Page 1 of 1

What's wrong with this Plextool Scan (PIE SUM8)

PostPosted: Tue May 17, 2005 11:33 am
by Halc
Take a look at this graph (particularly the graph and the status area average):

http://www21.brinkster.com/roina/nec_nd-3540_8x_ritek_dvd-r_8x_g05_plextor_px-716_sum8_hi.png

Anything odd?

PostPosted: Tue May 17, 2005 12:41 pm
by dolphinius_rex
Hrm... it's a little red x in a box... that's pretty odd I think :wink:

PostPosted: Tue May 17, 2005 1:52 pm
by Halc
Well, look at the graph.

What's the lowest PIE figure? 60+ something?

What's the PIE average in the Test Result window? 11.01

Does not compute, unless the graph is drawn based on maxima alone and there is a lot of minima being ignored for the graph.

I don't understand how the average can be so low, if the graph is right (or how the graph can be so high, if the average is right).

Also, I'm not sure I've stated this before, but for the speed Plextools scans at, I'm still stupefied that they _still_ have not adapted true floating SUM8 window as per specification requirements. They are reading and graphing consecutive SUM8 block samples.

<sigh>

Time to start trying PxScan....

PostPosted: Tue May 17, 2005 2:04 pm
by MediumRare
Halc: this is what we see until we explicitly open the image: Image

Halc wrote:Does not compute, unless the graph is drawn based on maxima alone and there is a lot of minima being ignored for the graph.

I think that is the case- the graph shows the maximum value in each sampling bin (pixel) - one of the reasons alexnoe developed pxscan.

CD-Speed goes the other direction: it bases the average on the the displayed values (and samples for the display like PlexTools). This means that scans with the same length and same PIE count can have "average" values differing by factors of 2 or more (depending on the distribution of the errors in the displayed sample bins).

G

PostPosted: Tue May 17, 2005 3:19 pm
by dolphinius_rex
Ahh, now I can see the picture... once I copy and paste the address into a browser! :o

I don't think it's out of the ordinary. I think the software is not showing the accounts of '0' PIE which will offset the overall average, even when there are large PIE values as well.

PostPosted: Wed May 18, 2005 2:21 am
by Halc
Sorry for the image hosting problem. Didn't understand it right away from Dolph's post, sorry :) [Edited to a link now]

I think the way Plextools graphs is brain dead. It has not statistical logic at all, at least in all the statistically accurate graphs I've seen in my life so far :)

Ah, well. Back to work.

PostPosted: Wed May 18, 2005 2:37 am
by dodecahedron
Halc wrote:Sorry for the image hosting problem. Didn't understand it right away from Dolph's post, sorry :) [Edited to a link now]

clicking the link i get:
The image “http://www21.brinkster.com/roina/nec_nd-3540_8x_ritek_dvd-r_8x_g05_plextor_px-716_sum8_hi.png” cannot be displayed, because it contains errors.

strangely, before i did see the pic!

FYI, your avatar isn't showing either.
it's like Image
also, in the past i would occasionally get a small warning icon in the IE status bar about cookies being blocked when loading a forum page, and the URL that was causing the warning was your avatar!

PostPosted: Wed May 18, 2005 8:05 am
by Halc
Thanks for the info. I had no idea about this. I haven't had any problems with brinkster before, but no it seems there are plenty. Also, I didn't know they are feeding cookies to even people looking at images from their hosting servers.

Weird :)

PostPosted: Wed May 18, 2005 8:07 am
by dodecahedron
what do you know...now i do see your avatar!
weird go figure.
and that cookie thing in IE that was on and off as well.

EDIT: and now (3 hours later) i don't! :o #-o

PostPosted: Tue May 24, 2005 2:19 pm
by Muchin
Halc wrote:Well, look at the graph.

What's the lowest PIE figure? 60+ something?

What's the PIE average in the Test Result window? 11.01

Does not compute, unless the graph is drawn based on maxima alone and there is a lot of minima being ignored for the graph.

I don't understand how the average can be so low, if the graph is right (or how the graph can be so high, if the average is right).

Also, I'm not sure I've stated this before, but for the speed Plextools scans at, I'm still stupefied that they _still_ have not adapted true floating SUM8 window as per specification requirements. They are reading and graphing consecutive SUM8 block samples.

<sigh>

Time to start trying PxScan....
The PIE averages were erroneously calculated as SUM1 by Plextools. PxScan is correct in the calculation.