1073 Vintage Style Microphone Preamp
Explanation of file names:
`Pre73jr_0HP_0Air_65gain_0out` =
* No high-pass filter
* No "Air" high-end boost
* +65 dB input gain
* 0 output gain
[C300] = COMPLEX, 300 Epochs, trained with a custom fork of NAM.
[RyH300] = REVyHi, 300 Epochs, trained with a custom fork of NAM.
[RyS300] = REVySTD, 300 Epochs, trained with a custom fork of NAM.
[RxS300] = REVxSTD, 300 Epochs, trained with a custom fork of NAM.
[S300] = STANDARD, 300 Epochs, trained with a custom fork of NAM.
[T3K] [S1000] = STANDARD, 1000 Epochs, trained on T3K, no secret special sauce.
Model architectures explained:
STANDARD = Baseline, "normal" NAM Architecture, should play nice with most hardware, about 290 kB.
REVxSTD = Basically the same requirements as Standard, but sometimes better quality, about 290 kB.
REVySTD = Similar to REVxSTD, but possibly better, a bit more demanding, slightly larger file size.
REVyHI = Close to COMPLEX in terms of accuracy, about 480 kB.
COMPLEX = The absolute highest NAM quality! You may need a more powerful computer, and some hardware will not be able to handle this. I recommend using a less demanding Architecture for tracking, and then using COMPLEX for offline rendering. Size is about 1900 kB.
Signal chain:
Ableton Live 12.3 Suite
ZOOM UAC-232 32-bit float 48 kHz 100% gain
ART Dual RDB 100% gain
GAP Pre-73Jr MKII Hi-Z DI input
GAP Pre-73Jr MKII XLR output
Mogami XLR cable
ZOOM UAC-232 32-bit float 48 kHz input 100% gain
Ableton Live 12.3 Suite
Converted to 24-bit integer 48 kHz mono and normalized to -3 dBFS TPL via dBpowerAmp.
Any recordings with excessive noise were de-noised using the Spectral Denoise tool in RX 11, at the highest quality settings.
ZOOM UAC-232 100% output, using a sine wave input at approximately 1.00 kHz measuring just barely under 0 dBFS TPL at the DAW output (double checked with YouLean Loudness Meter 2 Pro) measures 3.518 or 3.559 Volts RMS, which converts to 13.14 dBu or 13.245 dBu, respectively. I list 13 dBu. I used an AstroAI TRMS-6000 multimeter, and triple-checked the measurements.
Computer specs:
ROG Z13 ACRNM
i9 13900H
32 GB LPDDR5
Nvidia GTX 4070 Mobile
Windows 11 Pro
You can use my fork of NAM to make your own "Hyper Accuracy" aka "COMPLEX" architecture NAM profiles, as well as "REVxSTD", "REVyHI", and other Architectures:
https://github.com/sskalnik/neural-amp-modeler
