The Congard Code defines a way to layout a Home Theatre sound system that delivers predictable results consistent with what the movie or video producer intended for the audience to experience.
Its focus is on:
The entire design process is grounded upon these facts:
The original soundscape of a movie or video soundtrack can be reproduced with sufficient accuracy, such that the viewers believe they are part of the action taking place on the screen. At Audio Excellence we refer to this as « the Suspension of Disbelief. » Unlike natural soundscapes (not amplified), the original is artificial, recorded, and comes from an ensemble of loudspeakers. Natural soundscapes have sound source dimensions, directivities, distances that cannot be faithfully reproduced by an array of loudspeakers, but an artificial one can…
The human ear/brain detects the location of the sound source primarily by the “precedence effect”: If two sound sources send the same message with a little time difference, the identified one is the one emitting the first wavefront that reaches the ear.
The human ear/brain considers the first wavefront of the same message reaching it as the important information. Wavefronts arriving later are considered as reflections, and provide only environmental information. This means that we, humans, discern a difference between an original sound vs its delayed reflections.
In view of these facts, the design of a Home Theatre that aims to precisely reproduce the movie soundtrack needs to be inspired by the production studio sound system that mixed the recording.
Most production studios are purposely about half the size of a commercial theatre auditorium. This is to reproduce as closely as possible the conditions where the soundtrack will be played back but without the need to accommodate an audience. The acoustic design of the studio is such that no acoustic signature is created by the rear wall of the room.
In very large residential rooms, it is sometimes feasible to reproduce the layout of production facilities or commercial theatres, but this is extremely rare. In most cases, residential rooms are too small to accommodate such systems. The challenge is to adapt sound systems to residential interiors and faithfully reproduce the listening conditions of the production studio or the very best public or commercial theatres.
The two main issues are:
The Congard Code is based on the integrity of the first wavefront, which should be as close as possible to the one generated by the monitoring loudspeakers used during post production. The first wavefront is the only one perceived as essential information by the listener, as we established above. This crucial integrity is preserved by:
The Congard Code is a process through which all these criteria are met.
Here is its description:
A. L,C,R The most important loudspeakers are the Left, Centre, Right -- as these are « story telling » in a movie or video soundtrack. Their position is of utmost importance as they directly relate to the image on the screen. There is a clearly defined position for these: behind the screen. All commercial theatres and production facilities have adopted this positioning, which is achievable only with an Acoustically Transparent (A.T.) projection screen.
The three L,C,R loudspeakers should have their HF transducers in horizontal alignment, near to the horizontal median of the screen.
B. Bass Extension Unlike in a commercial theatre, the L,C,R, speakers are somehow too bulky to fit nicely behind a screen if they are to reproduce the frequency spectrum down to 40 Hz.
It is therefore ideal to use a separate subwoofer to reproduce the frequencies between 40 Hz and 100 Hz. This subwoofer (I’ll call it “Bass Extension”, as it does not need to go down as low in frequency as an LFE subwoofer) can be fed with a summed L+C+R signal filtered with a low-pass function at 100 Hz. This Bass Extension should be located in or very near to the vertical plane of the L,C,R main loudspeakers. The wavelength at 100 Hz is 343 cm (134 ‘’), so the Bass Extension should not be more than this distance away from the centre speaker. This allows the use of small-yet powerful-main L,C,R speakers and we’ll see that it also offers other advantages. In some installations, it can be convenient to array a plurality (i.e. several) of Bass-Extension speakers in a line below the screen. This can prevent standing waves in the width direction.
C. LFE The LFE is to be reproduced by a dedicated subwoofer, that delivers a low-frequency response down to 20 Hz, or at least 25 Hz. A very important aspect is that the its signal should never be mixed with the L,C,R low frequencies. The LFE signal –at the production stage- is made and mixed down separately from the main channel’s signals, and hence is not phase correlated with these. The result of mixing the LFE and the L,C,R signals would be an addition of random-phase signals with inevitable cancellations. If the walls are solid (I mean, not drywall partitions), placing the LFE subwoofer in a corner can bring an extra level up to +9 dB. Even if not needed, it is at least extra headroom that is most welcome. The LFE should be placed in one of the corners that are nearest to the screen, to avoid difficult time-alignment issues. An even better solution is to use two subwoofers in the two corners that are nearest to the screen.
It is often pointed out that a flat frequency response does not necessarily guarantee good sound reproduction. Also, some speakers with a not-so-smooth frequency response sometimes sound better than others showing a quite linear response. Both statements are correct. However, any loudspeaker that has been equalised properly (performed with parametric equalisation and monitored with MLSSA measurement) to provide a smoother response always sounds better than before equalisation. What is a further advantage is that such an equalisation performed in the frequency domain also improves the time domain response. This is due to the interdependency of the time domain and frequency domain. Such equalisation should be made by DSP with at least 24 bit/96 kHz converters, or higher.
The SPL requirements are quite high in Home Theatre applications. THX recommends 105 dB available at the seating position. Bearing in mind that the sound level decreases by 6 dB per doubling of the distance, at a seating position 4m away from the front speakers, the level is 12 dB lower than at 1m. So, the 105 dB figure requires an available SPL of 117 dB at 1m. Then we need headroom. When a loudspeaker is driven close to its maximum level, its sound deteriorates severely. This also applies to power amplifiers: clipping is one of the most unpleasant distortions and it is also likely to damage loudspeakers. So, a part of the Congard Code is the specification of the loudspeakers and power amplifiers that need to be potent enough to provide the desired SPL plus 6 dB headroom over the whole audience area. It is to be noted that the majority of home theatre speakers are adaptations of Hi-Fi speakers therefore not capable of providing the required SPL. When real power is required, the use of professional components, high in sensitivity, more power, bi-amplified designs, etc. is required.
The main causes of distortion in a system come from the loudspeakers and are strongly dependent on a) the cone displacement in the low frequencies b) the HF transducer being driven close to its maximum level
The cone displacements are proportional to the square of the inverse of frequency (1/f)2 Setting a cutoff frequency at 100 Hz – apart from the advantages described above-seriously limits the cone excursions and reduces LF distortion. The HF transducer simply needs to be able to provide much more SPL than what is actually used, in order to minimise its distortion. This means that it should be a horn & compression driver assembly. With sensitivities typically comprised between 105 and 110 dB/1W/1m, horn and compression drivers are capable of quite loud SPL figures, hence providing the required headroom and being always operated at levels where the distortion is minimum.
Since the beginning of talking movies (1929), it has been obvious that the sound of speech should correspond in time with the movement of the lips. The tool to achieve this alignment is the iconic clapperboard, still sometimes in use today (although in digital recording it is replaced by a time code). In a Home Theatre playback system, the separate processing of the audio and of the video creates a delay of the video versus the audio. This, of course, depends on the display and on the AV processor. Typically, such delay is comprised between 100 and 150 ms. Adjusting this can be done on the AV processor, but it is very difficult to use an actual movie to make the test and adjustment. The best solution is to use a specific video file comprising synchronised clicks and flashes. This file is provided by Screen Excellence to whoever registers on its website.
There are lots of new room EQ solutions appearing these days, and AV processor manufacturers seem to make this domain their battlefield. That might be good for them, but it deserves some thinking.
What room EQ does is:
a) Measures the sum of the original signal and some early reflections b) Combines the inverse response (sometimes in amplitude only, sometimes with a « smart » compromise between amplitude and phase) to the signal fed to the amplifier.
The problem lies in the early reflections, so let’s have a look at what a reflection is. When a wavefront (say, the original one) reaches a reflecting surface, this surface absorbs a (small) part of the sound, another (small) part goes through it (it is transmitted), but the majority of the sound energy is reflected. This means that a second wavefront carrying the same message as the original one is created. Because its path to the listener is not direct, it is slightly longer than the initial one, and also slightly attenuated (the sound pressure diminishes with distance). When this secondary wavefront merges with the first one, the difference in time arrival creates some phase cancellations at some frequencies. The theoretical response of this addition (provided the two wavefronts have the same SPL, which is not exactly the case) is called « comb filtering » because its appearance reminds of a comb:
In an actual room with various reflections and as many various amplitudes and delays vs. the original wavefront, the frequency response is a random combination of multiple comb-filterings. This means it can be anything!
However, what you hear first is the original wavefront and it is what really matters to your brain. It is not subjected to comb filtering
What a room EQ system does, is to feed the inverse response of the measured frequency response – including all reflections- into the first wavefront.
It is then easy to understand that it will severely degrade this original signal, inducing severe colourations and time discrepancies. Furthermore, the correction occurs before the phenomenon it is supposed to compensate. As nice as it might look on a measured amplitude frequency response curve, it just cannot sound right: The original sound, the one that we identify as the primary and most important source of information, has a modified response to compensate for reflections that will occur later. There is no way this can work! In the Congard Code, the only EQ that is applied is to optimise the very first wavefront generated by the speakers. Any type of room equalisation (correction, calibration, software) is not suggested as this colours the sound and is based solely around one seat location in the audience. Instead, the Congard Code approach concentrates on ensuring the initial wavefront reaches every audience member in the purest form possible, maintaining that room modes are solved using other means, such as acoustic treatments.
If the listening experience in the room is not as expected, it is often due to poor acoustics. As we have seen previously, no room EQ can solve this electronically. Only a proper acoustic design can provide good results. Although the ear/brain system decodes the arrival of successive wavefronts, identifying the first one as the essential information, the early following ones (with less than 20 ms delay vs the original one) as early discrete reflections, and the later ones as reverberation, it is always better to have good room acoustics. The early reflections generate some « comb filtering » that tends to blur the signal intelligibility. The reverberation is lower in level (the sound level decrease when it travels. After 20 ms, it has already travelled about 7 m and has lost about 16 dB) and there are so many reflections that they generally do not affect the tonal balance or the intelligibility.
In a Home Theatre, there are two acoustic domains on which to focus:
a) Minimising the early reflections. This can be achieved by combining absorption and geometry (following the Reflection Free Zone concept, developed for recording studios) b) reducing the low-frequency standing waves (also called « modes ») below 100 Hz which tend to occur in small rooms like residential ones. This can be done by using bass traps.
The Congard Code is a method and a guideline to implement a Home Theatre sound system in a way that is simple and provides excellent results. It focuses on optimising the initial wavefront that is generated by the loudspeakers. This is mainly by defining the allotment (routing) of the signals feeding the loudspeakers, DSP parametric equalisation if needed, and distortion reduction. Also it specifies the equipment needed to reach the necessary SPL level, adding headroom to it. Moreover, it is well suited to most room proportions, and does not require more parameters than the distance between the screen speakers and the first row of the audience.
Published by Patrice Congard, Audio Excellence CEO, 10th November 2017