kicad-developers team mailing list archive

[Craig Southeren <craigs@xxxxxxxxxxxxxxxxx>]

On 8/10/2012 9:20 PM, Vesa Solonen wrote:
> 08.10.2012 02:20, Craig Southeren kirjoitti:
>
> > I'm hoping to have patches against the latest Kicad code available this
> > week. I'd love some feedback !
>
> I'm waiting eagerly to see it working! I was also expecting some 
> public code somewhere or some status updates on the list to see if you 
> have found something else to do ;)

Sorry - I've been busy and I had to redo the user interface as per some 
suggestions from the Kicad developers.

I need to find the time to re-baseline the code against the latest 
Bazaar repository, and then recompile it, and then somehow get these changes
published. So much work to do, so little time to do it in.....

> > I've implemented bitmap scaling, but the same factor applies to both X
> > and Y.
> >
> > Do you know why Scott wants independent scaling?
>
> One reason for independent scaling is imaging equipment limitations. 
> For example a (lens + CCD) scanner has some depth of field and angle 
> of view creates a scaling error for far away objects. In the case of a 
> scanner the error is 1D as sensor is moved linearly in the other 
> direction. That means aspect ratio of the output is not exactly the 
> same as original.

These kinds of non-linear distortions are far better handled by a 
dedicated graphics program. The ability to do bi-cubic scaling and 
fractional degree rotation,
and a nice user interface for it, is already in GIMP/PhotoShop/other.

I have no desire to re-implement this code in Kicad.

> For ease of use I suggest just specifying the image size in both 
> coordinates, no matter the number of pixels. That gives the freedom of 
> working with odd sized crops and any scaling is done only once. No 
> need to even know the exact aspect ratio during image preparation as 
> it's only needed in the use phase.
>
> Lens correction and rotation is definitely something to do with other 
> tools, but view size scaling is a job for viewer application.

At the moment, there is the ability to specify the X and Y offset of the 
image relative to the pcb origin, and the DPI (actually, pixels per inch).
This assumes the bitmap has already been ortho-rectified, which I feel 
is reasonable.

When the image is first selected, the user is presented with a display 
of the bitmap under a scaled graticule that allows them to accurately 
calibrate the DPI
against known marks on the bitmap. For example, you can adjust the DPI 
so that the 0.1" pins on ICs are exactly the correct distance apart.

Craig


--

-----------------------------------------------------------------------
 Craig Southeren          Post Increment – VoIP Consulting and Software
 craigs@xxxxxxxxxxxxxxxxxxxx                   www.postincrement.com.au

 Phone:  +61 243654666      ICQ: #86852844
 Fax:    +61 243656905      MSN: craig_southeren@xxxxxxxxxxx
 Mobile: +61 417231046      Jabber: craigs@xxxxxxxxxx

 "Science is the poetry of reality." Richard Dawkins