I don’t use “reference systems.” I use reference reality.
Musicians in one room. Prototype speaker in another. Studio mics, flat preamps, zero compression, zero EQ, zero limiting. Wired straight to the speaker.
If the speaker can’t fool you into thinking the musicians are behind the wall, it dies there. No second chances.
This kills 90% of prototypes before they see a measurement mic.
Their “reference” is a $40k turntable playing a mastered, compressed, EQ’d record made in a different room, on different mics, by a different engineer, 30 years ago.
That’s not a reference. That’s a rumor.
» No recording engineer’s taste.
» No mastering voodoo.
» No compression hiding transient collapse.
» No EQ masking crossover flaws.
» No "golden-ear" electronics flattering the flaws.
» You get raw acoustic truth vs raw reproduction. Period.
1. Live musicians, now. Pass this, or you’re done.
2. My own 15-ips analog tape masters, then later on DSD masters. No NR, no EQ, no limiting. Brutal, but not live.
3. Commercial recordings. If this is your “reference,” you’re designing jewelry, not instruments.
A speaker can measure flat and still lie. My test rips the mask off:
Coloration. Dynamic strangulation. Transient smear. Crossover phase wreckage. Spatial collapse.
You can’t hide behind a Chesky Records track when the violinist is live next door.
Measurements, dispersion, balance, dynamics are all prerequisites. But they’re not the trial.
The trial is: Does the speaker transmit human intent?
If the musicians’ grief, rage, or joy doesn’t hit your gut through that box, it failed. I don’t care if it’s ±0.5 dB from 20 to 20 kHz. It’s a corpse with good specs.
I don’t build Altisonus to play recordings. I build them to resurrect performances. If your design needs a $20k DAC and a familiar audiophile track to sound good, keep it.
Altisonus has to survive a live trumpet at 100 dB SPL with no place to hide.
Note: AI was used only to remove reflections from the eyeglasses.