INTERVIEW WITH PAUL HALES
We put our questions about the structure of the Theory system direct to the company’s President and product designer, Paul Hales.
SOUND+IMAGE: Can we first address the duplication of processing? The Theory system requires a decoding processor prior to the ALC1809, in this case an Anthem AV processor, so there is calibration and optimisation available both in the Anthem and then again in the Theory controller. We gather that here the Anthem calibrated first, then this was fine-tuned with the Theory controller. Is that what you’d recommend?
PAUL HALE: There is no need to utilise the processor for calibration, and in fact we recommend against it to ensure the best audio performance of your Theory system. Our recommendation is to set all levels on the other AV processor to zero, and to use the very capable and user-friendly Automator software on the ALC-1809 — this will yield the best results.
No third-party surround processor-based automated system can properly calibrate a surround system because the calibration scheme has no knowledge of loudspeaker bandwidth or directivity, nor do they have any knowledge of the room acoustic. Plus microphones do not detect the sound in the same way as humans do, so the calibration ‘target curve’ is not what we hear, anyway.
Humans hear with our brains, not our ears. There is a lot of ‘processing’ of the signals that occurs in our brains, and that processing includes adapting to our acoustic environment — we settle in and ‘hear around’ the room instinctively.
Think about it. If you’re having a conversation with someone in a typical conference room, you are hearing the sound of their spoken words first directly from the speaker’s mouth, and then you get bombarded with dozens if not hundreds of delayed copies (reflections) of the words from all the hard surfaces of the room. If our brains did not adapt to ‘tune out’ those copies, we could never hear where the speaker was located, let alone understand what they are saying — our hearing would be overloaded by the copies. We’ve evolved to filter out the reverberation and tune into the sound source itself.
Automated calibration schemes that include room reverberation in their measurements and target are ‘compensating’ again for the room that we have already naturally compensated for. They are redundant, and they don’t work. Worse, if you have well-engineered accurate speakers, they can do nothing but make them less accurate, by definition.
Unlike ‘dumb’ automated systems found in surround processors — “dumb” in the sense that they have no knowledge of the speakers being used — Automator does know all about the speaker, the loudspeaker frequency-dependent directivity, sensitivity, bandwidth, power handling, and so on. It will get great results in all typical installations where the speakers are mounted in the usual locations within the room. Having the knowledge of speaker characteristics sets our systems apart from all the others.
S+I: How about the additional conversion stages in the system — going digital (HDMI) to analogue (Anthem outputs) to digital (ALC-1809 samples) to analogue (ALC-1809 outputs). It’s clearly undesirable extra conversion; why did you choose this doubled-up arrangement over, say, HDMI inputs on the ALC-1809 and doing your own conversion?
PH: Theory has made the explicit decision to not enter the surround sound processor market. We are not interested in that product category. For that reason, we could not take in the HDMI signal directly, because the surround sound format must first be decoded into its individual signals. The world has standardised on surround decoders for this, usually within surround processor preamps or surround receivers. All of these except the very most expensive rely on analogue to output the individual decoded surround signals. Since Theory is a value premium product, we must be compatible with the widest variety of attached equipment — and right now that’s analogue.
With that said, our DLC loudspeaker controllers have S/PDIF digital inputs for stereo signals and as soon as Audinate [maker of Dante] can catch up to demand, we will have full Dante and AES67 compatibility. In the DLC models you can ‘bypass’ the DAC/ADC at the preamp/amp interface. Since it’s unlikely that any affordable surround processors will have Dante or AES67, we’re going to keep seeing the analogue standard for the foreseeable future.
And that’s OK. If you’re getting the most thrilling, enveloping, dynamic and engaging sound possible for a system of that price, if the sound delights you, what does it matter how many domain conversions are taking place? And since you have no control over or can even know how many conversions took place in the recording process, you can’t control the total number of conversions anyway. Forget about the audio hardware — get lost in the content. Theory products are made to deliver the maximum enjoyment of the content, not the gear.
Lastly, the inaccuracy caused by domain conversions, especially if you use high-quality converters as we do, is swamped by the inaccuracies in other parts of the system, especially the acoustic part. Given high-quality converters and high-quality ancillary equipment, I don’t think you would hear the extra conversion. Whereas you would most definitely hear inferior speakers.