This is another one of those "note to myself" posts so I can remember how I got through some techincal problems with Autostakkert and Registax.
There are a number of reasons why you'll get artifacts in Registax while using wavelets. Hm...I mean that there are many pathways to artifacts in Registax.
(In the interest of clarity, I should mention that the artifact generally appears in Registax when using wavelets; or it appears in PS when sharpening. Sometimes, the artifacts are so severe that you can see them in the stacked TIF file. If you see the latter, it's not worth bothering with Registax or PS. This means you need to move some alignment points (sometimes you even need to start over) and stack again.)
Here's a list of things that can lead/contribute to artifacts:
1) In Autostakkert, putting alignment points OUTSIDE a planet's edge. This sometimes happens if you let Autostakkert automatically insert alignment points. It can obviously happen due to human error if placing them manually.
2) In Autostakkert(AS), putting alignment points ON the planet's edge. This sometimes is fine with really steady video footage. But when you play your video in the preview window and you see the planet's edge moving a fair bit, that edge alignment point can someties be on the OUTSIDE of the edge. (The lesson here is to move the alignment point INSIDE the edge by a fair bit.)
3) In AS, it's tempting to use a bunch of small alignment points all over. The reason I've done this in the past is that -- in my experience -- AS will produce a relatively more detailed and contrasty stacked file with a bunch of smaller alignments points versus a small number of large alignment points. But you are guaranteed to have a greater chance of artifacts with small alignment points. Again, with really good video footage, you can get away with a lot of small alignment points. With average data (or worse), you are asking for trouble.
4) In AS, there's a "Quality Estimator" (QE) setting that I've discussed in the past. It generates a graph that resembles a seismograph data sheet. I've noticed that when the graph touches the bottom axis for more than a moment (guessing for 3% or more of the total), you may encounter weird ovals in the stacked image. The fix is is to change the QE setting so that the generated graph stays off the bottom axis.***
5) This is also related to the above remark... If your QE graph is really compressed where it looks like a fat line instead of a zig-zag graph, then you may encounter artifacts that you cannot eliminate with ANY alignment point scheme. Usually, there is some QE setting that will "widen" the line so that it resembles a zig-zag.
5) The video footage was shot TOO CLOSE to the edge of the frame. I've had this happen due to laziness OR when trying to dodge a piece of dust in some part of the frame. It gets to be a REAL problem if the planet goes off the edge for any period of time. In the video footage preview window, you can "scrub" through the video by moving a slider. You can also hit "delete" (space bar) to remove any "bad" frames where the planet is too close to the edge or has moved off the frame. The problem is that if you had the planet move off the frame for more than a second, you will need to "spacebar" (remove) A LOT of frames. Sadly, I've spent 20-30 minutes removing frames on just one clip. It's monotonous and sometimes artifacts still appear.
6) Speaking of dust... a dust bunny will get stacked and create its own artifact.
7) Undetected moon on Jupiter surface. You really need to place an alignment point on the moon, but sometimes you can't quite see where the moon is located. AS will think the moon is some surface detail, but it will look distorted.
7) Really NOISY video footage... In an effort to really "push the edge" of the number of frames, I've shot at like 250+ fps with high gain settings. The resulting footage is often noisy and will produce artifacts. Also, if you are shooting through clouds or poor visibilty, you can end up with really noisy video with no details. Noisy images don't inherently produce artifacts, it's just that they are LIKELY to produce artifacts depending on your alignment points.
8) Overexposed image. If you make the mistake of shooting with the histogram pushed to the right, you can often get artifacts. This can easily happen when shooting Venus. It gets brighter and brighter as it rises and you forget to change your settings.
9) Alignment points near a terminator. I've noticed that when setting alignment points near the terminator of Mars, they will produce artifacts that look like partial circles. The solution is to try fewer alignment points with a much larger size.
There are probably tons more scenarios where you can get artifacts, but I'll add to this list as I remember more of them....
So, I'm kinda of getting to the point of usng small and medium alignment points with 8-10 larger alignment points interspersed around the edges. Also, I'm using alignment points sparsely near the terminator of Mars.
There are some "best practices" to list after reflecting on all of this, but I'm too tired to compile them right now.... :(
***Sometimes you can't avoid the graph from hitting the bottom. The fix has been to UNCHECK the "Laplace" option box. But I've found that this will oftentimes result in stacking bad files with good files. Not sure how to remedy this.
No comments:
Post a Comment