Collectible Card Game Headquarters

http://www.manaleak.com/mtguk/2016/06/s ... n-mystery/

We have 7 repeating token card backs in Eternal Masters. We also have 121 unique word puzzle borders. In many cases the borders were trimmed off-center, so collecting good specimens will be difficult. Personally I don't think it is worth our time to process all of these advertisement cards. Instead I propose that we use a script to generate the borders. In order to do this, we first need a spreadsheet with all of the border text. In some cases a symbol is used. These should be represented in the spreadsheet using [1] for the rose or [2] for the second symbol. I'm attaching a blank sheet. Luckily the complete text is available at the link above. Somebody just needs to split it up based on the printed cards.

As a starting point, I scanned 73 ad cards. There are at least 121 unique cards required to complete the word puzzle, so we need people to help fill the gaps. My scans are available here: https://mega.nz/#F!joNzzCLZ!kV6iIn-3e3xkayW8wv_aqQ

I'm curious if the backs are mated to the borders, or if they are randomized. In other words, are there only 121 cards or are there 7 * 121 = 847 possible combinations? To help solve this, I'd like people to analyze their tokens and report if theirs are the same as mine:

An easy benchmark is the 5/5 Elemental token because there is only one possible back (Show your true colors). On the one I scanned, the left border read "YOU TO CHOOSE BETWEEN LOYALTY TO". The Dack Emblem also has only one possible back (Show your true colors). It's left border read "OF A USURPER QUEEN: MARCHESA, THE". There are only 121.

skibulk wrote:This project will take time to complete. Unfortunately I'm too busy to participate in this task. I can identify and provide the correct font for rendering, once the Spreadsheets are completed. We also need somebody to write a script for this.

If you guys sort out the font, data, and algorithm I can very easily crank up a PHP/GD script to read an excel or CSV file with the texts and impress them on the borders (I will also need to know the offsets for the trimming of the text).

I can make the script very simple, like:
./impress_ema_texts.php <data_file> <images_folder> <output_folder>

Ping me when it's time to do it

I'm happy to host the script so we can just process and upload the blank versions and then you can specify the text or version and it will generate the images as needed (and cached). that way we wouldn't need to keep all 120 versions in the torrent.

Okay now I have a matrix with ~85 cards plugged in and the horizontal text divided up. Some of the text required guesswork. We need people to help fill in the holes with scans. Thanks!

Is that just in photoshop or was that scripted?

The photo matrix was manually constructed using Photoshop. I'm attaching a new spreadsheet. The Blue fields need to be populated with the vertical border text and the Green fields need to be populated with the name of the advertisement artwork. Once we have enough token backs scanned to verify all text, I will upload a higher resolution matrix to assist with data entry. After the spreadsheet is completed, we need a script to generate all 121 token backs. I have uploaded the rendering assets at my Mega link above. (The Rose and Crown Symbols are not yet completed.)

Specs at 600DPI
Font Size: 63.1pt
Cap Height: 44px (Except for J, Q, and some Punctuation)
Symbol Height: 86px
Space between text and symbols: 102px
Border Overlap: 50% of Cap Height

Artwork Options
Chrome Mox
Firebolt
Force of Will
Hydroblast
Jace, the Mind Sculptor
Memory Lapse
Timberwatch Elf

We need a scan of either (#5) or (#6 & #16) to completely verify the text in the puzzle. We also need somebody to assist with Data Entry. Once we have all the data, we can generate all 121 puzzle backs and share them on this forum.

Find the High-Res PDF checklist and spreadsheet data here: https://mega.nz/#F!joNzzCLZ!kV6iIn-3e3xkayW8wv_aqQ

The already-filled blue fields need to be verified. The empty blue fields need to be populated with the vertical border text. The Green fields need to be populated with the name of the advertisement artwork.

You can get the vertical text pre-typed and see the missing backs in low-res here: http://www.manaleak.com/mtguk/2016/06/s ... n-mystery/

Attached a completed spreadsheet, with both vertical and horizontal text filled in and all card images identified.

charlequin wrote:Attached a completed spreadsheet, with both vertical and horizontal text filled in and all card images identified.

Here you can find the first samples generated by the script i made.
It still lacks the automatic generation part, because I cannot understand the excel file to feed the script. But I think the most part is done, it might need some tuning on the font size and offsets. If you have a running PHP environment you can play with it yourself, once someone explains me in details how to feed the generator with the excel data. I still don't understand which layout goes with which text and how to output the result filenames.

Also I had to resize 25% the symbols I received in the assets folder, they need to be at the correct size at 600dpi, the one I was given were way too big. 25% resize just looked fine but I don't know if is the exactly right size.

@kudit: Feel free to host the script, i marked it public domain so it's completely free! I will finish it as soon as I get the missing information.

http://imgur.com/Y2wzR2p
http://imgur.com/Vqw87fi

Sample looks great! I might want to do a little post-processing on it (right now it's sharper than the rest of the card scan) but it's otherwise spot-on.

In terms of reading the spreadsheet, I made a hideous but hopefully useful image:



Essentially: each green cell is a card back, with the text specifying which back image is used; then the four blue cells surrounding it are the four lines of text half-printed on that card's four borders.

If you need, I can add a sheet to the xls later that's just a list of card image in column A and then the four text lines in columns B-E.

I have updated the symbols to the correct sizes in my MEGA folder. The output files should be named Back001.jpg... Back121.jpg.

I think the right and left text is getting cut off slightly too much. Later I will check the font sizes.

For simplicity of parsing the data, feel free to use a CSV version of the XLS (I added one to the rendering assets folder). You might traverse the data like this:

Code: Select all

$csv = array_map('str_getcsv', file('./EMA Puzzle Text.csv'));

$count = 0;
for ($y=0; $y<count($csv)-1; $y+=2){
   for ($x=0; $x<count($csv[$y])-1; $x+=2){
      $count++;
      $name = 'Back' . str_pad($count, 3, '0', STR_PAD_LEFT);
      $top = $csv[$y][$x+1];
      $left = $csv[$y+1][$x];
      $center = $csv[$y+1][$x+1];
      $right = $csv[$y+1][$x+2];
      $bottom = $csv[$y+2][$x+1];
      print("<pre>$name, $center, $top, $left, $right, $bottom</pre><br>\r\n");
   }
}

@charlequin: Do you think it would be worthy to code something in my script so you don't have to post process 121 files which are generated automatically? Maybe something from this link could be helpful?

@skibulk: I think I understood, I'll wrap it up tomorrow!

Nah, it's literally like five minutes to build a Photoshop action to batch-filter the borders if need be. I'll play around with your samples and if I find something good it'll be quick for me to apply to the whole batch once you're done.

Lets move this to PMs. Check your messages JohnnyTH & Charlequin.

skibulk wrote:Lets move this to PMs. Check your messages JohnnyTH & Charlequin.

I liked seeing the discussion