A Modest proposal for a High Resolution Monitor (Studio & Domestic)

Folks,

The earlier ATC debate over in HiFi got me thinking as to what I would like to arrange right now at my current stand of understanding of the matter a "true monitor" system.

I would like to first outline what I would want from such a system.

I recommend the study of the following document to get some idea where outline requirements are taken from:

SURROUND SOUND FORUM - RECOMMENDED PRACTICE SSF – 01.1- E-2002 - Listening Conditions and Reproduction Arrangements

I would consider the Monitor to be best considered "mid/farfield", that is around 3m listening distance. I would also consider the requirements for a heavily acoustically treated room counterindicated, as this is an unrealistic requirement for domestic settings.

Based on the recommendations for 2 Channel reproduction as offering per chennal 82db for a -18dbfs/rms setting we can translate this into 100db SPL per channel for notional full scale, with peaks of 103db. This referenced to a listening distance of 3m in the free field suggests an attenuation of around 9.5db so the Monitor must produce 113db/1m SPL Peak with low distortion and compression.

In order to reduce the level of early reflections (< 15mS) to at least 10db below the direct sound we require either a near sound dead room (usually canned acoustic non-envoironment room or anechonic chamber) or a highly directive system. In my view the use of a highly directive system is preferable, if set up correctly it will provide a sufficiently wide window in which the portrayal of the stereo image is accurate and where tonality is accurate. This window will be as wide for the accurate portrayal of the stereo image as with wide dispersion monitors, the window of tonal accuracy will be somewhat narrower.

Based on my own experience I would include several more requirements outside the traditional areas. I would first postulate that the system must be minimum phase and impulse/time coherent. I would secondly postulate that in order to judge tonal colours and tonal accuracy the entire formant range (frmants are the overtones of an instrument that so to speak make the "sound" of the instrument) of instruments must be conveyed from a single source of radiation using the same material throughout, this for example means that even coaxial systems with a metal dome compression driver and a paper/plastic cone are counterindicated if the crossover falls into the formant range.

The formant range is as low as 70Hz for the Upright bass, however more commonly it is reconed to start at around 150Hz for the upright bass and reaches past 3KHz for the human voice and certain woodwinds.

On further Items I would note that a reasonably flat and wide frequency response is implicit at any extent.

What would I propose to make such a system?

I would use a large format spherical wave horn with a suitable cone driver for the larger part of the spectrum. A prime example would be to combine the BD-Design Oris 150Hz horn with a Supravox 215-2000 EXC 8" Fieldcoil cone driver. The combination should exceed 103db/1W/1m efficiency and can produce easily > 116db/1m with low distortion and compression, requiring no more than 20W RMS amplification to achieve that.

Here an Oris Horn with a Klipsch La-Scala type bass bin.



BTW, I use a dipole version of ths driver at home on a dipole as a compromise in achieving good directivity control and domestically acceptable size/shape, this manages 97db/1W/1m efficiency and around 106db/1m using my current amplifiers for < 3% 2nd Harmonic distortion amplifier and speaker together. Not quite enough for studio monitoring but still pretty decent.

Anyway, the proposed Horn/Cone driver assembly can cover from around 150Hz to around 8KHz with a pretty well flat response as well as offering a well directivity controlled point source. I would utilise the Driver itself open back directly on the horn without further enclosure or attempts to remove the rear output, which would make for a nice hypercardiode radiation pattern. When combined with a simple 1st order highpass at 150Hz (line level or directly in series with the driver) we get a resonable phaseshift at low frequencies and a 3rd order highpass behaviour acoustically allowing us to operate the main assembly directly without intervening DSP correction for frequency or phase.

The problems start above 8KHz and below 150Hz. The High frequencies of the Horn/Driver assembly will roll off with a near 2nd order slope with 8KHz as -6db point acoustically, so we require a suitable high frequency device. Based on experinece I would choose a Focal device, the "Car HiFi" TN-52 is an excellent device, which due to the small diameter of only 2" can be suspended coaxially within the horn relatively close to the main driver. With a suitably formed rear cover it may indeed be used as diffuser to slightly improve the high frequency response of the Horn/Driver combo. Due to the relatively large diaphragm it also produces a nicely controlled directivity behaviour without adding a further horn above our 8KHz crossover point.

In order to achieve a time coherent system we need to delay the signal to the tweeter somewhat, we also require an amplifier as we have a 7db difference in efficiency between the main acoustic system and the tweeter. The amplifier should provide around 100W unclipped signal. Modern digital crossovers offer all this functionality and even allow us to shape the driver output of our tweeter such to equalise out some of the remaining frequency response uneveness of the main horn/driver system.

We now have a System that will cover 150Hz to >> 20KHz as point source, directivity controlled acoustic unit, requiring a main amplifier offering around 20 Watt or more of really high quality output power and a good "tweeter amplifier" with around 100W, implying that an absence of "grainieness" in the tweeter amplifier is most crucial.

The final missing part is the woofer section below 150Hz. As we have payed much attention to directivity control till now it would be stupid to stop. In terms of room interactions the most desirable solution is a hypercardiod LF system. This can be achieved in several ways, one of the easiest is to combine a dipole and a monopole radiator. If the output in terms of SPL is identical for both systems we will find that on axis the LF output is boosted 6db over using dipole or monopole alone, while behind the assembly the sound is in effect eliminated crealing a strong "null" in the output.

These days a number of 12" long throw drivers exist that are well suited to dipole operation and equally enough such exist to work well in small sealed enclosures. If we use a linkwitz style folded dual driver dipole with (say) 2pcs Peerless XLS 12" Drivers in a 15" cubed format and stack tree of these interleaved with 3pcs of 12" XLS in sealed enclosures we can then use our digital crossover to equalise both sealed and dipole woofers and adjust the levels.

Here a picture of one of these folded up dipoles:



To drive the LF Array we would need a pair of 2 Ohm tolerant high output (> 400VA into 8 Ohm) 2-Channel Amplifiers to drive this LF Array, the payback for this extreme application of power and drivers would be a relatively compact LF system (2.3m Tall tower, around 38cm wide) with very low distortion and compression and conrolled directivity (the alternative would be a 3m X 3m mouth horn!) to match our 150Hz and up system.

Such a system is of course miles off the beaten track for studio applications and requireds a free-standing position, but it has at least the potential for accuracy and resolution well past many of the best conventional systems. It is not designed for the maxmum output levels required in acoustic non-envoironment ooms however it also does not require such so, the field is again even.

Anyway, some of my thoughts, which may illutrate where I am coming from.

Ciao T

 

You can buy the Peerless XLS 10" for only £89 each from BK. I have them in my 3 ways. The Q is so low, that even after you have put a series inductor [xover] and allowed for the loop resistance of the speaker wire - it's still lower Q than most drivers alone. So an air core inductor is at no disadvantage versus an active speaker with another typical driver, except for a bit of sensitivity loss. Remarkable drivers..

 

Surely an easier way is to fire the woofer into the wall and employ one of those clever Tact thingies ;-)

 

Hi,

There are things TACT can do and things it cannot. For example, TACT cannot remove a room mode or reduce it's excitation and thus cannot alter the timedomain problems, it can equalise the frequency domain problems, which is better than nothing, but ideally an acoustic problem is dealt with acoustically.

Ciao T

 

Mmm, subjectively it can do a very good job IMO, provided speakers are setup correctly, although I would agree it cannot correct for the phase issues create by the monitor's backwave. Soffit mounting makes sense then?

What are your thoughts on Martin Logans? In particular Force Forward is designed to create a null behind the speaker, they offer controlled directivity, and the only crossover below 20khz is centred around 250hz.

 

3D

are you aware of the recent article in hobby hi-fi comparing several dipole sub construction principles? The Audioelevation thing did not ome out tops, they preferred a simple U or H frame, less distortion and not much worse directivity IIRC.

 

Hi,

No, I don't get to read HH or K+T here.

This may very well be down to the drivers used. I used the Pictures from Audio Elevation merely as illustratiuon of the principle. A friend build a dipole following the same principle (which BTW is derived from Siegfried Linkwitz's work, whatever little "design" work was requierd by yours truely) using a pair of 18" Pro Sound Woofers each, simply awesome.

Using overall 6 pcs Peerless 12" XLS drivers in a similar (but smaller) design should give almost the same cone surface as 2pcs 18" but more linear throw. And do not forget that I want to add another three sealed box 12" XLS drivers (both for directivity control and to increase SPL at very low frequencies) which together should make for an LF system with few limits.

Ciao T