After all those tests and measurements and with much hope in resolving the problem, I quickly acquired an IR block/stop/cut filter to see IF the problem could be eliminated:-
The filter is essential a glass filter with heat reflective coating on it. As such, it is also known as hot mirror, besides calling it a IR blocking filter. (More product brochures can be found here for the B+W and here for the Heliopan.)
A typical wavelength response on the light spectrum is as follows:-
It can be observed that such a filter does reflect red or green light when viewed from a large angle of incidence. In contrast, when it is viewed near perpendicularly, it is clear. Hence, these filters should not be used for wider lenses for an Angle of View greater than 60 degrees, i.e., at about 35mm 135 film equivalent. Theoretically, it has virtually no light loss for visible light and thus the EV drop should be zero and it is negligible.
Now, that's the most important part which we are looking for - the Results! Of course all tests were carried out under the same pure tungsten environment (at the same time). This time I only tested at 35mm of the focal, which was found to have a more dominant effect for the problem than the tele-side:-
| Condition | Record of the Focus Scale for where AF was Measured | Result Image (Central 100% Crop) | Result Image (Large and Uncropped) |
| Without IR Cut Filter, Optical PMAF |  | View | Download |
| With IR Cut Filter, Optical PMAF |  | View | Download |
| With IR Cut Filter, Live View CDAF |  | View | Download |
| Without IR Cut Filter, Focus Adjust at -10 (Max.), PMAF |  | View | Download |
| With IR Cut Filter, Focus Adjust at -10 (Max.), PMAF |  | View | Download |
Well, once again, the best focus was achieved with Live View. With the help of the IR block filter, the front focusing symptom has been relieved somehow, but only partially! :-( Now that the focusing error seems to be halved in this case and indeed it even did slightly better than making Focus Adjust at a full span backwards at -10 as allowed. Most possible focus correction was achieved with the use of IR cut filter together with a Focus Adjust at -10 - but it was still NOT there for exactly where the Live View could reach! :-o
To explain the half-failure (and half-success anyway!) of this IR cutting remedial, there are at least two possibilities:-
1. The IR block filter is not strong enough to reject all the harmful IR lights from tungsten and the SAFOX IX is just too sensitive to those;
2. The harmful IR light or near-IR lights are out of the flat rejection band of the filter such that the filtering is not as effective as desired. As such, some improvement can be seen but yet the problem is not totally eliminated!
No matter how, at the end of the day, Pentax must be blamed for the in-born latent defect and that complete design fault of the SAFOX IX system that they created. In fact, this fault is a very serious one and is unforgivable for the large amount of focusing error that is produced! >:-[
Nevertheless, I am still feeling somehow good to find out and know about the truth and then to report in details here for what I've discovered. It is a tragedy for all we K-r and actually all Pentax users after all as it is seen that Pentax created an inferior AF system like this. I do hope that it should *never* happen again in the Pentaxland, for any new Pentax body to come!
N.B. Last time I used an entry level Canon 550D (for a complete month), it never had this kind of front focusing problem under tungsten and the AF had been dead accurate! >:-(
Related:-
Pentax UK and Pentax Germany Admitted K-r Tungsten Front Focusing Issue (But there is No Solution, as Both Clearly Told)
Still Annoying K-5 and K-r Old Issues, Users are Pissed Off!
SAFOX IX Tungsten/Yellow Light Front Focus Issue Fully Investigated