Trying to Improve Print Images
Since our publishers will only publish in black and white, I'm experimenting with some modification of building images in Photoshop. Normally, this white motor disappears into a white background. I'd appreciate your feedback on my attempts to fix that. (Click on the links for larger images.)
Fay
Sample 1
Sample 2
Sample 3
Sample 4
Sample 5
Fay
Sample 1
Sample 2
Sample 3
Sample 4
Sample 5
Comments
As David Purdue remarked in a past post, he outlined his white parts with Mega-POV. However, Mega-POV doesn't work on all computers. Also, Mega-POV sometimes leaves the overall images too dark, in my opinion.
The Photoshop, outline glow of the pictures in Fay's upcoming "NXT ZOO" book are the clearest instructional images I've seen yet. I/we will post a sample later this month.
Rick
Did you try the updated NXT parts?
And there are more options than pov...
Sample POV and new parts.
Sample LDView and new parts.
Sample Google SketchUp.
Google SketchUp Building instructions.
Laurens
--
NXTMonger
Is there documentation for LDView?
I tried the new part and it is somewhat better---more like the first sample.
Laurens, thank you for the Google Sketch up instructions. Does it print building steps and parts steps?
Fay
Fay
http://aivaliotis.com/images/MA_NXTsample.png
There are two things of note:
- Removed all reflections. The reflections add to the confusion.
- Traced edges with black lines.
Another note about the books. I have purchased three titles and notice that the models are never rotated to focus on the build area of concern. The reader is left to guess as to what is going on. I think it's ok to rotate the model.
http://www.nxtguide.davidjperdue.com/resources/NXT_Motor.png
NXTmonger and others, the 4 samples that Laurens posted are created different then the ones from Fay. Laurens used different rendering programs, and Fay only changed the POV renderings afterwards with Photoshop.
Martyn
I just took a look at Laurens links and I like that much better. I like the Google Sketchup images and the LDView image a lot. The main positive point (in addition to absence of reflections) is that the edges are traced by a black line which helps distinguish the contours when smiler colored faces overlap. For example, white on white. That seems the way to go.
The only snag is when you are drawing black objects, the outline is black too so it doesn't help much. If the outline turned white (on black objects) that might help but I'd have to see a real example to judge if it really works.
In general, I think that photorealism is bad in black and white. You actually want an image that looks like it was drawn out.
I don't like the reflection either, but haven't figured out how to eliminate it. (David, thanks for the LPub hint. I'll see if I have the option you mentioned. )
Michael, drawing lines in such a custom manner would be incredibly time-consuming and almost impossible to do well. My experiments with Photoshop are an attempt to help publishers and authors find an acceptable solution that is practical.
Remember, publishers also require very high resolution images---at least 300 dpi. Does anyone know the dpi of images produced by these various programs. POV-Ray/LPub is only acceptable if the image is kept very large. A distance setting of 1000-1500 usually does it.
Just choose LDview for renderer in LPub (instead of Pov or LDGlite). More details here.
Default LDview preferences may need to be tweaked, see here.
Also have a look at this thread.
Philo
If you can add the number of holes in the straight connector and length to the axles next to each of these components, that will be very useful.
http://lrobotikas.net/instrucciones/joystick/barcos1-00.html
Koldo
http;//lrobotikas.net
I've worked with David, trying to get my computer to surrender to his settings, but Mega-POV rendered my computer useless. If I'm not mistaken, he used LPub 3 beta, but that gave me fits as well. (My husband would blame this all on Bill Gates.)
Geekytom,
The effect used in #4 was actually Torn Paper, with the original image overlayed and set to (I think) 80% transparency. (I don't have my notes at the moment.)