Become a Site Supporter and Never see Ads again!

Author Topic: CardTest card testing donationware program  (Read 2227 times)

0 Members and 1 Guest are viewing this topic.

Offline Ozpeter

  • Trade Count: (0)
  • Taperssection All-Star
  • ****
  • Posts: 1109
CardTest card testing donationware program
« on: June 12, 2008, 11:56:12 PM »
http://sourceforge.net/projects/cardtest is the place to download it.  I did a quick search and couldn't see a reference to it here.  Note that testing a card destroys any data on it and requires a subsequent reformat, so don't use the program on a card that has not had its contents copied off!

Offline Ozpeter

  • Trade Count: (0)
  • Taperssection All-Star
  • ****
  • Posts: 1109
Re: CardTest card testing donationware program
« Reply #1 on: June 13, 2008, 12:15:49 AM »
Interesting how card speed varies with data packet size.  For a Transcend 8GB class 6 card I am testing, I get this reading from the CardTest program:-

Packet size in bytes / speed of write

Edit - figures in [ ] are from a different card reader.

8k / 0.74 MB/s [1.??]
16 / 1.26 [16.38]
32 / 1.64 [16.38]
64 / 5.04 [2.85]
128 / 6.9 [11.92]
256 / 8.46 [11.40]
512 / 2.24 [11.16]
1M / 9.53 [4.46]
2 / 5.84 [5.83]
4 / 5.59 [8.13]

I wonder whether this could be why some cards behave better in some recorders - the packet size written by the recorder may or may not suit the packet size preferred by the card.  Note that with the two different readers, the packet size / speed relationship is different.
« Last Edit: June 13, 2008, 01:50:08 AM by Ozpeter »

Offline Ozpeter

  • Trade Count: (0)
  • Taperssection All-Star
  • ****
  • Posts: 1109
Re: CardTest card testing donationware program
« Reply #2 on: June 13, 2008, 12:58:06 AM »
More data... this time from a Panasonic 8GB card. 

Packet size in bytes / speed of write

Edit - figures in [ ] are from a different card reader.

8k / 8.19 MB/s  [1.64]
16 / 1.?? [1.??]
32 / 1.49 [32.77]
64 / 32.77 [9.36]
128 / 18.72 [43.69]
256 / 8.46 [17.48]
512 / 16.38 [8.32]
1M / 11.16 [16.91]
2 / 12.19 [4.47]
4 / 5.07 [5.07]
« Last Edit: June 13, 2008, 02:11:45 AM by Ozpeter »

Offline Ozpeter

  • Trade Count: (0)
  • Taperssection All-Star
  • ****
  • Posts: 1109
Re: CardTest card testing donationware program
« Reply #3 on: June 13, 2008, 01:05:42 AM »
Full range transfer speed test results:-

Transcend 8GB

Max read - 21.74MB/s
Max write - 16.13
Avg read - 14.43
Avg write - 8.95

Panasonic 8GB

Max read - 21.74
Max write - 21.74
Avg read - 14.51
Avg write - 11.52

Curiously tests on both cards terminated with a read error 75% through the test.  And the graph of the write speed also showed a significant speed reduction (by half) half way through.  I'm not sure what to make of that.  Still, the Panasonic figures are generally better than the Transcend, but perhaps so they should be, given that typically the Panasonic cards cost twice as much.

[Edit - results from a second card reader are very similar]
« Last Edit: June 13, 2008, 02:13:09 AM by Ozpeter »

Offline Ozpeter

  • Trade Count: (0)
  • Taperssection All-Star
  • ****
  • Posts: 1109
Re: CardTest card testing donationware program
« Reply #4 on: June 13, 2008, 02:05:47 AM »
Meanwhile, I've come across a mine of card info and test reports (in cameras) at http://www.robgalbraith.com/bins/multi_page.asp?cid=6007 - the introductory page there makes particularly interesting reading.

Fake memory cards and how to spot them - http://www.overclockers.com.au/wiki/Fake_Memory_Cards

« Last Edit: June 13, 2008, 03:25:43 AM by Ozpeter »

Offline Ozpeter

  • Trade Count: (0)
  • Taperssection All-Star
  • ****
  • Posts: 1109
Re: CardTest card testing donationware program
« Reply #5 on: June 13, 2008, 02:27:23 AM »
Heh, now I'm trying the same test using my R-44 as the reader/writer, and coming up with completely different results - it's much slower.  I'll post the results in a while (running the test is going to take some time!) but that opens up another whole area of consideration.  Is the slower speed a result of the R-44's USB communications being slow (though supposed to be USB2.0) or is it because its data recording/reading system is slow?  And if the latter, is that because it doesn't actually need to be fast?  And if it doesn't need to be fast and isn't, is there any point in using fast cards in it?

Interesting stuff!  Well, I think so anyway...

Edit - results are in for the Panasonic 8GB card using the R-44 as a card reader -

8k / 0.63 MB/s
16 / 1.49
32 / 1.93
64 / 2.62
128 / 2.26
256 / 2.60
512 / 2.58
1M / 2.39
2 / 1.72
4 / 1.85

Max read - 4.00
Max write - 2.56
Avg read - 2.81
Avg write - 1.76

Slow, huh?!  Yet it's running off the same USB Enhanced Controller that the card reader used in the first test runs off.
« Last Edit: June 13, 2008, 03:50:26 AM by Ozpeter »

Offline Ozpeter

  • Trade Count: (0)
  • Taperssection All-Star
  • ****
  • Posts: 1109
Re: CardTest card testing donationware program
« Reply #6 on: June 13, 2008, 04:28:15 AM »
Now moving on to some real world tests, and some thinking aloud about data rates -

I created a 20 minute file, 24/96, which is 659MB in size.

I transferred that to the Panasonic card in the R-44 via USB, and it took 4'34" to copy across.

According to my maths, that's a data rate of 2.41MB/s.

The required data rate of the R-44 when writing two stereo tracks is (659 x 2) / (20 x 60) = 1.1MB/s

So even though the USB data rate compared to a direct transfer to the card via a reader is very slow, it's fast enough.

However, putting the card in the card reader and transferring the file to it that way takes a mere 42" - 15.48MB/s I reckon.  So that's way faster than is necessary and way faster than the R-44 seems capable of writing at.

Therefore the logical conclusion would be that there's no point in using a fast card in the R-44 (or for that matter, any other 4 channel recorder running at 24/96).

But am I missing something?  Like, confusing bytes and bits?  Or does the R-44 write audio data in bursts and needs a faster card for that burst speed?

I guess the next test will be to use a standard speed card and see how that fares.
« Last Edit: June 13, 2008, 04:42:34 AM by Ozpeter »

Offline Ozpeter

  • Trade Count: (0)
  • Taperssection All-Star
  • ****
  • Posts: 1109
Re: CardTest card testing donationware program
« Reply #7 on: June 13, 2008, 04:49:41 AM »
Well, data transfer via USB to the R-44 using the same 24/96 20 mins stereo file took the same amount of time using a basic grade Sandisk 4Gb card.  I think they are rated as class 2.

I'm now recording 2 stereo files at 24/96 onto the R-44 onto that same basic grade card, and its buffer display shows nothing amiss.

So maybe low cost slow cards from reputable manufacturers are actually perfectly ok?


 

RSS | Mobile
Page created in 0.194 seconds with 36 queries.
© 2002-2018 Taperssection.com
Powered by SMF