Two things a room correction product must do

My research has led me the conclusion that there are a number of basic functional elements that ALL room correction devices should possess. These conclusions derive from a room acoustics viewpoint rather than a sound quality perspective. It is interesting that there are many devices on the market that do not meet these two basic criteria...

1. Must not apply correction filters above the transition frequency OR must allow the user to switch off room correction above the transition frequency

When a measurement microphone is placed at the listening position(s) and the source is a reasonable distance away (e.g. 8ft) what we see on a frequency response chart above around 300Hz (the transition frequency) is a combination of the direct sound from the speakers and the reflected sound from the surfaces of our room. The contribution of the reflected sound from the room to the measured frequency response can be as much as 60%. The measured response, as seen on the chart, is often termed the ‘power response’ of our speakers. The shape of the line normally ‘droops’ towards high frequencies; this is due to the increasing directivity of our tweeters as 20kHz is approached.

Each line on the graph represents a measurement at a different angle off axis. The chart shows 0,30,60 and 90 degree off-axis measurements. The red line is an estimation of the 'power response', or what the listener hears. These speakers maintain very good off axis response through 75 degrees off axis.

Above the transition frequency room correction products are actually correcting the power response of our speakers i.e. they are applying speaker correction. In some cases this can be a good thing, for example with lower quality speakers that possess variable on-axis frequency response. In most cases however, and especially with audiophile quality speakers, one should be hesitant about allowing a room correction product free reign to determine and apply correction filters above the transition frequency. Some reasons for this are listed below:

• Equalizing the power response will typically cause the direct frequency response to rise towards 20kHz; your nice speakers that measure flat on-axis at 1m will now exhibit a rising frequency response!
• Many speaker designers purposely engineer in some minor deviations in the frequency response. This is often a subjective voicing of the sound but is also sometimes required to control the sound quality impact of a driver breakup mode. An example of this is the BBC or 'Gundry' dip (see here).

These examples illustrate why any room correction product should either:

• Only apply correction to bass frequencies like Meridian.

OR

• Allow the end user to select which the frequency range to which room correction is applied like TacT.

2. Must allow the user to choose a target frequency response:

Room correction devices aim to fit the measured frequency response to a nominal target frequency response by applying correction filters. There is an element of preference in the shape of the target curve, some people preferring a ruler flat response whilst others prefer a rising response in the bass which adds a little more weight to music. To some extent this preference is informed by the volume at which a person listens to music and the sound level dependent nature of our hearing. Those who listen at lower volumes (e.g. 70dB) may prefer a lift of up to 10dB at 30Hz. See this information on the Fletcher Munson curves for more information.

In any event any room correction device should either:

• Allow the user a wide choice of target frequency responses

Or

• Allow the user to manually ‘draw’ their target frequency response. Any system that requires the user to specify the correction filters (e.g. Rives Audio PARC, Z-Systems RDP-1) is effectively providing a limited version of this capability.

There are three additional criteria that a room correction product should have: not performing a needless analog to digital conversion, providing measurement capabilities and providing filter generation capabilities. These will be covered in a future blog post.

Do you think there are any other functional criteria that a room correction product MUST HAVE? Let me know through the comments!

Comments

[Kenny B]

Nyal, as a long-time 2ch pure-analog owner, I have to say that switching to a DSP-based RC system has been a revelation. I would suggest that another desirable feature of such products would be an ability to measure the room post-correction. The benefit is obvious: One can see the effect of the filtering i.e. did it do what it's supposed to do? The TacT product "computes" the resultant FR, but that's not the same as accurate measurement, which must currently be done via third party solutions such as REW.

Nice site!

Ken

[Nyal at Acoustic Frontiers]

Hi Ken, that is a great idea for a functional criteria, otherwise how can you possibly assess whether the room correction product did what it was meant to do? There is a nice tie in with the fourth criteria, which states that you should not need additional measurement equipment such as a measurment microphone and software to measure the room and compute the correction filters required.

Are you using your room correction product with an analog to digital conversion?

[Kenny B]

Nyal, I feed the digital output from a transport to the TacT preamp, which then sends digital output to a TacT S2150 "true" digital amp, The only conversion is the single D/A which occurs at the very end of the amplification process in the amp (technically, more of a "power DAC", I suppose). I've found that additional A/D steps don't do happy things to the AQ. YMMV, of course.

Ken

[Nyal at Acoustic Frontiers]

Ah, a true digital system. Your system topology, Ken, is the future! We just have to wait for everyone else to catch up! Did you know that there are actually 'digital loudspeaker' technologies - see here http://en.wikipedia.org/wiki/Digital_speakers