Do you really think your model is superior ?
I give Scott all the credits to come with a piece of SW that is as least as good if not better.
The approach with a Hilbert transform is very capable of finding the instantaneous frequency, when avoiding some well defined traps to stay away from.
I was especially not impressed by you processing of Pano's virtual TT1.
The bulges at both ends were equal in size, although the lower one in your image should be half that size.
You mentioned that your model is bandwidth limited to 40 Hz, although flutter goes well up to 100Hz and also pick up of 50 or 60Hz mains may cause IM products.
I'm sorry to see that the discussion has left the rails in such a dramatic way, where so much knowledge and ideas could have been bundled.
Hans
Attachments
Last edited:
Yes, but only a small specific part of it. That part is amazingly accurate, fast, and has superior noise rejection. The rest of my SW is just cobbled together for my own use which is fine, but inferior.
Every thanks and respect to Scott and the polar plot python code in its entirety. It works well and is thoroughly useful and great. I really hope it's further developed and is widely used.
This situation is fairly insane, because I am passionately supportive of philanthropy. I've posted 1000s of open contributions into the public domain which I hope are being used. As far as the test LP is concerned, I only seek to avoid exploitation of philanthropic contributions for profit. I seek no payment. If it was ever simply said and agreed that the Test record project is not for profit and for Diyaudio members, I would have had no problem.
We are all on the same side of philanthropy and open contributions, I'm sure.
That's because I didn't find any significant FM above 40Hz in practice. When you think about the physics involved, and what has to move, that isn't a surprise. 60Hz IM products should be rejected because they aren't real FM.
Another time, Hans. Thanks, and all the best for 2018.
LD
That the LP would be sold at cost to everyone was always the understanding in my mind. I have discussed this numerous times with SY, making more than coffee money from these DIY adventures is a fantasy.
On the tracks; Locked grooves already exist on the Chardas test LP. Measuring the impulse response of a system is signal processing 101 far predating anyone here, though deriving it from a low frequency chirp is trivial if you insist. I suggest we replace the complicated trackability tests with ones selected from pre-existing test LP's there are plenty to choose from.
Thanks for the list.
2, 4, 5 no problem, not enough info to use them anyway
8 - just name for a 1K tone track.
1, 3, 6, 7 are all standard practice, no one has a claim on any of those.
well the thing was to be able somehow to remove eccentricity effects from a speed stability test ('cleanness' of BIM peak, bearing in mid different results from different test LPs on same TT as shown earlier in this thread)..or least assess it's impact. Wrt assess it's impact, this knowledge can be transferred to all tracks on same side. Yes reverse side would need another similar track if those signals are eccentricity sensitive as well
But if speed stability test is conducted on same track (closed loop) as the eccentricity test = you could deconvolve the deterministic part (eccentricity). Having closed loop easier forces a known revolution vs signal (maybe w marker)..having several revolutions recorded of exact SAME signal (as in closed loop) can provide suppression of some random effects (if speed stability is uncorrelated/ in-harmonic vs a revolution) = better eccentricity signal estimation + enforce a simple deterministic 2D eccentricity model on top = you could extract a clean as possible this effect -> apply it in a deconvolution model to clean up a BIM spectrum etc.
Where I see problem is ability to make closed loop good enough:
1) spline a signal .. well the higher the signal freq the less 'noise' problem a phase jump (or click) as spline point will be. Actually if phase jump at spline point is not a problem from 'noise' point of view = it automatically can act as marker for one revolution .. and you do not need to design & cut a signal being exact harmonic to a revolution at a specific circumference = easier to master
2) other factor is quality of a circle .. if the track is 'egged' then we have 'eccentricity' noise that is hard to remove, as we would need to have model for 'egginess' to do so.
I could try to simulate something (have no real data to play with) teh few days
No, so far we haven't gotten pressing only quotes. There are a number of mastering houses that don't press, but do regular business with pressing plants. We can try that, for sure. I'll revisit that. Many of the plants have prices posted online.
I'll also get in contact with Aardvark mastering in Denver. Wait! You're in Denver!
I'll also get in contact with Aardvark mastering in Denver. Wait! You're in Denver!
I've thought about this for our locked tracks. Since we know the duration of a single turn, the signal could be made a millisend or 2 shorter, so no spline or bump. There would just be a tiny silence once per revolution.
sad.... was out just a day, partying in paris, and what a discussion has evolved here! anyway i still hope we will not derail this (non profit) project, and i am actually glad i only proposed to handle the logistics, which seemed the more complicated part just a couple daysago , but no more 
by the way, no response yet from GZ. not surprising as these folks must be on christmas holidays now.
thought about the center hole tolerance, and obviously one way would be to make the hole larger and center with a disk clamp; i am not sure however that i could do that at 200micrometer or better...... so i prefer any electronic way to calculate out that error.
which raises another topic - with the test signals, would there have to be any software people should use , and how would write that and make it available? not to enter another copyright discussion......
by the way, no response yet from GZ. not surprising as these folks must be on christmas holidays now.
thought about the center hole tolerance, and obviously one way would be to make the hole larger and center with a disk clamp; i am not sure however that i could do that at 200micrometer or better...... so i prefer any electronic way to calculate out that error.
which raises another topic - with the test signals, would there have to be any software people should use , and how would write that and make it available? not to enter another copyright discussion......
Partying in Paris, eh?
Same for me. Got some auto-replies about being out until Jan. 2.
The polar plot is python and public, thanks to Scott. There are many other software choices for other analysis, some free, some paid. And hardware, of course. Compiling a list of what software to use and how will be an important part of the documentation.
Mulling it over I can think of a number of uses for a locked, modulated groove* and there may be more to come, so for me if this is one we can put to bed in terms of what can be cut and how to make it as useful as possible that would be a great start to 2018.
So sorry if I missed this, but if we pick the right frequency(ies) to be an integer number per revolution can it be cut to be a continuous tone or will there have to be a discontinuity at one point (which we probably use to our advantage anyway).
Have we had any feedback on this other than knowing it will cost more?
*So far, frequency stability, wear and coast down/friction.
So sorry if I missed this, but if we pick the right frequency(ies) to be an integer number per revolution can it be cut to be a continuous tone or will there have to be a discontinuity at one point (which we probably use to our advantage anyway).
Have we had any feedback on this other than knowing it will cost more?
*So far, frequency stability, wear and coast down/friction.
Partying in Paris, eh?
including a real concert with real instruments, now can i please have a testrecord to adjust my TT to sound at least close?