Instructions for Creating and Uploading Samples for
The Big Band Charts Group Manuscript Identification Project
If you would like to upload samples for this project, please follow these instructions.
If you don't follow the instructions, you could mess up the display and links.
1. Verify you have identified the handwriting sample
Please don't guess. Be absolutely certain that you have properly identified the person who wrote the manuscript sample you are identifying.
2. Extract one page from the PDF
To keep from corrupting your original PDF file, first create a copy of that file.
Next, use one of the PDF manipulation programs (list here) to delete all the pages except the page you want to use for your sample. The sample should contain a treble clef and some eighth notes, and any other distinctive characteristics of the writer's hand.
Save that file, then open it in a PDF viewer.
3. Learn the file-naming protocol
It is CRITICAL that you name the files properly, or this system will not work. You will be capturing three images - a thumbnail of a treble clef, a thumbnail of a sequence of eighth notes, and a full page of manuscript.
There are three segments to a properly named file: (a)The type-identifier prefix; (b)The writer's name; and (c) The file type suffix. Here's more information on each.
a. The TYPE-IDENTIFIER PREFIX - this consists of one or two characters plus a dash.
For a MASTER page image, the prefix will be "M-" (in uppercase)
For a TREBLE CLEF thumbnail image, the prefix will be "TC-" (in uppercase)
For an EIGHTH NOTE thumbnail image, the prefix will be "8N-" (that's the numeral eight, and letter N in uppercase)
b. The WRITER'S NAME - The second segment is the writer's name in "one word" with no spaces. Last name first, then the first name. Use a capital letter as the first letter of each name, and lowercase for all the rest. Here are some examples of how a writer's name should be listed in this segment:
WolpeDave
HolcombeBill
TchaikowskyPeterIlych
c. The FILE TYPE SUFFIX - All images must be in JPG format, so the final segment of every file name should be ".jpg" -- four characters: a period and the lowercase letters jpg
Here are some examples of acceptable filenames. Notice they all come in sets of three.
M-WolpeDave.jpg
TC-WolpeDave.jpg
8N-WolpeDave.jpg
M-HolcombeBill.jpg
TC-HolcombeBill.jpg
8N-HolcombeBill.jpg
M-TchaikowskyPeterIlych.jpg
TC-TchaikowskyPeterIlych.jpg
8N-TchaikowskyPeterIlych.jpg
Why do we have to do it this way? Well, if you're really interested, here's the explanation.
All the sample files are kept in the same folder on the server. Whenever anyone wants to look at the list of samples, the code that displays the page first looks in the folder where the samples are kept, and gets a list of all files that start with "M-".
The code then extracts the writer's name from each entry on that list, and creates one row in the displayed table for each name on the list.
With the writer's name isolated, it is easy for the code to add the "TC-" prefix and the "8N-" prefix and build the name of the thumbnails to display in each row of the table, and then to link the writer's name to display the full page sample.
This is why we have to adhere to the file naming protocol.
WARNING! If you do not name your image files correctly, in accordance with this protocol, our system will not be able to read them.
4. Create the two thumbnail images
Using a screen capture program (list here), select and capture a representative treble clef and save it to your hard drive. Then select and capture a sequence of eighth notes and save that to the same place.
For the eighth notes image, try to capture a single eighth note with a flag, in preference to beamed eighth notes, as these are more distinctive.
Rename the files in accordance with the file naming protocol outllined in step 3. above.
5. Create the full page image
Depending on which PDF handling program you have, you might be able to export the single PDF page as a JPG file.
If you have Adobe Acrobat, use File > Save As > Image > JPG
If you have a different PDF handler program, consult your help file or the provider.
If your PDF handler program cannot export a page to JPG format, then rotate the page (if necessary) so the long axis of the page is the same as the long axis of your computer monitor, and use your screen capture program as in step 4. above to select and capture the entire sheet, and save it to your hard drive.
Rename the file in accordance with the file naming protocol in step 3. above.
6. Crop, resize, and downsample the images
Using a graphics/bitmap manipulation program (I recommend Irfan view; download here), crop each image file to cut out any extraneous areas, rotate it if necessary, then resize the image, and finally downsample the image to B&W. Details on each are as follows.
Crop The Image. If you captured too much area in your image, you can crop out the extra by clipping (cropping) the edges.
To do this in IrfanView, first click and drag a rectangle on the image to the approximate area of the part of the image you want to keep. Using your mouse you can drag any of the four edges of the rectangle to refine the keeper area. When you are satisfied, press Ctrl-Y to crop the image to that rectangle.
Rotate The Image (if necessary). If you had to take a sideways picture of your full-page image, you'll need to rotate it back to the upright position.
To do this in IrfanView, simply press the "R" or "L" key as needed to display it right side up.
Resize the Image. The thumbnail images should not be more than 150 pixels on the long edge, and the full-size page image should not be more than 700 pixels on the long edge.
To do this in IrfanView, press Ctrl-R, or use the menu items Image > Resize/Resample, then make sure the "Set new size" radio button is checked, enter 150 (or 700) in the long dimension field (width or height), make sure the units are "pixels," and be sure the "Preserve aspect ratio (proportional)" check box is checked, then click OK.
Downsample the Image. This means reduce the number of colors in the image. Ideally, we want to convert the image to pure black and white, as these are the only colors used. However, in some instances, this will give your samples unacceptable "jaggies" on curves and diagonal lines, so in these cases, use 8-bit shades of grey to allow hinting and dithering.
To do this in IrfanView, use menu item Image > Decrease color depth, and select the 2 Colors option.
I did this with an image of a Treble Clef, sized 92x120, and it went from 4.37k in size to 3.27k - it doesn't seem like a lot, but it's a 25% size savings. For the eighth note image by the same writer, it went from 7.51k to 4.54k. When we have 400 or more images on one page, these savings make a huge difference in page load times. If your original image was scanned in color, the savings will be more like 85% to 95%.
Save Your Changes. Once you have made all the changes, don't forget to save them.
To do this in IrfanView, press the "S" key.
7. Upload the images to the site
Once you have three image files for each identified writer, upload them to this site. There's a possibility we'll just use FTP for this (it's simpler) but more likely I'll create us an upload page where you can select the files on your hard drive then click upload and they will automatically be put into the proper folder on the server.
8. Check your uploads
Finally, go to our Manuscript ID Project page and double check that the files you uploaded are displaying correctly. Check that the Treble Clef Sample and the Eighth Note sample are properly displayed in the samples table, and that the name of the writer, when clicked, properly displays the full-size sample page.
Samples:
This is how your uploaded images should appear on the Samples page if you've done everything right:
| Treble Clef | Eighth Notes | Writer |
|---|---|---|
|
|
Bob Eberhart |
If anything goes wrong and something didn't display properly, contact a site administrator and we'll work with you to fix it.
Please click on any of the links at the top of the page for more information on any of the listed topics.