This relates to the discussion at http://bugguide.net/node/view/35661].
Admin edit: the original text of this posting was lost in a database problem. The image and reference have been restored from other archives.
Features for identifying yellowjackets
Moderators: rjlittlefield, ChrisR, Chris S., Pau
Features for identifying yellowjackets
This relates to the discussion at http://bugguide.net/node/view/35661].
Admin edit: the original text of this posting was lost in a database problem. The image and reference have been restored from other archives.
-
rjlittlefield
- Site Admin
- Posts: 23224
- Joined: Tue Aug 01, 2006 8:34 am
- Location: Richland, Washington State, USA
- Contact:
Tony, this image does a great job of pulling together diagnostic features that are referenced throughout that key, and putting them all in one place for context. I also find it pleasant to look at -- lots of fine detail wherever my eyes happen to land, and hardly any stacking defects.
It's very hard to avoid some halo around strongly overlapping structures like that antenna versus the head and legs behind it.
Aside... Telecentric lens systems help to some extent by avoiding changes in perspective throughout the stack, but you're still left with the problem that when the background is sharp, it's contaminated by blur from the foreground. That's a very difficult problem to fix. In theory, there's a mathematical process to compute the contamination and subtract it out , but it requires knowing a whole bunch about how the lens handles OOF information and I've never heard of anybody actually pulling it off in practice.
Anyway, I suspect that most of the halo you have here could be fixed up manually without too much trouble.
One trick that I've used with subjects like this is to process the stack in a couple of chunks such that neither chunk has strongly overlapping detail. For example find the farthest forward frame in which the eye appears sharp. Process the stack from that frame backward, and also from that frame forward. The two partial results should line up perfectly because they have the same starting frame. Then put the two together in Photoshop, and paint a mask to reveal what's sharp in each one. Or run the two partial results back through HF, let it make most of the decisions automatically, and manually retouch only what it gets wrong.
Illumination looks good, placement of all the lines and numbers works well for me -- no trouble at all figuring out what goes with what.
It may sound odd, but I'm really looking forward to the yellowjackets this year. Finally, I know where to find the keys to identify what they are!
--Rik
It's very hard to avoid some halo around strongly overlapping structures like that antenna versus the head and legs behind it.
Aside... Telecentric lens systems help to some extent by avoiding changes in perspective throughout the stack, but you're still left with the problem that when the background is sharp, it's contaminated by blur from the foreground. That's a very difficult problem to fix. In theory, there's a mathematical process to compute the contamination and subtract it out , but it requires knowing a whole bunch about how the lens handles OOF information and I've never heard of anybody actually pulling it off in practice.
Anyway, I suspect that most of the halo you have here could be fixed up manually without too much trouble.
One trick that I've used with subjects like this is to process the stack in a couple of chunks such that neither chunk has strongly overlapping detail. For example find the farthest forward frame in which the eye appears sharp. Process the stack from that frame backward, and also from that frame forward. The two partial results should line up perfectly because they have the same starting frame. Then put the two together in Photoshop, and paint a mask to reveal what's sharp in each one. Or run the two partial results back through HF, let it make most of the decisions automatically, and manually retouch only what it gets wrong.
Illumination looks good, placement of all the lines and numbers works well for me -- no trouble at all figuring out what goes with what.
It may sound odd, but I'm really looking forward to the yellowjackets this year. Finally, I know where to find the keys to identify what they are!
--Rik
-
puzzledpaul
- Posts: 414
- Joined: Sun Aug 06, 2006 4:15 am
- Location: UK
- Contact:
<< Tony, this image does a great job of pulling together diagnostic features that are referenced throughout that key, and putting them all in one place for context. I also find it pleasant to look at -- lots of fine detail wherever my eyes happen to land, and hardly any stacking defects. >>
I agree - but on a minor presentational issue, I think I'd have tried to arrange the call outs to be at the same angle (45deg / horiz?), with item Nos in a vert column, maybe on a (small) bit of extra bg down RHS of pic?
No big deal, just would've made it even better - to my eyes
pp
I agree - but on a minor presentational issue, I think I'd have tried to arrange the call outs to be at the same angle (45deg / horiz?), with item Nos in a vert column, maybe on a (small) bit of extra bg down RHS of pic?
No big deal, just would've made it even better - to my eyes
pp
Well I guess I will play the devils advocate here, whatever that may mean but I really like it, the descriptive photograph that is. I found it interesting in knowing what to look for on yellow jackets because I have a bad habit of running into them for one thing and it always seems I find quite a number of them to photograph during the summer and fall, either by macro lens or stereomicroscope. Also I know very little about insects as a whole, to me bugs are bugs and bees, wasps whatever, are that which they are, something that will sting you if you annoy them. Especially if you run over their nests with a lawn mower. So it is nice to know what I am looking at. Especially this gender, species character thing, interesting, didn't know till now. Thanks 
