Help needed

[kuzes]

Hello, I'm using the code above but unfortunately sometimes the data doesn't seem to encoded correctly and i cant scan the produced QRcode. I'm using arduino mega2560 and this tft http://www.adafruit.com/products/335.
Do you think this problem occurs because of the tft library?
For example: When I try to encode this data:"1879999921" the produced qr code cant be scanned.
Thank you very much

[davidhend]

I noticed the same thing from time to time; the reason is that during the
data masking I am leaving off a step. Basically some QR codes are hard to
decode because there are too many of the same color squares too close to
one another. You are suppose to choose different masks and than rate the
final QR code for readability. I decided to pick a mask and stick with it;
I am not performing this last step. For what I was working on I did not
need 100% readability all the time.

Here is a excerpt from the tutorial I followed that re-explains what I said:
Step 6: Data Masking

Certain patterns in the QR code matrix can make it difficult for QR code
scanners to correctly read the code. To counteract this, the QR code
specification defines eight mask patterns, each of which alters the QR code
according to a particular pattern. You must determine which of these mask
patterns results in the QR code with the fewest undesirable traits. This is
done by evaluating each masked matrix based on four penalty rules. Your
final QR code must use the mask pattern that resulted in the lowest penalty
score. The masking process is explained in the data
maskinghttp://www.thonky.com/qr-code-tutorial/data-masking/section.
Here is a link for the scoring aspect:
http://www.thonky.com/qr-code-tutorial/data-masking/

Good luck and have fun,
David

On Sat, Oct 26, 2013 at 7:15 AM, kuzes notifications@github.com wrote:

Hello, I'm using the code above but unfortunately sometimes the data
doesn't seem to encoded correctly and i cant scan the produced QRcode. I'm
using arduino mega2560 and this tft http://www.adafruit.com/products/335.
Do you think this problem occurs because of the tft library?
For example: When I try to encode this data:"1879999921" the produced qr
code cant be scanned.
Thank you very much

—
Reply to this email directly or view it on GitHubhttps://github.com//issues/1
.

[budnail]

Hi David,

I've been using your code to learn how to make some basic Version QR codes work for a different display. Before I went on to the higher order codes, I at least wanted to port your code and get it working to prove I understood the details.

I was having a similar problem as Kuzes, and even added the 7 other masks to compare to other online generators but, I was still having issues. I finally tracked the problem to the ECC generator part of your code. Just got it working this morning.

If you make the change below on the line "//change by BN 2/6/14 " I think you'll find that the scanner will get the correct answer most of the time even if only using the mask 0 you implemented. I think what was happening is occasionally there was a zero other than the leading term, so the exponent was being decremented more than it should. This in turn caused the ECC to end before completing all the steps.

//Loop until last exponent in polynomial is zero
while(exponent > 13){
//Discard the lead 0
//Remove zeros & decrement exponent
for(int i = 0; i < 14; i++){
if(Result[0] == 0){//change by BN 2/6/14 was" if(Result[i] == 0)"
zeroFound = true;
exponent -= 1;
}
//discard lead term
if(zeroFound == true){
Result[i] = Result[i+1];
xorResult[i] = Result[i];
}
}

Thanks a lot for sharing your code and the link to http://www.thonky.com/qr-code-tutorial
Seeing how you interpreted the algorithms was very helpful.

BTW
I was able to port your code to a standard Arduino Uno, (something no one seems to think was possible) by being very frugal on the SRAM.
I changed the LogTable to a 2x255 rather than a 4x255 (the index columns were not needed).
I also changed the table to bytes from integers since no value was greater than 255.
Last of all I commented out all the debug lines to get rid of the strings. Using the F() function might worked just as well and keep the debug lines working.
The result is, that I think I'm only using about 1/2 the available SRAM now.
Now I'm off to see if I can implement the grading scheme to choose the correct mask.

Bud