Hi Sleiven
My observations:
1. We should take care with the "fine print", in other words "the devil is in the detail". As i mentioned in the past, it is all dependent on the recorded velocity. This graph will depend on the recorded velocity and the groove radius. As you see there, the graph is for 10 cm/s. This is a rather high velocity for treble frequencies. And, to keep things in focus, 10cm/s at 20KHz would require 1281g of acceleration.
2. 10% of complete 2nd harmonic distortion is not believed to be audible at all, in fact 10% 2nd-hd is the typical "limit" figure used to calculate the power of a valve amp for this reason.
Let's consier 10% 2nd-HD as our "limit", this means no more than 5KHz @ 10cm/Sec if it's going to be traced by a 18um conical (Also, please note that the conicals often discussed here, like most Audio-Technica conicals, are 15um, not 18um -- but let's assume 18um for now)
It is good to take a look at Jim Lesurf (hifi-news) pages, they have examples of the real recorded velocities on real records. I'm going to copy paste here:
" Bach’s Trio Sonata in C, BWV1037 as performed by Jaap Schroder, Stanley Richie and others on a half-speed mastered LP produced by Reference Recordings around ten years ago."
As you can see, 10cm/sec at 5KHz or more is never reached, there is no such combination (= acceleration) on the record.
A 18um conical would trace this with as low a distortion as any other stylus tip shape.
Let's see a more "challenging" record, "Mobile Fidelity LP of Pink Floyd’s Dark Side of the Moon"
10cm/sec (or more) at 5KHz (or more) is achieved, but less than 0.5% of the playing time of the whole record is at these conditions.
The higher velocities are often at the beginning of the record, for obvious reasons. The graph you cited was calculated for the middle of the record.
Thus, either our "limit " is never reached, or it is reached but only at 0.5% of the playing time.
All in all, this shouldn't be any news for cutting engineers -- it is well known that records are almost always mastered to be traced(read) by a conical stylus.
Of course, part of me wishes they would cut records at much higher velocities, to be traced by finer shapes, to better take advantage of the medium. But honestly most (95%?) of the records out there will be traced cleanly by a conical.
Even a recent pair of records I bought from a very interesting musician ("Slugbug"), analog recorded, mastered by him, cut by Paul Gold in NYC straight from the analog tapes... those records are rather loud or a bit louder than average on my turntables. At the beginning I played it with one of the Microlines (AT-VM540ML) and it sounded so good, but the surprise was that it also played cleanly with my 0.6mil conical cartridges here (i.e. VM-510CB, AT91R), right to the inner grooves. I was surprised, but I shouldn't be that surprised.