I thought Id throw this in Chat GPT and see what it would spit out. No idea if this is correct or not so take it with a grain of salt
TLDR - it estimates current profiling would be 2000-5000 points
my question - "Kemper plans to sample 100,000 points across the frequency range in their mark 2 profiling, based on existing patents and known information, how many sample points do they currently use"
Based on the original **Kemper Profiling Amplifier patents**, the number of frequency points used in profiling is not explicitly listed as a fixed number like "X points." However, we can deduce the general resolution and method from the profiling process described in **Kemper's U.S. Patent 8,619,026 B2** and related documents.
---
###
What the Patent Says (Summary):
1. **Frequency Response Capture**:
* The system measures the frequency-dependent transfer function using **test signals** (e.g. impulses, bursts, or swept sine waves).
* The result is a **frequency response** over a broad range (typically 20 Hz to 20 kHz).
2. **Resolution**:
* The patent describes an **FFT-based analysis**, implying that resolution depends on:
* **FFT size**
* **Sampling rate**
* **Windowing function**
* For example, a typical FFT length of 2048–8192 at 44.1 or 48 kHz sampling gives **tens to hundreds of frequency bins**, not 100,000.
3. **Dynamic Profiling**:
* Kemper doesn't rely on frequency-domain data alone. It supplements with **dynamic tests** (short bursts, tone pulses, playing clips) to capture **nonlinear behavior**, which reduces reliance on dense frequency bin coverage.
---
###
Estimated Number of Points (Original Kemper):
While the patent doesn’t say exactly, inference from:
* typical FFT sizes (e.g. 4096-point FFT @ 44.1 kHz gives \~2048 bins, but only \~1000 below 20 kHz),
* burst test responses,
* and the **lightweight profile file sizes** (\~few KBs)
Suggests the **original Kemper used fewer than 2,000–5,000 points** for frequency-dependent profiling data — likely **log-spaced or perceptually weighted** to match amp/cabinet characteristics.
---
###
Contrast With Mark II (Rumored 100,000 Points)
* A massive jump in resolution, likely tied to improvements in CPU/DSP power and a desire to capture **much more detailed IR-like behavior**, possibly moving closer to **full-resolution convolution** territory or hybrid modeling.
---
Would you like a direct excerpt from the patent that outlines their test signal method or the type of FFT analysis they reference?
.
