Francinstien is a filter which improves the quality of reproduced stereo from loudspeakers. Perfect Pitch Music's interest in improving stereo derives from my work in Virtual Reality. People don't appreciate that a good home stereo system is the only practical VR system that most of us possess! The aims of a hi-fi system are more limited than true VR because it only aspires to address the auditory sense but the goal, to transport the person to a virtual "other" place, is the same.
What does it do?
The brain, when it experiences a stereo recording, receives signals from the ears which are very similar to those it expects to hear when it listens to real instruments in real space. But only very similar because conventional two loudspeaker stereo ('drawing-room' stereo) cannot create a perfect illusion. The passive Francinstein filter was developed as a result of psychoacoustic research - involving controlled listening tests - designed to improve the quality available from any stereo system. It acheives this manipulating the stereo image to be nearer to that experienced when listening to actual instruments in authentic space.
Why has no-one done it before?
The drawbacks of conventional stereo have been understood since the time of Blumlein and the earliest days of stereophonic research. Many researchers have endeavoured to find a solution. But, whilst it has proved easy enough to bring about changes in stereo image quality, this has never been achieved, prior to the Francinstien filter, without incurring penalties of colouration, distortion and deterioration of transient response.
Can you explain simply what Francinstein does?
Yes, it's not complicated. But first you need to understand a little about how humans hear and interpret the direction of a perceived sound. You see, Francinstien is about understanding psychology....... it's not about understanding electronics!
OK, how do we hear sounds from different directions?
Psychologists know that the psychophysical cues we use to determine the direction of a sound depend on the nature of that sound: At high frequencies, the head casts an effective acoustic "shadow" which acts like a low-pass filter and attenuates high frequencies arriving at the far ear. The result being that the relative loudness of a sound at the two ears is different. Psychologists say, in this case that the nervous system makes use of interaural intensity differences to determine direction. At low frequencies, sound diffracts and bends around the head to reach the far ear virtually unimpeded but there is a delay between the sound reaching the near ear and the further ear. The nervous system compares this relative delay of the signals at each ear (known technically as interaural delay difference) and uses this to determine direction. In the case of steady-state sounds or pure-tones, the low-frequency delay manifests itself as a phase difference between the signals arriving at either ear. The idea that sound localisation is based upon interaural time differences at low frequencies and interaural intensity differences at high frequencies has been called Duplex theory and it originates with Lord Rayleigh at the turn of the century.
When rock and pop records are recorded, the instruments are recorded separately and are mixed and "panned" (steered into stereo position) electronically within the recording console onto a two-track master. Yet, the pan-pot is a simple device which controls the relative contribution of an identically phased signal between the left and right stereo channels in a complementary fashion; just like a "balance" control does on an amplifier. It doesn't introduce any delay type manipulation. Similarly, many classical recording engineers use simple crossed-pair microphone arrangements and an essential feature of this "coincident" microphone technique is that all the time difference information between the microphone capsules is erradicated by positioning the capsules of the left and right hand microphones as close together as possible.
So, how does a stereo signal, which only contains channel level differences, fool the brain into believing its hearing instruments from different positions between the loudspeakers?
Well, it should be immediately obvious that left/right channel intensity differences result in high frequency interaural intensity differences when listening to a stereo loudspeaker system. The conceptual problem is for bass frequencies. In fact, it is far from obvious that these same level differences do, in fact, translate into low-frequency interaural phase differences as well! Indeed the common misconception that bass frequencies contribute little to stereo imaging is largely due to misunderstanding this very important principle.
When two spaced loudspeakers produce different intensity, yet identically phased, sounds, the sound-waves from both loudspeakers travel the different distances to both ears and therefore the signals arrive at each ear at different times. This diagram
illustrates the principle involved: The louder signal travels the shorter distance to the right ear and the longer distance to the left ear. But the quieter signal travels the shorter distance to the left ear and the longer distance to the right ear. The result is that the sounds add vectorially to the same intensity but different phase at each ear. The brain is able to interpret this information in terms of interaural delay. Two channel stereo is more subtle in its operation than is generally supposed. A purely intensity controlled stereo signal actually produces two stereo images at the same time: A predominantly high-frequency interaural intensity-derived image and a predominantly low-frequency delay-derived image.
How very fortunate that these two mechanisms produce a perceived sound at exactly the same place in the stereo picture.
Ah, that's just the problem, they don't! These simultaneous, and coexistent, stereo images are not in exact perceptual register. So that, when two-loudspeakers reproduce a stereo image from a standard stereo CD for example, the high frequency components of each instrument or voice emanate from one place within the stereo image and the low-frequency components emanate from nearly - but not quite - the same place. The result, to use a visual analogy, is a slight smearing or blurring of the stereo image.
And that's the problem that the Francinstien filter solves?
That's right. The important qualitative fact to appreciate is that, for a given interchannel intensity difference, the direction of the perceived auditory event is further from a central point between the loudspeakers when a high-frequency signal is reproduced than when a low frequency is reproduced. Since music is itself a wideband signal, when two-loudspeakers reproduce a stereo image from an interchannel intensity derived stereo music signal, the high frequency components of each instrument or voice will subtend a greater angle at the listening position than will the low-frequency components.
Blumlein mentioned in his 1931 patent application that it was possible to control the width of a stereo image by matrixing the left and right signal channels into a sum and difference signal pair and controlling the gain of the difference channel prior to re-matrixing back to the normal left and right signals. He further suggested that, should it be necessary to alter the stereo image width in a frequency dependent fashion, all that was needed was a filter with the appropriate characteristics to be inserted in this difference channel. He termed this filter the "Shuffler". After his untimely death, the post-war team working at EMI on a practical stereo system and attempting to cure this frequency dependent "smearing" of the stereo picture implemented just such an arrangement. Unfortunately this circuit was found to introduce distortion and tonal colouring and was eventually abandoned. Other derivatives of the "Shuffler", using operational amplifier techniques, have appeared, but the act of matrixing, filtering and re-matrixing is fraught with problems since it is necessary to introduce compensating delays in the sum channel which very exactly match the frequency dependant delay caused by the filters in the difference channel if comb filter coloration effects are to be avoided. Furthermore the very precise choice of the correct constants is crucial. After all, existing two-loudspeaker stereo is generally regarded as being a tolerably good system, the required signal manipulation is slight and failure to use the correct constants, far from improving stereo image sharpness, can actually make the frequency-dependant blurring worse!
Others have taken a more imaginative and unusual approach. In the late 1980's, Dr. Edeko conceived of a way of solving the problem acoustically (and therefore of side-stepping the problems which beset electronic solutions.) He suggested a specially designed loudspeaker arrangement where the angle between the high frequency loudspeaker drive-units subtended a smaller angle at the listening position than the midrange drive-units and these, in turn, subtended a smaller angle than the low frequency units. This device, coupled with precise designs of electrical crossover network enabled the image width to be manipulated with respect to frequency. I can vouch for Edeko's methods because I built a prototype system and was surprised by the improvement in stereo image quality it brought about. Indeed, it was Edeko's articles which first alerted me to the fact that 'drawing room' stereo could be improved. But I didn't get any further until I was involved in my first recording which appeared on vinyl and CD. That gave me an opportunity to compare analogue and digital reproduction because I also had a copy of the original master tape. Unhesitatingly, I should say, the CD was 'nearer' the master, but vinyl produced a 'better' stereo image - in fact, better than the master tape! A conundrum indeed.
After much deliberation and armchair theorizing I set about doing some experiments. Late nights with an oscilloscope eventually uncovered that electrical and mechanical crosstalk within the cartridge and pre-amp were causing a stereo image manipulation which was similar to that brought about by the Blumlein 'Shuffler' circuit and Edeko's loudspeakers - all the important narrowing of the stereo image at high frequencies. It supported what I and so many hi-fi fans knew to be the case, that vinyl really does sound better than CD - especially in LP's presentation of a realistic soundstage.
The excitement for me was discovering that high frequency crosstalk emulated the effect of the 'Shuffler' circuit and improved the stereo image without incurring tonal penalties from sum and difference processing. It might be sixty years late, but to my knowledge, Francinstien still represents the first commercially (and sonically) successful implementation of the last piece in Blumlein's stereo jigsaw.