Line6 is creating their own tech that will need a lot more resources. But that makes sense for them because it could be another steam of $$$ selling captures.

 

[Dimiguitar]

Cliff said they couldn't make it run in real time on the Axe-FX III. He didn't sit this one out.

I know. The comment I was replying to was talking about allowing NAM to run on fractal devices, independently of gen. "Unwise to sit this one out" refered to the possibility of fractal not incorporating NAM at all, which is of course not what they're doing.

(But I understand how what I said can be misunderstood/is unclear.)

 

[mikah912]

Cliff said they couldn't make it run in real time on the Axe-FX III. He didn't sit this one out.

He doesn't care for captures. He's stated that multiple times. But he supports his customers, and tried to implement it on the current hardware that wasn't designed to run NAM. Apparently he's been able to make that work on the next gen hardware in the works.

When Kemper was an anomaly, it was easy to dismiss it. But hybrid modelling/capture hardware is becoming the norm now. NAM is the perfect way to add it without having to create an entire ecosystem. Line6 is creating their own tech that will need a lot more resources. But that makes sense for them because it could be another steam of $$$ selling captures.

I think they're taking fundamentally different approaches. Everything Line 6 has said about Proxy indicates that they want to offer more than a me-too solution. They're also setting up a cloud infrastructure to make training and executing the captures hardware-agnostic. And of course, there are hints that they're looking to capture effects and more as opposed to just drives, amps and cabs.

Cliff has only indicated a narrower focus, and his engineering muscle looks to stay firmly aligned with his modeling.

 

[KHAAAAAAANNNNNNN!!!!!!!!]

I know. The comment I was replying to was talking about allowing NAM to run on fractal devices, independently of gen. "Unwise to sit this one out" refered to the possibility of fractal not incorporating NAM at all, which is of course not what they're doing.

(But I understand how what I said can be misunderstood.)

So you're saying that even though they can't make it work currently. it is wise to do it for upcoming products. Gotcha. I misunderstood.

 

[Dimiguitar]

So you're saying that even though they can't make it work currently. it is wise to do it for upcoming products. Gotcha. I misunderstood.

Yes, that's what I mean. But I was quite clear.

 

[Dimiguitar]

Yes, that's what I mean. But I was quite clear.

* Quite UNCLEAR

 

[LP59]

I think what he's getting at is that there's nothing outside of the Dimehead and maybe a couple others that actually run NAM profiles without converting them to "crappier" quality. Bringing that in cheaper will be a little challenging, since it requires a decently powerful chip.

That's right. The only NAM players out there are the Dimehead at $600 and the Anagram at $1200. That's evidence that it's going to be very difficult for anyone to come out with an inexpensive NAM player.

 

[Sedaxel]

For me, NAM is not the future, but almost the past. Yeah, it´s great... but let´s face it, static capturing is already a mature tech. Capturing has to evolve into hybrid solutions, parametric, tweakable and whatnot.

Besides, NAM is still too CPU hungry to be easily implemented in cheap units. It has to be optimized. Or... wait a couple years to hardware to have enough power for cheap. And gain calibration will never be fully implemented until the user could solve it without studying "how-to" guides.

Fractal featuring NAM support means a risky step for them, I think. At some point, they´ll find sounding the same as a 300€ unit. Sure, they´ll still have over the top sim algos to keep the edge over the competition, but in absolute terms... NAM sounds equally good (if not better). I also think that Line6 has realized about this and that´s why they went for their own capturing tech. Knowing the capabilities of Line6, they´ll surely evolve much faster and will probably take the leadership in that field. If you think about it, all their competitors (in capturing tech, I mean) are much smaller (IK, NDSP, Kemper... NAM). If they manage to build a good capturing platform, in a few years they´ll be ahead of everyone.

Of course, I´m just especulating here.

 

[laxu]

For me, NAM is not the future, but almost the past. Yeah, it´s great... but let´s face it, static capturing is already a mature tech. Capturing has to evolve into hybrid solutions, parametric, tweakable and whatnot.

Besides, NAM is still too CPU hungry to be easily implemented in cheap units. It has to be optimized. Or... wait a couple years to hardware to have enough power for cheap. And gain calibration will never be fully implemented until the user could solve it without studying "how-to" guides.

Plus IK has shown that their Tonex V2 is pretty much on par with NAM yet runs on the cheap Tonex One just fine.

Fractal featuring NAM support means a risky step for them, I think. At some point, they´ll find sounding the same as a 300€ unit. Sure, they´ll still have over the top sim algos to keep the edge over the competition, but in absolute terms... NAM sounds equally good (if not better).

If you just stick with the static settings paradigm of captures as they are. Fractal's modeling will outperform the moment the user wants something different. With captures you then go back to trawling for a different capture or try to EQ the existing one to your liking, which then takes it out from the "sounds like the real thing" territory. Not that it truly matters for the end result though!

Hotone is probably one of the leaders for combining parametric component modeling + machine learning models with the stock amp models on the Ampero 2 range. And you know what, some bugs aside they work! You can use them very much like their real amp counterparts and they will sound in the ballpark, with some caveats like master volume amps not being able to go into poweramp distortion. Interestingly, that's an issue with the ML Amped Volcano as well since it has no master control.

So even for the cheaper stuff there's not some massive gap between the top tier stuff at least for amps. Effects are a whole another thing and that's where Fractal still absolutely rules.

I also think that Line6 has realized about this and that´s why they went for their own capturing tech. Knowing the capabilities of Line6, they´ll surely evolve much faster and will probably take the leadership in that field. If you think about it, all their competitors (in capturing tech, I mean) are much smaller (IK, NDSP, Kemper... NAM). If they manage to build a good capturing platform, in a few years they´ll be ahead of everyone.

Of course, I´m just especulating here.

I think Line6 has figured the key pain points of capturing. Finding what you want is a shit experience on all current platforms where you can never quite know what you get without auditioning everything. While things like custom DI clips help a lot, it's still a pain in the ass.

At the same time processing in the cloud makes a lot of sense when the on-unit capabilities are not suited for complex models, and not everyone has a powerful GPU that is compatible to make the process short. Hell, these days a lot of younger folks don't even have desktop/laptop systems at all!

I'm guessing they want to build something like: Capture the amp on HX Stadium using its I/O, mic pre etc (known input/output system, levels etc) -> send it to cloud for processing -> keep playing while you wait -> capture comes back -> try capture.

 

[AZG]

For me, NAM is not the future, but almost the past. Yeah, it´s great... but let´s face it, static capturing is already a mature tech. Capturing has to evolve into hybrid solutions, parametric, tweakable and whatnot.

Besides, NAM is still too CPU hungry to be easily implemented in cheap units. It has to be optimized. Or... wait a couple years to hardware to have enough power for cheap. And gain calibration will never be fully implemented until the user could solve it without studying "how-to" guides.

Fractal featuring NAM support means a risky step for them, I think. At some point, they´ll find sounding the same as a 300€ unit. Sure, they´ll still have over the top sim algos to keep the edge over the competition, but in absolute terms... NAM sounds equally good (if not better). I also think that Line6 has realized about this and that´s why they went for their own capturing tech. Knowing the capabilities of Line6, they´ll surely evolve much faster and will probably take the leadership in that field. If you think about it, all their competitors (in capturing tech, I mean) are much smaller (IK, NDSP, Kemper... NAM). If they manage to build a good capturing platform, in a few years they´ll be ahead of everyone.

Of course, I´m just especulating here.

I'm just not seeing the risk. It's just another option to make amp/drive/etc. tones to use in a signal chain. Most of the time I will use an amp model, but if I find a great capture I may use it. It's just another option incorporated with all their other excellent effects, routing, etc.

 

[MirrorProfiles]

Capturing has to evolve into hybrid solutions, parametric, tweakable and whatnot.

I don't think this is the case really. I can't really see why end users would need to be able to have fully parametric and accurate models of their own amps. I think the main benefit of capturing tech is situations like being in a recording studio and cataloguing your tone to use or come back to later. Or to be able to take on the road. Making deep parametric models, or even a perfect algorithmic model wouldn't really satisfy this use case, where the ideal is just having a 1:1 snapshot of a particular sound.

It's not dissimilar to IR's - individual IR's are significantly less useful (IMO) to a fully captured cab with various mics and every position. But sometimes a single position and sound thats locked in is required. One doesn't negate the other.

Fractal adding NAM support doesn't really have any downsides. It's there if people want to use it, and it vastly expands what's possible on the HW device. There are likely people already using various capture devices in tandem with Fractal gear - being able to do it all on one device would certainly be an upgrade for those users.

 

[Dimiguitar]

Fractal adding NAM support doesn't really have any downsides. It's there if people want to use it, and it vastly expands what's possible on the HW device. There are likely people already using various capture devices in tandem with Fractal gear - being able to do it all on one device would certainly be an upgrade for those users.

I use tonex and fractal fm3. Had some VH4 captures made that are very difficult to replicate on the FM3, despite my efforts. Even when they sound relatively close, after creating EQ compensation IRs and such, tonex feels considerably better.

But I want that tone on the FM3 too, since it's my main unit. Tonex has been more of a backup (even though it gets a lot of use currently).

Having NAM loading in future units would fix issues like this. Could just load a NAM capture and be done, in scenarios where it's hard to get the desired tone out of the modelled amps.

 

[Sascha Franck]

Plus IK has shown that their Tonex V2 is pretty much on par with NAM yet runs on the cheap Tonex One just fine.

Yeah, it's actually quite amazing just how little CPU Tonex captures consume. Just for giggles I was inserting 15 instances (mono, DI plus IR) on a track in Logic (15 is maximum) and another 15 on a bus I was running the channel into, so that's 30 instances playing serially on a single thread of a Macbook Air M3. All at 44.1kHz, 32 samples buffersize.

Finding what you want is a shit experience on all current platforms where you can never quite know what you get without auditioning everything. While things like custom DI clips help a lot, it's still a pain in the ass.

While I absolutely agree in general, so far I found the Tonex experience to be pretty satisfying. I'm finding myself way less stuck in rabbit holes than back when I was trying to find whatever "perfect" IRs. Sure, when ToneNet is slow, it's just horrible, but at the moments it's fast, it's quite an enjoyable experience. Sure, there's plenty of duds, but I just scroll on and ignore them.

 

[Desertdweller]

Whether it’s a NAM capture or another kind of capture, it’s great to see Fractal moving forward. I love how with the Nano Cortex, I can just run the Capture OUT to my amp head and run the output of my reactive load back to the NC Capture IN. It’s so easy and it’s done in less than 5 minutes. No complex routing like the Tonex, no need to refine over and over again. It just works, and if I hear any variability, I can dial it on the amp and re-capture. My Road King II capture had too much bass, so I just rounded down the bass knob and recaptured, and it was perfect.

If I can have easy amp captures and Fractal effects in a box like the Axe FX or FM9, I’d consider it a near-perfect piece of gear.

 

[Chocol8]

Adding NAM support seems very low risk but also low reward. It's not going to sell a lot of units if any, but maybe it will prevent the loss of sales to competitors. At a minimum, it will get Fractals in to the game and keep possibilities open.

The biggest issue with NAM is how resource intensive it is which is why there are no cheap native players. Tonex solved that, and I am sure Proxy will address it as well, but also through proprietary software. For Fractal, if their next gen machines will already have the horsepower, it makes sense to let it run NAM, and for a small company, it makes sense not to try to develop their own capture tech.

It will be interesting to see how this plays out, because NAM is still flawed, and the rest of the capture market is proprietary and somewhat fractured, but I really do think having top quality capture tech integration is going to be table stakes going forward. The naysayers are loud but stuck in the past. There are too many use cases and the sound quality is too good to simply ignore it.

 

[Desertdweller]

I really do think having top quality capture tech integration is going to be table stakes going forward.

Agreed 100%. It takes a lot of tweaking to get models to sound like amps. It is easy to get captures to sound right the first time, and then you can just focus on playing. More time playing music, less time tweaking and giving yourself ear fatigue.