NAM: Neural Amp Modeler

[Sedaxel]

For me, NAM is fantastic when you want a lightweight version of your own amp. In my world, it's just this, full stop.

My captures are only useful to me (so to speak), because I tweak my amp for the sound I need, with my specific guitar and effects. And this sound is not necessarily a standard setup of my amp or for usual rock tones... So, if someone else loads my capture, and he doesn't put the post EQ that I use, and he uses a Strat instead of a PRS, the sound won't be like the one I get. It would sound, probably, as a weirdly tweaked amp. If someone would want to play a capture of my amp with a "more standard" setup, I should tweak the amp on that way and generate a new capture. And still, nobody would know that unless I make a good description of the capture for people to read and understand (which AFAIK is very uncommon). Otherwise, people won't know what kind of amp setup they're playing at all.

And I'm not even talking about the input gain Kaos of online captures.

So I have zero interest in tonehunt or others captures. And it's not because I might think they're rubbish (of course I don't), but because I don't want to guess what setup of the amp was used, and what the input gain was.

The only way for me to be interested in that, would be the kemper approach. Standardized gain, professional made profiles with description of each setting of the amp for each profile. And that's in my opinion the biggest advantage of Kemper... There are tons (and then another ton of tons) of professional quality profiles, with full setup spectrums of each amp, and zero guesswork.

 

[2dor]

For me, NAM is fantastic when you want a lightweight version of your own amp. In my world, it's just this, full stop.

My captures are only useful to me (so to speak), because I tweak my amp for the sound I need, with my specific guitar and effects. And this sound is not necessarily a standard setup of my amp or for usual rock tones... So, if someone else loads my capture, and he doesn't put the post EQ that I use, and he uses a Strat instead of a PRS, the sound won't be like the one I get. It would sound, probably, as a weirdly tweaked amp. If someone would want to play a capture of my amp with a "more standard" setup, I should tweak the amp on that way and generate a new capture. And still, nobody would know that unless I make a good description of the capture for people to read and understand (which AFAIK is very uncommon). Otherwise, people won't know what kind of amp setup they're playing at all.

And I'm not even talking about the input gain Kaos of online captures.

So I have zero interest in tonehunt or others captures. And it's not because I might think they're rubbish (of course I don't), but because I don't want to guess what setup of the amp was used, and what the input gain was.

The only way for me to be interested in that, would be the kemper approach. Standardized gain, professional made profiles with description of each setting of the amp for each profile. And that's in my opinion the biggest advantage of Kemper... There are tons (and then another ton of tons) of professional quality profiles, with full setup spectrums of each amp, and zero guesswork.

While I did come across a handful of NAM profiles others have shot which I liked, I tend to agree - NAM's greatest strength is being able to shoot pixel-perfect profiles of the gear you own & love (for a mix, for live, for songwriting etc.).
A dedicated hardware unit that could at least handle the reamping piece and playback of profiles would take it to another level.

 

[Sascha Franck]

And it's not because I might think they're rubbish (of course I don't), but because I don't want to guess what setup of the amp was used, and what the input gain was.

Personally, I don't think it's that much work, even if the "required" input gain is all over the place between captures. I mean, I don't care whether I'm using some captures "wrong" as long as they sound great (as you know, I don't care much about authenticity), also, after all turning the input gain knob of whatever plugin isn't too much of a big deal once you've recorded something.

But, and that's my main complaint about all NAM plugins: In case I'm going through these "guesswork-and-let-my-ears-be-the-judge" things, allow me to sort out files much quicker. Allow me to very, very quickly go through a folder with large amounts of files and easily listen to them, put some in quarantine, delete others and mark yet some others as "great". Also offer some "auto save input gain" function for this kinda pre-sorting process.
That'd make sorting out myriads of files heaps easier.

 

[Sedaxel]

I don't care whether I'm using some captures "wrong" as long as they sound great (as you know, I don't care much about authenticity),

Yeah, I think that's perfectly fine.

But when I want to go that way, instead of searching for the most accurate representation of an amp, I just feel that Sims are much more convenient, and the sound is good. So I don't even bother with long tries of captures and guessing work.

When I want a JCM800 tone, I've never went to tonehunt to search for a capture. I much prefer the simplicity of a simulator, choose the model, tweak a little and done. I'd have the same doubts about its accuracy anyway.

Unless I had a Kemper. In that case I'd choose a JCM800 collection of profiles, check 4 or 5 of them, get the one I like the most, and done.

 

[Sascha Franck]

But when I want to go that way, instead of searching for the most accurate representation of an amp, I just feel that Sims are much more convenient, and the sound is good.

Basically I agree and it's what I'm usng most of the time. However, I found some NAM captures that are most excellent, especially when "enhanced" by the capture loading block in Genome. I'm only not using it much because there's zero ways to organize captures.

 

[Elric]

The tech is amazing, vst players are good enough (we dont need more imo) but I'm still not using it much....

Like everyone says it really needs to be harnessed properly and ideally placed into a modelling/fx ecosystem. IMO 70-80%+ of the reason playing back NAM profiles is hit and miss is because of random capture levels. A hardware capturer would solve this or a software guided trainer checking levels along the way would also solve it (in my mind like if NDSP made the tonex/tonocracy walkthrough). Even with all that you're still at the mercy of the user dictating "whats good", and then you also only have so much lateral movement with a single profile (the lateral movement with gain and EQ is actually respectable, but still nothing like modelling).

If something can even nail all that stuff then we still ideally want custom routings / fx / stompboxes / reverbs etc etc.

It takes me 3 seconds to fire up an ampsim and have a fully modelled amp at my fingertips with everything else ready to be engaged to build a final tone. When so much of NAM is hit and miss its kind of a cool thing I like to mess around with but aren't using day to day. Where it shines for me is when I grab an amp and reamp a tone for a track, I'll grab a snapshot of that exact tone to pull up later if something needs to be retracked I can slot it in very easily. I guess I can also just pull up those tones for messing around with, but they've been created around a mix already so its hit and miss in the cold light of day.

TLDR Best capturing tech but needs to be harnessed top to bottom by a new vst, or adopted into an existing mature platform.

+1. The incremental improvement is not worth it with the inconsistency across captures.

It might be cool to capture my own stuff, at least the levels would be consistent but my 'screwing around with guitar gear' time is limited and the capture process is not automated enough for me; especially when I have a QC sitting here that can nail everything in a few minutes with a line out from the amp, it's so easy and is well within my current standard for accuracy.

NAM needs a hardware standard or standardized capture process to be more than a 'nerds only grass roots' type thing.

 

[northern_fox]

Too bad the Dimehead unit failed so badly on many levels (IMO at least) because in terms of latency and accuracy it's simply astounding. Could've easily been the best thing ever since the KPA, but unfortunately they apparently didn't ask the right folks for their opinions beforehand.

I've used it on a minimal board on a bunch of gigs and it's easily been my favorite in terms of sound and feel of all my other modelers. Just depends on what you want really - it wasn't ever a full featured multi-fx in the first place, I think the original idea was a pretty literal adaptation of the native NAM plugin to hardware. They've since expanded on that a bit, but it's a limited platform for sure. If you need a bunch of FX or a ton of different patches for songs or are used to the convenience features of flagships, it's not a good choice. But if (like myself) you use maybe 1 or 2 sounds for a whole gig and minimal FX then none of that is really important. I will say there's something about the simplicity and "rawness" of the whole setup that I've come to appreciate after years of using full modelers - that's not really any selling point, but it's just something I like about it (applies to tonex as well).

The price was mostly dictated by the cost of hardware. It may not be cost-competitive, but only you can decide if that's a dealbreaker for you. For me, it was definitely worth it just to be able to use my captures live and not have to deal with tonex - my FM3 is literally collecting dust in my closet.

 

[Sascha Franck]

Just depends on what you want really - it wasn't ever a full featured multi-fx in the first place,

I never expected it to be. But there's some hardware design decisions that are just plain bad IMO. And it could've been different easily.

But if (like myself) you use maybe 1 or 2 sounds for a whole gig and minimal FX then none of that is really important.

In that case, why have 4 switches crammed into too little space?

 

[northern_fox]

I never expected it to be. But there's some hardware design decisions that are just plain bad IMO. And it could've been different easily.



In that case, why have 4 switches crammed into too little space?

I don't need 4 switches - 3 at most is what I generally use. I'll grant you they're close like the QC - I'd rather it had 3, but I haven't had any issues regardless. Next gig I'm actually going to leave it back on the amp and just run my small midi controller up front which is spaced wider and takes less space then my board.

 

[northern_fox]

Also I genuinely appreciate the capability to organize my captures how I want and just simply drop them on a USB and pull them into the player. I can understand how that's not appealing to some people who want full software and editors - but frankly, it's actually faster and simpler than even using tonex because of their poor design. At first I thought it was gonna be a pain, but I actually quite like it now. Not saying it wouldn't be great to have a software side, but just that its vastly less annoying than using the TX ecosystem. Owning your own content (nams and IRs) outright vs having them tied up in the cloud like NDSP and IK do lets you organize as you please on your PC too, which is something I've always been put off by with those platforms.

 

[Sascha Franck]

Fwiw, I think that way of organizing captures is fine.
But still, the hardware (or rather: the design) almost sucks. Which is a shame, given the technical guts behind it.

 

[2dor]

Also I genuinely appreciate the capability to organize my captures how I want and just simply drop them on a USB and pull them into the player. I can understand how that's not appealing to some people who want full software and editors - but frankly, it's actually faster and simpler than even using tonex because of their poor design. At first I thought it was gonna be a pain, but I actually quite like it now. Not saying it wouldn't be great to have a software side, but just that its vastly less annoying than using the TX ecosystem. Owning your own content (nams and IRs) outright vs having them tied up in the cloud like NDSP and IK do lets you organize as you please on your PC too, which is something I've always been put off by with those platforms.

My take on that is it should have came with the ability to hook it up to a PC via USB and just transfer files over like it (the NAM player) is the USB drive (which hooks up to your PC).

 

[Sascha Franck]

My take on that is it should have came with the ability to hook it up to a PC via USB and just transfer files over like it (the NAM player) is the USB drive (which hooks up to your PC).

Defenitely. Also, most devices in that price range (and even cheaper ones) feature a very simple onboard audio interface - which can come in very, very handy. I'm not exactly doing all that much mobile recordings, but all my previous and current devices have seen quite some audio interface usage, too (GT-10, Amp Academy, Zoom G3, HX Floor, HX Stomp). These chips seem to cost next to nothing anymore, so that's something I defenitely wanted, especially in case it's a small unit fitting into any bag.

 

[jellodog]

My take on that is it should have came with the ability to hook it up to a PC via USB and just transfer files over like it (the NAM player) is the USB drive (which hooks up to your PC).

They can probably add that in a firmware update. Here's hoping.

 

[2dor]

They can probably add that in a firmware update. Here's hoping.

I'm not sure about that; IIRC the underlying USB hardware needs to be host and slave spec-ed. Right now, the NAM player is a USB host.

 

[Sedaxel]

If some of you are nerdy enough, you should take a look at Mike Oliphant's implementation of NAM for the Raspberry Pi. It's called "Stompbox". His work is outstanding.

He has developed a software that includes full signal chain effects (comps, gates, OD, mod, delay, reverb, EQ, level modules...), freely assignable, and NAM loader plus IR loader. It has MIDI implementation that lets you switch presets, effects on/off, parameters... Plus, it can be controlled with a smartphone, with his app "StompboxUI".

I built it with a Raspberry and a Hifiberry (audio interface). My latency is 3ms (not as impressive as Dimehead, but perfectly good though). Total cost, maybe 150€.

Now, it takes time, some Linux fight, and you must be brave enough to dare to use it live. I'm just using it at home so far, but it's being reliable. I'm probably using it live soon...

 

[jellodog]

I'm not sure about that; IIRC the underlying USB hardware needs to be host and slave spec-ed. Right now, the NAM player is a USB host.

Oh really? Wow!

Okay, I had just assumed that they didn't have time to sort this detail out with the rush to market and that more would follow. That's a real shame.

 

[Elf]

Raspberry

Sounds very cool! What Raspberry did you buy?

 

[Sascha Franck]

I'm not sure about that; IIRC the underlying USB hardware needs to be host and slave spec-ed. Right now, the NAM player is a USB host.

Yeah, this. And while not a scientific proof by any means, the USB-A socket is kinda giving it away, too. That's just *the* (pre-USB-C) standard hosting socket format.

 

[Sedaxel]

Sounds very cool! What Raspberry did you buy?

I've got the 4B. It can perfectly run standard NAM captures.