| Author | Message | ||
     Henrique Feio New member Username: Hfeio Post Number: 1 Registered: 02-2005 |
I've read Den's "Three Tone Range Adjustment Procedures" and "Making Complex Tone Range or Density Masks" excellent tutorials. However, I haven't figured out how Den calculates the blur radius. What does Den mean by "25-35% of the pixel width of the largest masked area diameter"? Henrique | ||
| den Unregistered guest |
The concept here is to select a blur such that the ADD portion of the tone-range/density mask edge has a fairly large soft gradient transition into the white area of the mask while the SUBTRACT portion(s)of the tone-range/density mask has a fairly small hard gradient transition, hence the Feather = -2 and Blur = 1. It is difficult to see with the lo-res mask example, but notice the slight graying above the treeline and surrounding the mast and rigging.... This is what you are trying to achieve and it is dependent upon the size of the area being masked. If the tone range being masked includes 1/3 of the image..... Then an ADD Blur = 100 would be suitable.... but in the case of SHADOWs, the image areas here are generally small and scattered throughout the image.... if the largest of these small scattered areas is around a 100 pixels in diameter [you could use the Readout tool but I generally will try a blur amount and if it provides a soft edge both into and out of the mask edge, use it; if not UNDO and adjust the Blur amount and try again].... then a Blur = 25 may be suitable. These hard and soft gradient mask edges are needed to make it easier to prevent auras/halos and posterization.... the ADD and SUBSTRACT curve slopes also play a role here. For this reason, the inverse of a tone-range/density mask using the suggested procedures will not as work well as generating another mask of the inverse tone range. Experiment and vary the suggested numbers and procedures… As not all images have the same type of content.... | ||
| Tony Gamble New member Username: Tonygamble Post Number: 88 Registered: 04-2003 |
Den, I had a good chance to use your Three Tone Range worksheet last night. It worked splendidly and I was surprised how it removed the need for any added sharpening, whereas the original image suffered from the lack of it. I would feel more comfortable about making progress if I understood why you first ADD a mask and then SUBTRACT a variant from it. I think you were attempting to explain the reason above. Would you be kind enough to explain again and I the answer may click, for me, the second time. I also could not understand when you were asked:- " What does Den mean by "25-35% of the pixel width of the largest masked area diameter"?" You replied " If the tone range being masked includes 1/3 of the image..... Then an ADD Blur = 100 would be suitable...." Where does the figure of 100 for the Blur come from? Thanks for all your all hard work. It is a great system. Tony London UK | ||
| den Unregistered guest |
Perhaps the illustration below will aide in understanding the previously posted explanation. Because of the nature of the Image used, the ADD and SUBTRACT curve slopes do not affect the mask transition edge... only blur and feather. The ADD Blur = 100 was arbritary and happens to be the maximum blur that PWP31h offers. If more blur were desired, the 100 amount could just be applied multiple times. Again note, that because of the asymmetry of the mask edge transitions that arise from this methodology, use caution when apply as an 'inverted' mask.  | ||
| Tony Gamble New member Username: Tonygamble Post Number: 89 Registered: 04-2003 |
Thanks Den, Getting clearer Tony | ||
| den Unregistered guest |
Recent experimentation with the mask edge illustrated above has resulted in tighter back-white mask edge transitions without degrading the gradients when the: Feather = -2 & Blur = 1 step is replaced with: Feather = -2 & Feather = 2 ..... Seems to work better especially where foreground tree branches and leaves are being masked in front of a contrasting sky..... I would recommend this as an alternate option to the tone range or denisty masking workflow.... |
