Is there any such thing as "too many frames"

[Fordos Moon]

I'm just curious really. When I have been capturing Jupiter with a webcam I tend to get around 3000 frames.

Now is there any point in capturing 4,5, 6000 frames?

I guess there will be a point where the software doesn't like having too many but is there actually an exhaustion point where there is no point capturing more?

[long_arms]

Depending on the frame rate of the camera, too many frames will mean that there will have been significant rotation in Jupiter.

Obviously for the moon it doesn't matter and Saturn the rotation isn't noticeable for a longer period of time.

How long would it take you to get 5 or 6000 frames? Someone can comment on how good de-rotation techniques are...I've never tried it myself.

Dan

[Fordos Moon]

Ah now that's a good point about Jupiter rotating! I reckon it would take me about 90 mins to get 6000 frames so I think I can see where you are coming from re Jupiter!

So if the moon, would 6000 frames be worth the time to improve an image created from 3000 ( so aside from the planet rotation issue)?

[long_arms]

90 minutes ! Do you mean 9? I thought the amount of time before rotation is a problem with Jupiter is around 2 minutes!!

I am not sure with something like the moon, where there's a lot going on in the image (lots of craters, features etc) the difference would be THAT noticeable. Certainly the more frames the better, but after 3-4000 frames I cant see the point. I'm no expert though, just guessing. Someone more knowledgeable may have the answer to that one.

I'll do a test tomorrow, I will be imaging the moon and I'll see if it makes a difference

Dan

[JamesF]

How long you can capture for depends on quite a few different factors such as the speed of rotation of the target and the resolution of the camera and telescope, but for Jupiter there's a program called WinJupos that can sort out a certain amount of rotation to improve the overall image.

In terms of the number of frames then yes there's a point of diminishing returns with the stacking beyond which there may be very little point adding more data, but it's probably better to accept that another few hundred frames isn't going to make much of a difference and capture them anyhow rather than find that you don't have enough data once you get to processing.

I wrote up some calculations for capture times here though these are theoretical limits only. I suspect that for most practical purposes the capture time for Jupiter can be at least doubled. The rest of the planets aren't that much of a problem as they allow plenty of time anyhow.

James

[Fordos Moon]

Thanks Dan, looks like we will have a clear night tomorrow.

I have been setting 10 frames per second in SharpCap and before today I hadn't modified my dob base so was always fighting to get Jupiter on screen.

This is my latest so I'm guessing rotation is what is causing me to not be picking out "features".

[tich]

When I image Jupiter I tend to set up the capturing software at 2 minutes with an SPC at 10 or 15 FPS, apparently according to watching an Episode of the sky at night any more than 2 minutes will show rotation, although I'm sure someone will jump in to correct me. When I use the QHY I have I still set it for 2 minutes although at certain video settings I can get around 50FPS & when I've finished processing it does show up more detail.

When I've used the SPC I can get around 1800 frames to stack when I've used the 2 minute method at 15FPS bearing in mind Registax will stack around 80% unless you change the settings but should be good enough for a decent image.

[Fordos Moon]

James - great info as always, I will check out that WinJupos :-)

[Fordos Moon]

Tich - thank you, interesting stuff as I may look to purchase an ASI, QHY type camera in the future!

[JamesF]

I used to go for about 1800 frames with the SPC900. I think that's a reasonable target with the Xbox cam too. Up to that point I think you're more likely to benefit from the additional frames rather than lose anything by having them.

James

[Fordos Moon]

So I should stop once I get 2000 frames with the Xbox cam to avoid rotation issues when capturing Jupiter?

[JamesF]

So I should stop once I get 2000 frames with the Xbox cam to avoid rotation issues when capturing Jupiter?

Unless you can sustain a much higher frame rate, I think so. My experience with the Xbox cam was that it wouldn't really produce much that was useful at more than 15 fps, so setting a limit of 2000 is not at all unreasonable.

James

[Fordos Moon]

Am I right sticking with 10 fps?

[JamesF]

Am I right sticking with 10 fps?

It's worth experimenting with 15fps, say, but I really struggled to get the camera to work at all nicely above 10fps.

James

[Bizibilder]

When imaging Jupiter I now take .avi's of 10,000 frames - using a fast camera that averages 65fps - so a "run" is around 150seconds (2min 30sec). I suppose this is a little long but it seems to work ok. I've not tried derotation using WinJupos yet - but I intend to give it a try. The main advantage of taking so many frames is that I must have collected a few good ones! Thus I can stack the best 10% of them and still get a decent signal to noise ratio as I'm using 1000 supposedly "good" frames.

If you take 2000 frames with a slower camera it is likely that you will need to stack at least 40-50% of them or suffer from noise problems, you will also be forced to use some "not so good" frames in your stack.

When I did use my SPC900 I took 2000 frames as my "normal" number and stacked 500-1000. You will get image compression with the SPC900 at rates above 10 fps.

It is quite tricky to get the camera to operate at these rates. I can get it to work with very small image "regions of interest" (ie only using part of the whole chip to record my image - something you can't easily do with a webcam) for planets and individual sunspots. I would need to use a much larger "ROI" capture area for lunar mosaics etc and i would probably have to settle for a frame rate around 30-40 fps with the same camera.

[neil phillips]

Hi Guys, its interesting question that i thought a lot about. The main considerations as experianced imagers will know. Are indeed rotation, and noise control, ( assuming here histograms are healthy ) Its not really just a question of more frames. Because of course we can get more frames two ways, one using longer captures ( winjupos ) two using high frame rates. This is where the real questions start, as there has been much discussion already about how long we can shoot using winjupos.

High frame rates and faster exposures do help beat poor seeing, thats a good thing. But of course theres a price to pay namely higher gain and hence noise. But here again another question arises if there will be more frames will they counteract the higher noise. My experiance has been to a degree yes they do. But not totally. I belive its better to try to keep the noise down, rather than try to cure it with extra frames. The seeing beating aspect is another matter, and often faster exposures will indeed produce more sharper frames. If the target is bright, i am thinking jupiter or even more so the moon, then there is more scope for manipulation of all these considerations.

In general if a high frame rate can be got with its faster exposure times, with noise fairly under control. Then it will most often do better than slower frame rates, slower exposures, and less noise.You can not really get away from the terrible effect the atmosphere has on planetary and lunar images. Especially in the uk. While under excellent seeing. slower exposures slower frame rates can do very well. For the most part, for most of the year in the uk. Higher frame rates and faster exposures will produce tigher sharper frames.

Another consideration that shooting longer captures can do. Is enable the imager to cherry pick better frames, saving the first third of the best quality sharpest frames. and binning the two third blurry ones. Can bring benefit to the overall sharpness and detail to the stacked image. I have often done this on lunar hi res. Producing acceptable results under less than ideal seeing.

By cherry picking the thousands of frames captured. And stacking say the top 10% ( this can vary somewhat ) sharpest frames, to beat the uk conditions somewhat.

Of course this can only go so far, if the seeing is terrible. Then even the top 10% of the captured frames will still be under par. Just not as ugly as the other 90% of course.

So its the moderate seeing that benefits the most from this idea, as the fleeting moments of good seeing on the moon will happen. And those 10% of frames from many thousands can be useful.

Some math is required though. If you think you will need 3 or 400 frames to control the noise for any given exposure and gain. Then of course you have to decide how long you will need to capture to aquire that many frames, with binning the worst 90% in mind.

Luckily the moon doesnt cause rotation problems. So its just something you need to consider. I often just take a rough guess. And aquire more than i will likely need. This is for high focal length crater work. Full mosaics may have time constraints. So will have to be thought out a bit more carefully. Jupiter of course is not as friendly. And we do have time contraints in capturing. Even if winjupos is being used.