IK Multimedia TONEX

[Byrdman]

...haven't even tried NAM.

Or did and said Fuck This! I want to play guitar, not geek out for a few hours trying to get something to work properly.

For all the shit IK has taken over the past 2 years with their UI, and still some of the stuff makes me say WTF, they have a box that hooks up and works within 5 minutes of showing up on your doorstep.

 

[paisleywookiee]

Or did and said Fuck This! I want to play guitar, not geek out for a few hours trying to get something to work properly.

For all the shit IK has taken over the past 2 years with their UI, and still some of the stuff makes me say WTF, they have a box that hooks up and works within 5 minutes of showing up on your doorstep.

The best thing I can say about NAM is that I was able to easily make a few quick caps, and load them into Genome, and that was that. But the hardware is definitely the Wild West right now.

 

[Orvillain]

Someone sent me some NAM recordings to train once. It wouldn't train. Could not for the life of me figure it out. Something to do with file length complaints, but I checked and they were the right length. Maybe a sample-rate mismatch, I dunno. But it aint plain sailing, nosir.

 

[paisleywookiee]

Someone sent me some NAM recordings to train once. It wouldn't train. Could not for the life of me figure it out. Something to do with file length complaints, but I checked and they were the right length. Maybe a sample-rate mismatch, I dunno. But it aint plain sailing, nosir.

Yeah they both have to be 24/48 mono, exactly the same length.

 

[EOengineer]

I’m burned out from work and excited to blow $30 and see what some of these amalgam caps are all about this weekend.

I’m mostly just interested in creating my own captures but I want to hear what he was able to pull off.

 

[Dom Mcsweeney]

Someone sent me some NAM recordings to train once. It wouldn't train. Could not for the life of me figure it out. Something to do with file length complaints, but I checked and they were the right length. Maybe a sample-rate mismatch, I dunno. But it aint plain sailing, nosir.

if you still have the audio files, I'd be happy to take a look

 

[Dom Mcsweeney]

Yeah they both have to be 24/48 mono, exactly the same length.

Since the release of v0.10.0 of the NAM trainer, the trainer now accepts inexact reamp file lengths upto 1sec/1000ms longer.
It won't accept a shorter file though, so it needs to be at least the same length or just over.
This is helpful for people who didn't untick the additional tail length settings in their DAWs before rendering.

 

[EOengineer]

It would be nice if there was some sort of utility that could be used to upgrade existing caps to V2. This is perhaps another area where NAM has a leg up on IK in that NAM lets you hand it a clean guitar DI file and the amped audio file and it will build a cap from the files. It’s nice to be able to handle the source files.

Cap makers can just save their audio files when they do their micing sessions and then basically batch automate kicking out new V2, V3, V4, etc captures. Tonex seems to abstract all that away and I’m realizing that may ultimately prove itself a mistake.

I’m glad I’m primarily into capture tech for capturing and replaying my own amps because this really doesn’t make sense to me from the standpoint of a capture creator.

 

[MirrorProfiles]

It would be nice if there was some sort of utility that could be used to upgrade existing caps to V2. This is perhaps another area where NAM has a leg up on IK in that NAM lets you hand it a clean guitar DI file and the amped audio file and it will build a cap from the files. It’s nice to be able to handle the source files.

Cap makers can just save their audio files when they do their micing sessions and then basically batch automate kicking out new V2, V3, V4, etc captures. Tonex seems to abstract all that away and I’m realizing that may ultimately prove itself a mistake.

I’m glad I’m primarily into capture tech for capturing and replaying my own amps because this really doesn’t make sense to me from the standpoint of a capture creator.

I think going forward that may be easier for IK to do, but a big part of what's possible from training is dependent on what is in the training file to begin with. So if there is some information not present in the reamp, the training might not stand to improve any further (at least in a meaningful way). On V2 the training files get saved and archived locally on the computer so I'm sure it's possible to retrain off the same reamps easily now.

(Preaching to the converted on this forum but maybe someone from IK will give in) One aspect IK really should get a grip on (that can be retro-added to existing models) is the input level thing, NAM has demonstrated clearly how advantageous it is on a platform that has the same issues of everyone capturing using TOTALLY different gear and levels. When using captures on a HW device, having that parity between what someone hears in the studio, and what they hear coming out of their pedal is so easy to solve with metadata. Instead we still have to deal with guesswork and fudging around, largely on models where NO-ONE even knows what the original levels are.

 

[EOengineer]

I think going forward that may be easier for IK to do, but a big part of what's possible from training is dependent on what is in the training file to begin with. So if there is some information not present in the reamp, the training might not stand to improve any further (at least in a meaningful way). On V2 the training files get saved and archived locally on the computer so I'm sure it's possible to retrain off the same reamps easily now.

(Preaching to the converted on this forum but maybe someone from IK will give in) One aspect IK really should get a grip on (that can be retro-added to existing models) is the input level thing, NAM has demonstrated clearly how advantageous it is on a platform that has the same issues of everyone capturing using TOTALLY different gear and levels. When using captures on a HW device, having that parity between what someone hears in the studio, and what they hear coming out of their pedal is so easy to solve with metadata. Instead we still have to deal with guesswork and fudging around, largely on models where NO-ONE even knows what the original levels are.

Actually from what I’m watching right now I suspect the capture files are just wrapping the source and target wav files with some additional instructions - so I’m thinking I’m wrong. I bet you can keep those around and just retrain them through future iterations. Just guessing though.

 

[2dor]

I think going forward that may be easier for IK to do, but a big part of what's possible from training is dependent on what is in the training file to begin with. So if there is some information not present in the reamp, the training might not stand to improve any further (at least in a meaningful way). On V2 the training files get saved and archived locally on the computer so I'm sure it's possible to retrain off the same reamps easily now.

(Preaching to the converted on this forum but maybe someone from IK will give in) One aspect IK really should get a grip on (that can be retro-added to existing models) is the input level thing, NAM has demonstrated clearly how advantageous it is on a platform that has the same issues of everyone capturing using TOTALLY different gear and levels. When using captures on a HW device, having that parity between what someone hears in the studio, and what they hear coming out of their pedal is so easy to solve with metadata. Instead we still have to deal with guesswork and fudging around, largely on models where NO-ONE even knows what the original levels are.

Yeah - the "input calibration" is definitely a must nowadays especially given how well it works in NAM.

 

[EOengineer]

Yeah - the "input calibration" is definitely a must nowadays especially given how well it works in NAM.

It’s weird they haven’t addressed this yet. It’s a glaring issue that impacts quality, consistency, and is probably the single biggest contributor to tonenet being filled to the brim with garbage caps that will not transfer across systems.

 

[BenIfin]

I told you all it wouldn't be far off

 

[BenIfin]

..... and here's " TEMU Leo " now also jumping on the same bandwagon ....

 

[paisleywookiee]

On my Mac M1, it's actually far, far worse in terms of time on the more advanced training. I gave up. It ran for a couple hours, and was nowhere near being done.

 

[Chocol8]

I told you all it wouldn't be far off

The best part of that video is he quietly demonstrates how useless his null tests are. "Look, if we use a different input file we get materially different results." Uh, no shit, and neither of his test files is a very comprehensive look at the amp response to normal guitar playing.

 

[Chocol8]

On my Mac M1, it's actually far, far worse in terms of time on the more advanced training. I gave up. It ran for a couple hours, and was nowhere near being done.

Hmmm, that's an interesting result. If I get a chance this weekend (wife out of the house for long enough) I will try to train the same capture on my M3 MacBook and my PC with Nvidea and see how they do.

 

[paisleywookiee]

Hmmm, that's an interesting result. If I get a chance this weekend (wife out of the house for long enough) I will try to train the same capture on my M3 MacBook and my PC with Nvidea and see how they do.

Yeah, it's basically unusable, if I want to do advanced training. I guess I could install ToneX on my HPE Z4, with an nVidia Quadro. It's older though.

Honestly, I zipped through four NAM profiles on Tone3000, and then decided to throw in the towel on Tonex. It's a bummer, because what I did yesterday on Tonex sounded quite good.

 

[Chocol8]

It's a bummer, because what I did yesterday on Tonex sounded quite good.

I don't think it should be taking that long, so I am curious if other M1 users are having similar results, or if there is a bug.

 

[paisleywookiee]

I don't think it should be taking that long, so I am curious if other M1 users are having similar results, or if there is a bug.

I can try again later. I just got annoyed.