Test record issues are abundant so that could be a part of it but there is no conclusive answer as to why some results are more "squiggly" than others (especially with respect to the left channel CA TRS-1007 sweep). Perhaps record warp plays a role. I've even seen high pass filters affect the higher frequencies so the individual set-up plays a part in addition to the test record, and that introduces a lot of variables.
However, it's always an amalgam of things, including the manner in which the data is processed. One needs to have knowledge of FFTs and coding in order to discuss this more seriously. There is no one to one ratio when it comes to the recorded sweep and graphed FR. Please see the "How It Works" section on the script github page for more information regarding the issues encountered when creating the measurement code and the strategies used to overcome them.
Also, a quick check of CA TRS-1007 measurements posted by various members shows that this is a recurring phenomenon. (Simply look at page 1 for numerous examples. MANY results show a left channel with "squiglier" results.) The test record is essentially a copy of the JVC TRS-1007 so a lot of this may be due to the manufacturing process and exacerbated by individual playback conditions in addition to the manner in which the script analyzes the sweeps. Simply, the CA version of TRS-1007 is technically inferior to the JVC. (And no test record is perfect to begin with.) There is not anywhere near enough evidence to draw any conclusions with respect to cartridge condition and it is reckless to do so. You are all encouraged to explore this phenomenon but please do so though proper experimentation and analysis.
I ask that we move any conversation regarding the script to the appropriate thread.
Yes, you are right, the reasons for this can be different,
There is a very large tracking error at the beginning of the record.
It would be necessary to take several measurements with different cartridge settings, with different tracking error settings.
It would be necessary to take several measurements with different settings, such as different azimuth, different pressure, different VTA angle, and different combinations of the above settings. Perhaps some sensible conclusions could be drawn from such comparisons.
It is a lot of work.
The measurement script itself is very good, but the specific measurement results of the same cartridge models will always differ, because there are many mechanical variables that are known to affect the measurement result. Not to mention various electrical variables.
The measurement results should therefore be treated as approximate.
If the manufacturer provides the parameter Frequency response 20 Hz - 20 kHz +/- 2.5 dB, plus other mechanical variables, then it is obvious that with such a large tolerance, the measurement result for this given model can and often is very different.
If the measurement result looks like the one in the graph below, and in the case of measurement with the same record but a different cartridge, this jerking effect does not occur or is minimal, then either the cartridge has been set incorrectly or the cartridge is in poor technical condition.
A few more measurements with different settings and it will be clear whether it is the poor condition of the cartridge or the poor setting that is to blame.
I am convinced that it is the poor technical condition of the cartridge that is to blame.
I personally treat this script as a great tool for configuring a given cartridge, not as a measuring tool.