CDRLabs.com

I see you saw the problem

alexnoe wrote:I see you saw the problem

uh, yeah.... geez...

you're going to need Rainman to get the pattern out of THAT group!

I had a look at them out of morbid curiousity. I'm probably not telling you anything new, but there seems to be quite some regularity in the numbers after your "(12) F3 1F 23 00 00 04 00 x0 FF FE 00 00" commands.
My guess is 2 byte timing values for lands and pits with the high bit set to distinguish them. Have you tried doing a histogram based on that?

G
(who isn't Rainman )

Not yet... but there should be either a certain structure within the data (for 3T..11T, 14T), or this is really a plain histogramme needing some processing by software

So far I've only done a numerical histogramme, but it seems to make sense Seems like the scale on the left is log(number of occurences)

MediumRare wrote:G
(who isn't Rainman )

Perhaps not Rainman, but you have an understanding of numbers that both shocks and amazes me!

Someone should send you a PX-716a

I've got a proper (numerical) histogramme working, but i've serious issues now detecting minima in crap disc scans like

alexnoe wrote:...but i've serious issues now detecting minima in crap disc scans like...

As do the drives themselves. Now you know how your poor drives feel when you feed them such crap discs...

In seriousness, you are doing good work. Thanks!

cfitz

L0 inner:


L1 outer:


EDIT replaced picture with current one

wow.. Nice Alexnoe!!

But you need to add more scale info to the display I think.

Simply wow!

Well, the vertical 1,2,3,4 scale in PlexTools seems to be ln(number of occurences)/ln 10.

The visual impression (i.e. narrow peaks vs wide peaks and the position of the peaks relatively to the red lines) is much more important than some scales. The only reason to add a few numbers would be to make it look more like PlexTools

What I could do is adding a lable to each chart what it actually is (like "L0 inner zone PIT T Freq Distribution"), as well as some "separation" lines between the charts

alexnoe wrote:
What I could do is adding a lable to each chart what it actually is (like "L0 inner zone PIT T Freq Distribution"), as well as some "separation" lines between the charts

That would help a lot!!

Also, the ability to scale the images would be nice (although maybe it's already there, I don't have a PX-716a so I can't play with it).

config.ini file:

button1_x=480
button1_y=520
button2_x=560
button2_y=600
button3_x=800
button3_y=860
button_height_correction=0
logscale=1
autoscale=1
pisum8scale=2000
pisum8lines=0
betascale=200
pifscale=200
pieheight=0.6
pifheight=0.4
betaheight=0.5
taheight=0.4
talayerfactor=1
jitterminscale=1200
jittermaxscale=800
capture=1
update_interval=250
displayjitter=1
displaybeta=1
displaymetadata=1
strong_vertical_line_interval=1024
weak_vertical_line_interval=256
strong_vertical_line_intensity=96
weak_vertical_line_intensity=32
vertical_line_weight=0.5
autoadjust_log_scales=0


Virtually everything is scaleable:

- the window can be resized at will
- the ratio between the single tests can be changed at will
- there is a layer multiplicator for double layer media TA tests, i.e. you can select wether or not the TA test should use twice as much space for double layer media
-

can you maybe put some space in between test types? the graphs all kind of mesh together a little too much I think (just my opinion). One thing I would REALLY like to see is maybe a button beside each graph type, that you could push, that would enlarge that ONE test to fill the entire program window, and then another button on that window to return the test to its' regular size and viewability. That would allow for easy examination of various tests without having to resize the window quite large just to examine a single test more carefully. Is that possible??

What do other people think?

(Thanks for the test files BTW!!)

@alexnoe:
kudos for an excellent job!

@dolphin, alexnoe:
good suggestions.

I think that it possible, but more work than what it might sound like, especially the button

Very nice job! Almost makes me want to go and buy a Plextor.

alexnoe wrote:Well, the vertical 1,2,3,4 scale in PlexTools seems to be ln(number of occurences)/ln 10.

Isn't that just the decimal logarithm log10?

alexnoe wrote:... the position of the peaks relatively to the red lines) is much more important

If the red lines are the nominal values, then I think that the relative position is a very important quality index. You could probably chalk up some characteristic numbers from the distributions- e.g. shift and "Q" (= full width at half maximum) for the various peaks.

Keep up the good work!

G

Isn't that just the decimal logarithm log10?

Yep, and as everyone uses different ways to write that (like log10, log, lg, log(10, x)...), i prefer that way to write it

If the red lines are the nominal values, then I think that the relative position is a very important quality index.

Right, but an analysis is more difficult than what it looks like

alexnoe wrote:Right, but an analysis is more difficult than what it looks like

I'm very much aware that this can be difficult. Do you have any groups (e.g. nuclear or atomic physics) doing spectroscopy at the university? This is the sort of thing that occurs there all the time.

Another thing just struck me while looking at your big image- it would be nice if you could indicate where in the disc the TA data is taken, e.g. with an arrow or other marker for each position added to the PI and/or beta/jitter plot. And maybe even make it an option to look at or near a position of your choice. That would be very useful for analyzing problem spots and is something the PlexTools don't have.

G

MediumRare wrote:Very nice job! Almost makes me want to go and buy a Plextor.

you took the words right out of my mouth.

This tool should get pretty popular, given how much it focuses on a higher level of professionalism in media testing (not to be confused with professional testing!).

I also like the idea mentioned, about being able to easily compare problem spots. Maybe if you could use the mouse pointer to "draw" a straight line between graphs, so you can easily compare where different spots corraspond?

Of course, you know the only thing holding this tool back from being the perfect tool for the "professinal amateur" media tester? The fact that it can't easily incorperate a BenQ scan for increased comparison abilities

Could you imagine having a BenQ scan with PI/PIF/POF and Jitter scores all comparable to the Plextor's? Having all that Data in one place, with disc locations easily identifiable between graphs would be almost TOO good! But yeah... that would require a LOT of work... not to mention the ability to input two files into PX View. Maybe you could work with Erik the author of CDSpeed to get him to support outputting data into a file type, that PXView could open to compare?? That would be awesome!!

As the commands to initiate a TA scan do not seem to contain an explicit LBA, it might be impossible to choose regions other than "inner", "middle", and "outer"....

Concerning Benq scans: I'm getting a MSI benq clone for reviewing soon, so when I find the time to reverse engineer their commands for scanning, i might be able to add that to PxScan. That would require renaming the tool of course

alexnoe wrote:
Concerning Benq scans: I'm getting a MSI benq clone for reviewing soon, so when I find the time to reverse engineer their commands for scanning, i might be able to add that to PxScan. That would require renaming the tool of course

Wow... I was *THIS* close to offering to buy you one too!

Anything I can do to help you, just let me know!

Why don't you call the renamed tool "Noe-Scan" ?

alexnoe wrote:. That would require renaming the tool of course

dolphinius_rex beat me to it, byt i was gonna suggest PxBqScan

Nah, since it can do it all, call it: OmniScan

cfitz