Equalizers & M/S

Unless someone is an absolute 'audio newbie', principle and use of stereo equalizers is definately nothing to talk about. However, using equalizers in the M/S format instead of the L/R stereo format is an entirely different story.

While the standard L/R stereo format simply uses one channel for the left side and another one for the right side, M/S stereo uses the M channel for the mono signal and the S channel for the spatial information of the stereo signal. L/R can be converted to M/S and vice versa, using a matrix.

With a common music mix instruments like kick drum, bass, and the vocals are located at or close to the center of the mix in almost all cases. These sources also produce a very high percentage of the total level; exceptions confirm the rule, of course. Checking the M and S levels of mixes you'll find that the level of the M channel is usually between 10 and 20 dB higher than the level of the S channel. The more 'centered' the mix is, the lower is the level of the S channel in relation to the M channel. The other way around, the wider the mix, the lower is the level difference between S and M.

When using an equalizer with a boost at a certain frequency band only in the M channel the setting results in a 'real' eq that affects only the center of the mix. In addition, the stereo base width in the particular frequency band is reduced. Using a cut, the 'real' eq effect is still limited to the center of the stereo panorama, but with enhanced base width.

When using the same equalizer setting, again with a boost, only in the S channel, the 'real' equalizer effect appears only in the room while the center is not affected; however, you get a base enhancement in the affected frequency band. The base enhancement is caused by the boost curve of the eq in combination with the phase shift in the rise and fall of a bell type eq curve. Since the M/S format is a lot more sensitive to phase differences between M and S than the L/R format, the phase response of the eq is an important factor of effects that seem to increase the depth of the room behind the speakers. Equalizers with improper phase response for this purpose offer poor results.

Since the level of the S channel is - as explained above - usually a lot lower than the level of the M channel, the audible equalizer effect is low but the effect on the base width is high. In the S channel boosting results in base enhancement while cutting results in a reduced base width.

So, the easiest way is just to use an eq in the S channel, play around till you get some nice effects on the spatial image, and live with the audible eq effect that comes along as colleteral damage. This is the way the TM134 M/S EQ basically works.

With delicate mastering processing or such a treatment in the mix master or a subgroup, you usually would like to benefit from the base width enhancement; however, not so much with the audible eq effect that comes along. When using a standard stereo eq to compensate the eq, you are affecting the entire mix identically, which doesn't work for the compensation. The only way is to use exactly the frequency and q-factor setting of the eq in the S channel and adjust a cut in the M channel (or a boost if you have used a cut in S channel - let's stay with a boost in S and a cut in M for the moment). To compensate just the eq and go back to the original frequency response of the mix, the cut in M must be smaller than the boost in S. The exact relation of boost and cut depends on the level difference between M and S; therefore some kind of automatic compensation is not possible. It needs to be done by hand and by ear. Some issues make this compensation more difficult than it sounds. Setting up precisly matching center frequencies and q-factors with two fully parametric mono equalizers with sufficient ranges is not as easy as it sounds. Tolerances between mono eq's are usually pretty high; setting the pots to the same scale values will not result in the same setting in both channels but in settings that are quite similar but not identical. In addition, with most fully parametric eq's the q-factor is not constant over the boost and cut range. Checking the frequency response with different boost and cut settings shows that a higher boost and cut usually is steeper than a lower value. Even if you succeed in adjusting the same center frequency and q-factor, the necessarily different values for boost and cut result in different curves and therefore not in the required compensation. That's were the TM133 M/S mastering equalizer comes in.

If you don't care about the more or less coincidental result and the changed frequency response, you can stop reading here. However, if you want to get to the point that you can modify the base width frequency dependent without altering the frequency response of the mix, the TM133 offers a way to do this without the problems mentioned above. The four fully parametric bands of this M/S equalizer are precisely matched for the M and the S chain. The separate boost/cut pots per band for the M channel and the S channel allow fast and precise base width regulation and compensation. The q-factor regulation is designed to allow compensation with different boost and cut values. Since the TM133 has an integrated L/R to M/S matrix at the input and an M/S to L/R matrix at the output, it comes as a full package that can be used without additional matrix units with standard L/R format stereo signals.