Klaus is correct. Changes in zenith error impact other sonically important parameters when navigating warps - just as raising/lowering a tonearm changes overhang and so on. The tonearm/cartridge/record groove relationship is a multivariate environment. What we need to sort out for ourselves is, what changes are worth worrying about? The answer often depends upon your stylus profile, but sometimes statistical analysis will help too.
There is nothing wrong with Baerwald, Loefgren B, Stevenson or UNI-DIN. They are each simply efforts to determine where you want to apportion your tracing distortions. I think it is more important to make a decision on your innermost and outermost groove radii than to choose a single alignment scheme. I've done a 1,000 plus record measurement analysis (see "
Tracing Distortion, Angular Error & Innermost Grooves"
HERE) and my data shows that exceedingly few of my records go less than the IEC standard of 60.325mm radius.
% Incidence of Innermost Groove less than 60mm - 1950 and later pressings
Classical 5.2%
Rock 7.8%
Jazz 5.0%
I decided that I wasn't interested in increasing tracing distortion for ALL of my records to benefit the last couple minutes of play time on a tiny percentage of my records. This decision was solidified when I heard a comparison that Fremer did for me of the SAT tonearm and a linear tracking tonearm on a track that went to 56mm radius with an orchestral climax. The SAT tonearm uses an innermost groove assumption of 75mm, which is FAR away from the statistically significant figure I found in my record set for records made after 1990 of 68mm. As Klaus correctly points out, the tracing distortion at the inner portion of the record gets worse FASTER as it moves inward than it does anywhere else on the record. The SAT soundstaging collapsed the closer it got to the end of the record. High frequencies weren't nearly as extended and delicious as they were at the center of the playing area and things sounded more "SET"-like (I'm sure due to the higher second order harmonics resulting from the phase error on signals coming from the center of the two speakers). Imaging suffered a bit too, but it was still certainly a high-end experience. The net result was, which arm you liked better depended upon where you were playing it on the record surface!
I use an industrial microscope to view the contact edges of the stylus and inspect their angular relationship to the cantilever. I regularly see more than 2 degrees zenith error - even on the most expensive cartridges. I have the prototyped WallyZenith that I can then use to compensate for this zenith error. I have blind tested others and they use the same description I have come up with for the difference between corrected and uncorrected zenith: improvement in soundstage depth and width, increased imaging locational specificity and a generally "cleaner" upper midrange/high frequency experience.
With that said, I look forward to reading the AES paper that Klaus was involved in that determined error needed to get up to 12 degrees before it was audible. I certainly would agree that it would need to be that high before IM distortion (sounds like "breakup") was audible, but one can calculate that a "mere" 2 degrees of zenith error causes phase distortion on a 10kHz signal on a 33rpm record at the inner null point (Loefgren B) to be just under 3%. This causes harmonic and IM distortion of its own, but the IM distortion would be absolutely negligible. Harmonic distortion would not be. So, what was Klaus and company listening for? Changes in soundstaging, imaging and high frequency response? Did they use a severe stylus profile that had a minor radius less than 6 microns and a major radius over 40 microns? (I'm not suggestion there is anything magical about these figures. I'm just trying to illustrate the need for a severe profile to be able to hear the greatest difference.) Further, the math suggests that if dynamic SRA isn't spot-on FIRST, your efforts at discerning zenith error will be much tougher. Was this accounted for first? I look forward to finding out!
We have calculated that a 5 degree change in cantilever angle OFF of tangency at the null point causes a decrease of motor output in the neighborhood of 0.4%. It also causes some other IM distortions to increase as well. However, it is our opinion that the negative impacts at less than about 5 degrees are hardly worth worrying about. The cure is most certainly better than the disease inside of an acceptable tolerance. I aim to share the data behind this and the process for measuring zenith error that does not require an ultra-expensive microscope this year. I just need the microscope to validate the approach!
