Pick the real amp out from the emulations

[MirrorProfiles]

I posted the results to mine in the YouTube comments but if you CBA to click:

Spoiler: RESULTS

A Helix
B Amp
C NAM xSTD
D ToneX
E NAM STD
F Amphub

I found it interesting that not a single person identified the real amp, but almost everyone who took guesses could identify Helix (and to a lesser extent STL). I was a little disappointed by the test in some ways - Helix and STL got some characteristics of the amp correct that IMO the ML models couldn't quite get right, but with some trade offs that sound a bit digital (for want of a better word). The ML models all sound fine to me, but ToneX almost sounds like another amp (the character of the gain and the low end response doesn't have what I'd think of as a definitive part of the uber tone). NAM STD was slightly better but still not quite right. xSTD got closer still but when going between the real amp and the xSTD model the low end and upper mids still have some small differences.

Slammin Mofo kindly trained some Hyper Accuracy models for me, so I'm going to do one more video just comparing those against the real amp (and NAM STD, xSTD, and ToneX). This time I'll label the real amp, so there is a frame of reference to compare the captures to. It'll be interesting to see whether HA models close the gap further.

Latest here:

 

[Ipkz]

I posted the results to mine in the YouTube comments but if you CBA to click:

Spoiler: RESULTS

A Helix
B Amp
C NAM xSTD
D ToneX
E NAM STD
F Amphub

I found it interesting that not a single person identified the real amp, but almost everyone who took guesses could identify Helix (and to a lesser extent STL). I was a little disappointed by the test in some ways - Helix and STL got some characteristics of the amp correct that IMO the ML models couldn't quite get right, but with some trade offs that sound a bit digital (for want of a better word). The ML models all sound fine to me, but ToneX almost sounds like another amp (the character of the gain and the low end response doesn't have what I'd think of as a definitive part of the uber tone). NAM STD was slightly better but still not quite right. xSTD got closer still but when going between the real amp and the xSTD model the low end and upper mids still have some small differences.

Slammin Mofo kindly trained some Hyper Accuracy models for me, so I'm going to do one more video just comparing those against the real amp (and NAM STD, xSTD, and ToneX). This time I'll label the real amp, so there is a frame of reference to compare the captures to. It'll be interesting to see whether HA models close the gap further.

Latest here:

To my ears what really stands out is the sound on the fading palm mutes, that's why I said both A and F reminded me of Helix.

They have a blooming effect in some frequencies, almost like the simulation doesn't know what to do with the fading notes.

That's what bothers me the most when I use Helix and I feel that this is reduced in captures (NAM, tonex, Genome and even Neural DSP).

Unfortunately, as you said, some characteristics of the amps are not replicated by captures, meaning component emulation is still very important. I almost bought the Lion plugin this week because I miss that cranked sound with ghost notes, even though captures offer more clarity with my guitar.

Anyway, please keep these going. I like the hyper accuracy slamming profiles sounds, but the m1 chip on my MacBook complains a lot

 

[DrewJD82]

I posted the results to mine in the YouTube comments but if you CBA to click:

Spoiler: RESULTS

A Helix
B Amp
C NAM xSTD
D ToneX
E NAM STD
F Amphub

I found it interesting that not a single person identified the real amp, but almost everyone who took guesses could identify Helix (and to a lesser extent STL). I was a little disappointed by the test in some ways - Helix and STL got some characteristics of the amp correct that IMO the ML models couldn't quite get right, but with some trade offs that sound a bit digital (for want of a better word). The ML models all sound fine to me, but ToneX almost sounds like another amp (the character of the gain and the low end response doesn't have what I'd think of as a definitive part of the uber tone). NAM STD was slightly better but still not quite right. xSTD got closer still but when going between the real amp and the xSTD model the low end and upper mids still have some small differences.

Slammin Mofo kindly trained some Hyper Accuracy models for me, so I'm going to do one more video just comparing those against the real amp (and NAM STD, xSTD, and ToneX). This time I'll label the real amp, so there is a frame of reference to compare the captures to. It'll be interesting to see whether HA models close the gap further.

Latest here:

Whatever clip E was sounded pretty damn spot on to the real amp.

 

[Eagle]

In the room it will be obvious but start adding huge filters and you homogenise everything.
Put a microphone on a 25k hifi and a Bluetooth speaker , play it back on your phone and all the important differences are lost in translation.

 

[MirrorProfiles]

In the room it will be obvious but start adding huge filters and you homogenise everything.
Put a microphone on a 25k hifi and a Bluetooth speaker , play it back on your phone and all the important differences are lost in translation.

IMO the differences may be subtle but they’re definitely audible. Keen to hear your thoughts as you have quite a nuanced take on guitar tones.

In the room there’s so many variables that make it almost impossible to compare without any kind of bias. The experience may be fun in the room, but IMO “what sounds good” has much more validity once it’s under a microphone.

 

[Eagle]

IMO the differences may be subtle but they’re definitely audible. Keen to hear your thoughts as you have quite a nuanced take on guitar tones.

In the room there’s so many variables that make it almost impossible to compare without any kind of bias. The experience may be fun in the room, but IMO “what sounds good” has much more validity once it’s under a microphone.

It depends on what you care about most. The player experience or the final recording. To me if you take care of the player experience first and foremost the rest takes care of itself almost.
If I am inspired by a fantastic dynamic,” alive “ tone I definitely play better which in turn improves the tone even more and is evident in the final recording. Picking the difference between my X88ir and my Axe3 through headphones is going to be difficult but in the fingers it’s easy.
Clean sounds are very easily identifiable as digital. Some people like “ grease “ on the overdriven tone more than others and may even prefer the slight disconnect in the dynamic and easier more forgiving feel but if you have ever played through a real Dumble you would experience that exact thing that today’s digital is no where near capturing.

 

[Eagle]

It’s like viewing a piece of art from a distance and when you get closer you find it’s a print. Viewed in a particular way it looks identical but the closer you get and the more you look you see the lack of texture and three dimensional brush strokes in the paint. Also from the player perspective "FRFR" is different at best. Enlarge a print and no matter how well you do it all you get is bigger dots.
Art is so much more than a superficial 2D print

 

[MirrorProfiles]

Many others have been able to listen and offer an opinion. I thought you might be able to join in.

 

[Sascha Franck]

Fwiw, I only now listened to the first video - and when F came up, I instantly knew I liked this somewhat better than the previous sounds. Turned out to be the real amp. Which sucks, as I'm a die hard modeling believer by now.
However, fortunately the differences are so small, I personally don't care anymore.

It’s like viewing a piece of art from a distance and when you get closer you find it’s a print.

Broken. Record. Repeat. Ad. Nauseum.

 

[Eagle]

Many others have been able to listen and offer an opinion. I thought you might be able to join in.

The other problem is often the real amp is not well recorded and the digital ones are Di again making the comparison irrelevant. In terms on picking out a tone to like the sound of ; sounds good = is good end. If you’re the player it opens up hugely or does if you care about it. But you are essentially pleasing yourself to be inspired.

 

[MirrorProfiles]

The other problem is often the real amp is not well recorded and the digital ones are Di again making the comparison irrelevant. In terms on picking out a tone to like the sound of ; sounds good = is good end. If you’re the player it opens up hugely or does if you care about it. But you are essentially pleasing yourself to be inspired.

The test is about as like for like as I can do. On the most recent video I’ve labelled the real amp, I’m just curious where you personally would draw the line (as there are audible differences with the others, which in turn all affect feel). It takes maybe 5 mins to listen through.

 

[Eagle]

The test is about as like for like as I can do. On the most recent video I’ve labelled the real amp, I’m just curious where you personally would draw the line (as there are audible differences with the others, which in turn all affect feel). It takes maybe 5 mins to listen through.

Ok when I get to some monitors later.

 

[Eagle]

They mostly sound ok with minor EQ differences but a huge part of the deal on the real amp has replaced with an ir . Real amp, real cab in the room compared to a digital version "FRFR" through what ever you think is good is the actual test. The whole premise of the test is not showing how digital is holding up to the real thing.
Running my X88ir directly in to my "FRFR" can’t hold a candle to the analog out in to my VHT and a 4x12”
If you are saying you can use it to record and no one can tell that’s been true for years in various ways from old transistor amps to early digital. It’s a totally meaningless argument. It sure as hell does not prove the quality and dynamics of digital is a match for analog yet.

 

[MirrorProfiles]

Real amp, real cab in the room compared to a digital version """FRFR""" through what ever you think is good is the actual test. The whole premise of the test is not showing how digital is holding up to the real thing.

No that’s not what I was testing here. I’m trying to see what the most sensible trade offs are when using NAM, xSTD, Hyper Accuracy, ToneX etc and you’re talking about "FRFR"’s for some reason.

 

[Orvillain]

The comparison is from a recording perspective, not a playing perspective. I think we all know - deep in our souls - that from a playing in the room perspective, a real amp still trounces all of the digital options on the table.

 

[Mongillo19]

The comparison is from a recording perspective, not a playing perspective. I think we all know - deep in our souls - that from a playing in the room perspective, a real amp still trounces all of the digital options on the table.

Do you think digital device into a good power amp comes close to the real deal?

 

[Orvillain]

Do you think digital device into a good power amp comes close to the real deal?

Close, but no cigar.

I will say though, the best experience I've had along those lines was Axe FX III into a Fryette 100watt Powerstation. I tried the QC at the same time, which is when I discovered just how low output the QC is - even into a 100watt valve poweramp, I had to push the Powerstation twice as hard as I did with the Axe FX III !!!

I tried Helix too. It was okay. But the Axe FX III was a lot better.

 

[Eagle]

The comparison is from a recording perspective, not a playing perspective. I think we all know - deep in our souls - that from a playing in the room perspective, a real amp still trounces all of the digital options on the table.

And I said from a recording standpoint it doesn’t matter. There has been di fake options that work for decades.

 

[Orvillain]

And I said from a recording standpoint it doesn’t matter. There has been di fake options that work for decades.

Right, but I think that's where I disagree. From a recording standpoint, real amps, real cabs, real mics, and real air moving in a real room, still rules.

 

[[Nathan]]

The comparison is from a recording perspective, not a playing perspective. I think we all know - deep in our souls - that from a playing in the room perspective, a real amp still trounces all of the digital options on the table.

I haven’t done any in depth tests but I’ve always been very impressed how well plugins, digital units and preamp pedals sound running through an amps power section into a cab. It instantly falls into the category of “the real deal” to me…. But again it’s not a test really it’s just enjoying things as a player.