Video summary

Amp Sims are Over.

Main summary

Key takeaways

Product Review

Product/Technology Reviewed

Neural Amp Modeler (NAM) Architecture 2 (A2) — an open-source, free amp-capture/modeling system positioned as a major upgrade over earlier NAM versions and as a direct challenger to high-end hardware amp-sim boxes (e.g., Quad Cortex) and premium software/plugins.

Key Features Highlighted

  • Open-source + free No subscription, no license manager—described as avoiding a “no hostage situation” setup.

  • Improved accuracy vs. original NAM Presented as “ever so slightly more accurate.”

  • Major efficiency gains

    • A2 runs ~30–40% more efficiently than prior NAM (as claimed by the builder in the video).
    • A2 Light can run on very low-cost hardware (claimed: runs on a ~$3 ARM chip).
    • “Slimmable” model approach: one model can operate at different CPU/accuracy levels.
  • Capture fidelity & “blind test” positioning
    • Claims a blind listening test with 1000+ listeners, using BBC-grade methodology (as referenced).
    • A2 scored a perfect 100, with listeners supposedly unable to distinguish it from the real amp (“tied with reality” is also mentioned).
  • Out-of-the-box usability
    • A giveaway plugin/capture is described as simple to use, featuring:
      • One main gain knob
      • Switches for boost, gate, and impulse response
    • “Serious processing is under the hood,” intended to sound good immediately.

Comparisons Made (Central Theme)

The video frames NAM A2 within broader “hardware wars” and the premium plugin/software era.

Hardware Competition Mentioned

  • Quad Cortex
  • Axe-Fx
  • Kemper
  • Line 6 Helix / Helix Stomp
  • ToneX

The claim: NAM A2 can match or beat these systems in tone while being far cheaper.

Software/Plugin Era Context

  • Early software such as Amp Farm, Guitar Rig, AmpliTube, and ReValver are described as often poor/uneven.
  • Neural DSP is acknowledged as sounding great (e.g., Fortin Nameless, then Quad Cortex).
  • The video later positions open-source NAM as the “real” breakthrough.

NAM A2 vs Quad Cortex (Practical AB + Graphs)

  • The reviewer runs the same amp capture through:
    • NAM A2 (blue)
    • Quad Cortex (green)
  • Frequency response comparison
    • Dead-on match around ~1 kHz, described as a particularly sensitive hearing region.
    • Quad Cortex reportedly has slightly more low-end down to ~20 Hz.
    • NAM reportedly has slightly more top-end (difference described as small—roughly ~½ to ~2 dB, with exact figures presented loosely).
  • Verdict
    • Differences are characterized as minor EQ-level tweaks.
    • The audio conclusion is left to viewers: which sounds better is up to you.

Pros (User Experience / Value)

  • Excellent sound quality Supported by the video’s blind-test claims and AB-match discussion.

  • Very high value: free, open-source.

  • No subscription / no complicated licensing.
  • Low CPU requirements, especially with A2 Light.
  • Builder-friendly / accessible design
    • An interview with Steve Atkinson (builder) says A2 was aimed to be easier for third parties to integrate, including into embedded or limited-CPU devices.
  • Works for both full mixes and home practice (Stormblade example used for both.)

Cons / Risks Mentioned

  • No direct technical “cons” are clearly stated for NAM A2 itself.
  • The video emphasizes market/financial disruption risk:
    • If NAM A2 performs as claimed, paid amp sims become harder to justify.
  • Some broader caution is implied about hardware depreciation in general (not NAM-specific).
  • The video heavily critiques subscription/licensing complexity (especially Waves), but those are treated as industry issues rather than NAM A2 drawbacks.

Ratings / Numerical Scores

  • Blind test score: Perfect 100 (Allegedly indistinguishable from the real amp.)

  • Efficiency: 30–40% more efficient (for full A2).

  • Cost for A2 Light: runs on a ~$3 ARM chip.

Overall Verdict / Recommendation (as Stated by the Video)

  • The video strongly recommends NAM A2, framing it as a replacement for paying for amp sim plugins/hardware.
  • Core message: great amp tone shouldn’t be behind a paywall, and with NAM A2 it isn’t.

“Great amp tone should never be behind a paywall” (as presented in the video)

Unique Points Mentioned About the Product (Consolidated)

  1. NAM A2 is free and open-source.
  2. A2 is more accurate than previous NAM versions (slightly).
  3. A2 is 30–40% more efficient than earlier NAM versions.
  4. A2 Light can run on ~$3 ARM hardware.
  5. “Slimmable” models allow tuning CPU vs accuracy.
  6. Blind listening test: 1000+ listeners, BBC-grade methodology, score 100 (can’t distinguish from a real amp).
  7. AB comparison: frequency response matches closely around 1 kHz; other differences are minor and described as largely EQ-like.
  8. A2-based giveaway plugin is simple to use (gain knob + switches) and designed to sound good immediately.
  9. A2 intended to improve accessibility for musicians and developers integrating it elsewhere.
  10. Claimed impact: makes $1,800 hardware and paid plugins feel unnecessary.

Speaker-Specific Contributions

  • Glenn Fricker (main reviewer/interviewer)
    • Provides historical amp-sim context (Axe-Fx → Kemper → Helix → Quad Cortex).
    • Drives the critique of subscriptions/licensing and positions open-source NAM as the solution.
    • Presents the blind test discussion, AB comparison, and frequency response analysis.
    • Covers Stormblade capture/plugin giveaway details.
  • Steve Atkinson (builder of NAM A2)
    • Explains A2 goals: accessibility, less CPU, slightly more accurate, and slimmable compute options.
    • Confirms the efficiency improvement (30–40%) and design intent for builders/integrators.
  • Christian Bay (referenced musician/collaborator)
    • Appears as part of a segment describing A2 results as especially convincing for edge of breakup and also shred tones.

Original video