I am confused by the inaccuracy of plotting where PrintString is printing.

I was originally using a NewPage(850, 1100) setting for printing my checks, but that grid was too coarse. It caused my signature graphic to look rough. I increased the NewPage values to get a finer grid, in order to provide a better quality signature. That improves the image quality, but causes the X Y coordinates to become increasingly inaccurate the finer I make the grid.

I set up the code with variables that will multiply all of the coordinates by the same factor, which I thought should maintain the correct ratios for the X Y coordinate placement of the printed data. The larger the factor, the shorter the distance from the top and left edges of the paper the data is placed. (I hope I am explaining this clearly.)

This is the code I am using. What am I overlooking here?

var vCheckDate as String = @Rt("00" + @Str(@Month(Check Date)), 2) + "-" + @Rt("00" + @Str(@DOM(Check Date)), 2) + "-" + @Year(Check Date)
var vAmount as String = @Decimals(Amount, 2)
var vFont as String = "Courier New"	//Font Face
var vSize as Int			//Font Size
var vDPI as Int = 100			//Dots Per Inch

vSize = .16 * vDPI

//Format is Inches * DPI
NewPage(8.50 * vDPI, 11.00 * vDPI)
PrintPagePaper(28)

//Check data
PrintString(Check No, 7.60 * vDPI, .50 * vDPI, 0, vFont, vSize, 0)
PrintString(vCheckDate, 7.10 * vDPI, .85 * vDPI, 0, vFont, vSize, 0)
PrintString(Name, 1.10 * vDPI, 1.35 * vDPI, 0, vFont, vSize, 0)
PrintString(vAmount, 7.00 * vDPI, 1.35 * vDPI, 0, vFont, vSize, 0)
PrintString(Dollars, .40 * vDPI, 1.70 * vDPI, 0, vFont, vSize, 0)
PrintString(Memo, .70 * vDPI, 2.85 * vDPI, 0, vFont, vSize, 0)

//PrintImage(file, xpos, ypos, width, height)
PrintImage(".\Data\Sig.png", 5.40 * vDPI, 2.55 * vDPI, 2.80 * vDPI, .60 * vDPI)
 

Hi Bob,

I multiply each value by the vDPI factor so that I can change the dots per inch setting once, and have all of the coordinates scale to match it.

Regarding Alec's article: I just finished reading it, and that's when I added the PrintPagePaper command to the code. This problem has been bugging me for a while, and I thought that maybe this command would somehow help the situation. When I did not, I finally became motivated to find out why scaling the values for a finer grid causes everthing to shrink closer to the starting edges of the paper.

Mark,

Yes, 850 x 1100 is what I want to use right now, as it is equal to the original values when I first programmed the code. So currently this is the setting that gets my data to print where it needs to. The only reason I wanted to change the grid resolution was because my signature graphic was too choppy-looking at this setting. I added the vDPI variable was so I could experiment with different resolutions by changing just one number.

The problem is when I use higher resolutions, the printing locations of the text moves further to the left and higher. But if the X Y coordinates are being scaled up by the SAME factor as the page dimensions, why wouldn't the end result be that the text prints to the same relative position on the page?

Carl,

I see two things that may be affecting you.

1. You need a space before the font name to indicate a normal (not bold or italic) font.
" Courier New"
This can have an effect, in that, if Sesame can't find a matching font name, substitution may occur, which can affect text measurement.

2. At the ratio shown in your code, your font is scaled to be only 1/16" tall. This is very tiny. Courier New is not a fully scalable font. This means that the OS must substitute the closest available size it can find. Again, this can affect measurement.   

Try it with a space in front of the font name and a larger font size and see what happens. You may also see a difference by using a TrueType font, such as Lucida Console.

Quote:


Yes, OK. 

That's what I get for trying to do math without mechanical assistance. *sigh*

What printer are you using?

Carl,

After playing with this on a number of different printers, we think that you are just seeing the effect of an accumulation of factors. The factors include things like:
1. The printer is mechanical. It will not pull/feed two sheets of paper exactly the same way.
2. On a inkjet a "dot" tends to be considerably taller than it is wide, rather than being square. Inkjets generally have much higher horizontal resolution than vertical.
3. Your particular printer is really 600dpi. It uses "color layering" to get higher resolutions for color.
4. Font descriptors have their own rules about which fonts can scale to what sizes and how well they can do any particular size.
5. On MS, scaling is done with integer math. This can introduce "inaccuracies" due to truncation which can accumulate.
6. Your printer dirver has settings like "Best" and "Economode" which can drastically affect how it renders any given dot.
7. Your math may cause a dot to land at a place that is not mechanically accessible by the printer, especially if NewPage is used to create a higher set of divisions than the printer can do. The printer can only print on certain mechanical divsions.   

Etc...

Bottom line, you can't always work the math with printers based on "it should theoretically work". You should be able to use the math to get it basically into place, then adjust as needed. If your printer produces the results you prefer at a scaling factor of 100, then that's the one you should use. To improve the quality of your signature at this resolution, what you want to do is scan the signature at a high resolution, then scale it down for print. Images tend to get "rough" when they are scaled up from their natural size.

Problem solved!

I modified the code. It now produces VERY PRECISE results. I can simply measure where I want the text to print, plug in the coordinates, and be sure that it will print where I want.

I added two more variables to the code (vX & vY) that deal with the aspect ratio differences of the X Y coordinates. Now, if I enter 7 to indicate 7" on the X axis, I get the text starting at EXACTLY 7" from the left edge of the paper.

I also provided a calcluation so that I can enter the font size expressed in points, that matches the true point sizes specified in MS Word and other Windows programs.