Official K-Probe Discussion (Tool for Scanning C1C2/PIPO)

[rdgrimes]

the graph settings are now saved when you close the program. Very nice.

[CDHero]

Hi rdgrimes :
Yes , I add the function !
In fact , I also fixed some bugs and add some features.
1. KProbe will push on and try to find next readable sections when it encounter "seek error".KProbe will jump the defect area.And the interval is 75 blocks.

2.I add DVD Write feature in Tilt Analysis Tab , but I am so sorry you cannot experience it !

3.Now the settings can be saved.

[rdgrimes]

There's something very odd going on with version 1.1.5. I'm seeing many messages about "read errors, jump to xxxxxx" and it freezes at some points requiring an end task to close it. This is on a disc with high error rates. I suspect that Karr is trying to address the issue of the test aborting with high errors. It appears to be trying to skip over errors and not report them. The result seems to be that it's skipping and not reporting most of the C2 errors, and freezing the program with higher rates.

hi Karr, you posted while I was writing this. I think the fix may be worse than the problem was. I have to set the read speed very low to complete a scan with high error rates.

[CDHero]

There's something very odd going on with version 1.1.5. I'm seeing many messages about "read errors, jump to xxxxxx" and it freezes at some points requiring an end task to close it. This is on a disc with high error rates. I suspect that Karr is trying to address the issue of the test aborting with high errors. It appears to be trying to skip over errors and not report them.
hi Karr, you posted while I was writing this. I think the fix may be worse than the problem was. I have to set the read speed very low to complete a scan with high error rates.

Mmm, I want to add readability and tolerance originally , but as you said ,
the fix may be worse than the problem was.
You should be able to hear the drive try to read again and again.Maybe I must modify the algorithm of KProbe when it encounters high error rate area !!

Thanks for your comment!!!

BTW , I am glad to cooperate with you.

[rdgrimes]

I can hear the drive slowing down. But the program freezes very predictably when the errors reach a certain point. I think it's an interesting problem, how to deal with higher errors. CDDoctor just skips over them after re-reading a couple times. Personally, I consider the disc to have failed the test if it can't be read at full speed, but I prefer that the program would complete the test and display the full graph with the missing information being evident.

Also, one time the program froze at 103% !! while trying to read errors. Interesting because the disc is only 79 min.

[cfitz]

Mmm, I want to add readability and tolerance originally , but as you said , the fix may be worse than the problem was. Maybe I must modify the algorithm of KProbe when it encounters high error rate area !!

I noticed this new feature right away and was happy to see you implementing yet another new feature we have requested. You are very responsive. Thank you! The algorithm may not be perfect yet, but it is a first try. Please keep working, and I know you will get it to work well. I imagine trial and error will be required to tune the algorithm.

Here are some sample plots for comparison:

K's Probe v1.1.4:


K's Probe v1.1.5:


CD Doctor:


K's Probe still stops earlier than CD Doctor, but I hope you can overtake CD Doctor with more work.

BTW , I am glad to cooperate with you.

And we would like to cooperate with you too. If there are things we can do to help, please ask us.

Personally, I consider the disc to have failed the test if it can't be read at full speed, but I prefer that the program would complete the test and display the full graph with the missing information being evident.

I agree on both counts.

Also, one time the program froze at 103% !! while trying to read errors. Interesting because the disc is only 79 min.

It probably read a bad LBA that caused it to try to seek beyond the end of the disc.

cfitz

[rdgrimes]

I think there is some value, from a pure testing perspective, to having the program read at a fixed speed with no re-reading or skipping, no slowing down. We know that in reality, a drive will be able to slow and re-read those sectors, but for testing purposes why not have it just plow through and report the errors? It would be faster too.

What about having the program go ahead and abort the test, but display the full graph and pop up a "read error" message to inform the user that the disc has failed? Obviously it would be better if the program did not abort at all, but if the error rates exceed the ability to report or cannot be read at the set speed, it has to do something.
The pop-up error message could say something like "read errors, the disc has failed the test, please lower the read speed and try again". A little wave file of crash sounds could even be added. Too silly?

[cfitz]

I think there is some value, from a pure testing perspective, to having the program read at a fixed speed with no re-reading or skipping, no slowing down.

I'd still like to see it at least skip over and try to continue on. I had earlier asked if it was possible to indicate where uncorrectable errors occur (see here). Perhaps unreadable sectors could be indicated as well, if they aren't the same thing.

A little wave file of crash sounds could even be added. Too silly?

Yes, definitely too silly. Let's not take up Mr. Wang's time adding things that don't contribute to actual test and measurement.

cfitz

[CDHero]

Dear Rdgrimes and cfitz :
Can you describe the disc you test ? What's the type of defect ? scratch ?
finger print ? or black area ? And I also want to know the size of defect area ,thanks !!

I made a scratch disc , and I test it both on KProbe and CD Doctor ,
Kprobe attempt to jump behind the defect area , it failed finally , but
at least you can push eject button to stop the test.
CD Doctor didn't jump , it retry to read the same sector , and finally my
system crashed !!
So I am hard to duplicate the situation you met.

And , I have an issue.While seeking error or reading error , drive is unable to get any information from disc , even wobble address , how
can the testing program report C1/C2 error ?

[cfitz]

Dear Rdgrimes and cfitz :
Can you describe the disc you test ? What's the type of defect ? scratch ?
finger print ? or black area ? And I also want to know the size of defect area ,thanks !!

Mine are black marks drawn with a permanent ink marker. I drew them in arcs circumferentially along the disc to provide maximum grief for the drive. I drew them by hand, so they aren't exact, but they are increasingly longer with each mark. I would say roughly 0.5mm, 1 mm, 2 mm, 3 mm, 3 mm, 5 mm, 12mm and 20mm. I should try to do a better job and make a more careful "bad" disc.

And , I have an issue.While seeking error or reading error , drive is unable to get any information from disc , even wobble address , how can the testing program report C1/C2 error ?

I don't think it can, can it? From where would it get the information? It seems to me it would just have indicate by some means (another color, perhaps) that the area is unreadable.

cfitz

[rdgrimes]

CDspeed allows the drive to do it's normal slowing and re-reading and finally reports the sector as unreadable (red). This clearly is a more "real-world" sort of approach, and reflects how the disc might be expected to behave. I still prefer the "full speed ahead" approach because that's how i expect my discs to perform. CDSpeed can report the sectors as unreadable and then moves on (in most cases) to finish the test.

One other observation: would it be better to eliminate the options to select colors and stick to the "green(C1) - yellow(C2) - red(unreadable)" to make it consistant with the others?

Yes, definitely too silly.

OK, I dadn't had enough cafeine yet. It's morning here.

Can you describe the disc you test ?

Karr: this is the same disc that I have tested and posted in the CDFreaks thread. The colored TY(Fuji) disc with 2 large spikes of C2 and unreadable sectors. at 8x read speed, the C2 peaks at 120.

[cfitz]

CDspeed allows the drive to do it's normal slowing and re-reading and finally reports the sector as unreadable (red). This clearly is a more "real-world" sort of approach, and reflects how the disc might be expected to behave. I still prefer the "full speed ahead" approach because that's how i expect my discs to perform.

Agreed. Can my recent comments be misconstrued to read as if I meant otherwise? That wasn't my intention. I've said it before:

By the way, I personally feel that K's Probe and CD Doctor's strategy of not allowing the speed to drop to improve readability of difficult sectors is a reasonable choice for a testing program where you want to control the variables as much as possible.

(From your fine home forum here )

One other observation: would it be better to eliminate the options to select colors and stick to the "green(C1) - yellow(C2) - red(unreadable)" to make it consistant with the others?

Everyone has different opinions on GUI's, and once given choices it's hard to take them away... But, I personally have no objections to standardizing color schemes if it makes Mr. Wang's job easier, and your suggestions are reasonable choices. I would only request that if line colors are standardized, please choose the background color suitably so that all the line colors are easily visible.

cfitz

[cfitz]

By the way, Mr. Wang, am I being too formal by calling you Mr. Wang? Would you prefer to be called Karr?

cfitz

[rdgrimes]

Here's my infamous colored Fuji:

Kprobe will freeze at full speed read, at 32x it aborts at about 77min but does not freeze. At 24x, Kprobe has never worked for me on any scan. Here's 16x, I also note that max and total error counts are quite a bit lower than in earlier versions.


Full speed in CDDr




CDDr will not usually finish the test at full speed, but it usually gets a lot farther than Kprobe does. It also appears to be skipping large chunks though.
I guess my preference would be to have those unreadable sectors displayed as unreadable, say red, in the way that CDSpeed does. The disadvantage of the CDSpeed method is that it removes the read speed control from the user, but trades that for the pretty RPM display so you can at least see what the drive was doing. CDSpeed also fails to complete on some of the really bad discs, but it will re-read at 2x several times before failing. ( I presume this is the drive itself controlling the speed?)

[MediumRare]

I agree that maximum reading speed is preferred- both cfitz and rdgrimes ave made excellet suggestions. I have no strong preference re standard/flexible colours.

Comments on V1.1.5, the good things first:
- I like that the plot settings are saved from run to run. Maybe Karr Wang can also save the chosen drive and reading speed ?
- When the charts are combined, you no longer have to choose how many or which charts to save!

But I ran into the same problems as rdgrimes and cfitz did with poor quality disks.
I tried my "bleached crappy audio disk" with 1.1.5 too. This is an 3-year old 4x burn with audible hiss at the end, probably due to UV (not deliberately damaged). It has a couple of peculiarities beyond the other samples shown above. There were a lot of "seek" gaps at the beginning, where the error rate was low in vers. 1.1.4. At about 75 min. there was a long pause followed by the maximum/minimum error and the funny 2-part time scale seen on the plot:


Obviously the Probe is getting confused here. At present, the older version works better, but I'm sure that Karr Wang will find a solution!!

G

[rdgrimes]

Notice how everybody has a few of those really horrible CDR's laying around to use to torture testing programs?
It seems that the CDR makers will never tire of selling us those kind of discs.

[MediumRare]

Notice how everybody has a few of those really horrible CDR's laying around to use to torture testing programs?
It seems that the CDR makers will never tire of selling us those kind of discs.

I guess otherwise we wouldn't have as much to talk about. Personally, I went through my junk box after CD Doctor showed up and tested everything in sight.

G

[rdgrimes]

I have every disc I've ever tested in spindles on the shelf, don't ask me why. There's a couple hundred there.

[CDHero]

I have a lesson today , "WDM USB Driver Programming",
so I won't update KProbe.

[rdgrimes]

Study hard!

[dolphinius_rex]

I have every disc I've ever tested in spindles on the shelf, don't ask me why. There's a couple hundred there.

I also have every CD I've ever burned. I'm really happy that I kept them because I'm now going through them all and learning who manufactured each one, and how they are standing up quality wise. It makes for great beta testing media for K-Probe too

[Bhairav]

Awesome work guys! Keep up the testing, and lemme find some webspace so I can post some scans with some crappy Princos

Hey Ian, how about giving Mr Wang a custom nickname, say something like "Karr Wang, KProbe Creator Man"?
Sounds nice don't ya think

[CDHero]

I update KProbe to 1.1.6
I change the algorithm for seeking or reading error.
Now KProbe shoud have more tolerance for bad disc.
And I also make an "awful" testing disc.
KProbe will test full disc .But CDDoctor will hang at 23min.

[rdgrimes]

Hi Karr. Keep up the great work! LiteOn clearly doesn't know how lucky they are.
Same links, I presume?

[CDHero]

I also make some change
1.The default value of "Realtime Chart" is now unchecked.
If you still want to show realtime chart , you maybe lost its accuracy. I didn't describe this issue before , sorry !!
Why realtime chart will lost accuracy ? It is because KProbe waste some time to update chart.
2.I add user-defined "Error Value" and "Error Color" for seeking and reading error.KProbe will show these error with different color and value,
default value is 1.