The Big KProbe2 Thread

Yesterday i tried KP2 and everything works fine, today i have installed my new 812s and running Kp2 it restart my PC. Anyone can help me?

My PC: Abit Nf7, 512DDR400 TwinMOSS, 2400+, 9700Pro, LTR 48246S, SOHW 812S, Maxtor DP9 60gb, 120gb, 160gb (EXT).

I know,as for the speed,i use CD Bremse,i managed to maintain the speed from about 2,3x to 5x at the end.

My question is,that since i use the same drive,the same disc,why on earth with KProbe 1.29 the PO appears lower than 10 and with KProbe 2 appears to reach easily 3-4 (24,32 with the "old "ECC 8 measurement),indicating a disc "On the limit".

The PI seems to be similar with the 2 versions.And also quite logical.When i burnt an AN31 i had 700 PI and raising PO at the end of the disc confirmed by NEro DVD speed with a net reduction of reading speed.

I think i ll keep using the 1.29.Seems more reliable to me with my DVD ROM drive.I mean,there i was getting a PI and PO was always lower.

Now with KProbe 2 seems the opposite,it simply doesn't make sense.


Thanks for the info about the singular spikes.

Unfortunately, Kprobe2 still hangs on startup when you run it on a system with SCSI CD/DVD drives present, like 1.x did.

To all those who suffer from this problem, you can get around this by doing the following:

• Create an empty file "wnaspi32.dll" in the directory you installed Kprobe2 to (you don't have to do this if you don't have ASPI installed).
• Close all programs that are using ASPI (this may or may not be necessary, I didn't try).
• Go to device manager and disable all your SCSI CD/DVD drives.
• Run Kprobe2. You'll get an error message saying that wnaspi32.dll is not a valid Win32 image. Click OK.
• Kprobe2 will start in SPTI mode. You can delete the wnaspi32.dll file now (you won't need it on subsequent runs).
• You can now reenable all SCSI CD/DVD drives. Remember to disable the drives each time before you start Kprobe2 or it will hang.
• Never switch to ASPI in Kprobe2. This will cause it to hang, even when you have disabled your SCSI drives in device manager.

This has been tested with Adaptec's ASPI 4.71.2, it should apply to all other Adaptec ASPI versions as well, but might be different with other ASPI implementations.

To the author(s) of Kprobe:

Is there any possibility that future versions of Kprobe will pop up a list of available controllers on first startup (resp. whenever a change is detected) and let you select which ones to include in the bus scan and which ones not?

no problem ive seen alot of persons using the 166s they seem to be pretty accurate for dvd readers with the hacked speed controls.

I too have noticed that k2 seems to be more harsh on all errors overall..little higher with both. maybe its more accurate?

heres a example with spikes removed burned at 6x with my nec 2500

Thank you for this trick!
Hope that next Alcohol's VDevice will be eide =)

Does the spike in the burn, whose scan is attached, indicate a problem with the 851S itself? I seem to recall that Lite-On DVD burners had a problem with spikes on CD-R media?

Here's the scan..

nothing wrong with burn because of spikes. just remove the highest point by right clicking until the high spikes are gone.

heres the same disk scaned with Kprobe 1.29 noticed that the POs are about the same even at 8ecc not sure why that is though i would think they would be less with the k2 scan using 1ecc. and yes the PIs seem to be about 10 to 20 points less than the k2 scan.

@Jamos

I feel i am taking advantage of your patience,but i m looking at your results,and the same thing comes to my mind.Yes the PI shows top values higher with 2.0.1 but the average is close and the pattern is similar.

The issue is the PO. Again you have very similar results,a spike (not singular) at about 12 and a couple at about 8.And similar patterns (profiles).

Now,the problem is that with 1.0.29 the ECC in PO is at 8 (so the limit is 32) and so you 're fine.

But in 2.01. the ECC is 1 BUT you still get the 12,8,8 error peaks,which in this case IF i am not mistaken ,in ECC 8 should be like 96,64,64 ! This is a hardly readable disc in theory ,isn't it?

I mean,with 2.0.1 your PO values should be (for the non singular spikes) up to 4 ,correct? But you get almost the same as in 1.29 ,right?

Well,if i have understood the limits correctly,(32 if ECC is 8 and 4 if ECC is 1) ,then there are 2 cases:

1)KProbe 1.29 is correctly calibrated and you have a nice burn ,supported from the fact you have a RITEK there.

2)KProbe 2.0 is right and you have a poor burn that probably shouldn't even get read with POs like 12,8,8 (I mean 12 is 3 times the limit,it's pretty high!)

Do i understand well the situation or am i making a mistake somewhere?

The DVD standards are for 1x scanning speed.

Yes,i have read the relative FAQ,but as in KProbe 1, the 4x was conventionally agreed to be the default speed to velocitate a bit the scans,remaining to the limits of 1x as standard.

And the 1x rule doesn't change what i ask here,because BOTH scans (with KProbe 1.29 and KProbe 2) in theory should have been done at 1x ,but both are done at 4x ,so the results are comparable.

But it simply seems to me there is something WRONG,if i have understood the limits.

The same disc,in both cases at 4x gets.

In 1.29 : Max a 12 spike (non singular) with limit 32 ,so it's a nice burn

In 2.01 gets AGAIN a 12 spike with limit 4(!) ,so it's a disc to throw to the garbage.

The issue,is similar to the one i saw with my results in the previous page.

With the "old" Kprobe,my burn was very good,max serious PO at 10-12 i think,the rest under 5 as average with 32 as limit.

With the new KProbe,i have the PO often at 3-4,that should be 24-32 with the old KProbe.

Pitty i threw away the AN 31 disc i had,i 'd be curious to see what the new KProbe would read.It was already garbage with the old one,but still readable.

I think i ll make more tests tonight.I ve saved all the results of every DVD i 've burnt with the old KProbe,i ll try to see how the new onw reads them.

PO is the problem.PI seems more or less coherent.

You don't need to worry to much about single spikes exceeding 4. But larger areas exceeding 4 could be problematic.

I know.In fact i speak of the "Non solitary" spikes in Jamos' result:
There is one not so "fat" right on the start and almost immediately after,a bigger one.These aren't "solitary" are they?
According to the limit of 4 in ECC 1,that area should AT LEAST have a reduction of reading speed (ex in Nero DVD speed) ,correct? BUT the same 12 peak (non singular) appears in Jamos' KProbe 1 result,but there the limit was 32,so he shouldn't have any reading problems at all.So i think on of the 2 KProbes is miscalibrated or doesn't read correctly.I don't know how to explain better...

http://club.cdfreaks.com/attachment....&postid=583234

Hi...
I can tell you all about the the ECC sum in Kprobe2

1.> MediaTek Chip returns the PI/PO value every 1 ECC block

2.> KProbe1 collects 8 PI/PO values and sums them to a PI/PO record.

3.> KProbe2 collects 8 PI values and sums them to a PI record. And it saves every PO value as a PO record.

So, I am sure that KP1 and KP2 are all correct.
You can zoom in and zoom in the Pi/Po chart to see the interval. You will find the chart should be correct. All data are returned by chip. That is a very simple algorithm.

In fact , I don't care the exact number, but the bad zone.

Observe also the "fat" spikes on Jamos' results to the right (near the end) of the scan.

In Kprobe 1 he has a "fat" 4,2 peak.And before it another series of similar ones.

In 2.01 the peaks appear almost in the same positions(so seems they are not artefacts,but they trully exist) ,BUT their value is a little below 4.

Now the counts don't turn here,do they?

4,2 with a limit of 32 is fine.

3,8 with a limit of 4 is a different story.Corresponds to 3,8x8=30,4 in ECC8 of the first KProbe ,correct?

Well,this is like night and day!

Ah,Thanks Karr,this seems to change things,i always hated mathematics ,so it will give me some headache to think about it,but i understand it is almost a different way of counting the errors ,not just a difference od 8 to 1 scale.

I wish it was simpler I expected something like "if in KProbe one you 're well below the limit,then the same will be in KProbe 2 proportionally"


Thanks and sorry for the trouble to all.God,i hate maths

Ill rerun the test at 1x speed. i have a feeling it will make a difference with k2 more than it did with k1. also i will run a nero check on this same disk to see how the graph is. maybe we shouldnt just say less than 4 POs at 4x speed but relax it a bit as the disk seems to read fine on all my players..and i have some cheap ones. Also im really kind of weary of the quality of the reads by this drive..whats the best liteon drive to get accurate scans from in your experience?

Ack! runing kprobe 2 at 1x getting whacky results way higher than normal..like 1000s for PIs and 100s for POs..also getting the puh skipped at LBA xxxxx ignore 1 ecc block message (whatever that means).

Ill just stay with 4x hehee..

BQuick
You're making too much of nothing. These errors do not physically exist on the disc, they are much less well-defined. The same disc will never produce the same scan twice. Small differences like these are insignificant in terms of disc quality and read-ability. And yes it does make a differenc that the scan is done at 4x rather than 1x. So just relax, and don't worry till you see PO much higher than that at 4x speed.

PI/PO measurment is all relative. It's not so much about what the values are, but whether disc A is better than disc B and what combination of media, drive, and speed works best for you.

Don't fall into the trap of looking for absolute accuracy in anything, it doesn't exist, and it's not required. Your best choice is to remove KProbe1 and move on.

proportionally in KProbe2 ? It is impossible.
As I said before, chip returns the data every 1 ECC block(16 blocks).
But chip is passive, Kprobe must issue command to get the
data. But it is impossible for kprobe to issue command every
1 ECC block(16 blocks).. the best situation is exactly 1 ECC
block(16 blocks), but most of situations is about 19,20 even 30
blocks at the end of disc. So Kprobe will lose its precision in low performance system ,especially when the max speed is selected.
Because you know, the DVD speed is fast !!
But I can say, Kprobe2's performance is better than Kprobe1.
So K2 should be more accurate than K1.

great news, so we need to get enough of these probes done now to compare good to bad burns so we can get a general idea of where we will start getting read problems..ie prolly around 12 for POs id say from the looks of my scans (I know 4 would be the standard on a 100% accurate system) but coming up with a good number like 280/32 for k2 would be cool (yes i know it should be 280/4 on a perfect system) but like Karr has said its impossble for kprobe to issue a command every 1 ecc block..im gonna retest with the PO sum set to 2ecc or 4ecc and see how it goes.

also im using SPTI drivers cause my drive is external not sure if that infuences anything.

a hot key to delete highest error would be great too..it would releave right clicking on graph and clicking to get rid of the high spikes one at a time..very nice work though!

---

Last edited by Jamos; 27-04-2004 at 19:18.

After installing kprobe2 my PI values are 6 times higher then measured with kprobe 1.1.29. PI errors went up from 100 to 600.
(Ofcourse with same disc burned this morning) Installed 1.1.29 again next to kprboe2 and did a scan on 1.1.29. Now that version also gives PI values of 600.
Im getting tired of my lite on 811
could this be caused by installing the new aspi layer by kprobe2

same disk scaned again with all the single spikes removed..its pretty consistent with the Kprobe 1. i really like the remove highest error feature. it shows this disk is a pretty good burn. Im gonna just run kp2 from now on.

could be..did you install the adaptec ones? try it again using the spti drivers..ive been using them wihout any problems. (well i have to i have a external drive)