Rendering time in UF5

[Lennard]

I'm a novice to UF5 (a couple of days). I created a few images that I quite liked and proceeded to render them with the intention of printing. The 1st one I sized to 16" X 12" and saved as a PSD file (4800 X 3600 MB). It did not take very long(minutes), and informed me that it was rendering at several K Pixels/sec.
The 2nd one that I am attempting to render at a similar size is telling me that it is rendering at only appx 180 Pixels/sec(!) and that it it will take appx 27Hours!!! Can someone please explain? The 2 fractals are of different types, is this relevant? 

[bib]

Hello, and welcome to the forum

Of course the rendering time depends on the fractal formula and its parameters. If your second image is much longer the the first one to render, you should also see it in the preview : no surprise. If you can't wait overnight, try using less iterations or a different coloring formula.

...or buy a new PC !

I design fractal on my laptop, but I can't use it to render, because it crashes each time because of overheating! So I render some animations on a very old win98 PC. It can take a week for a 5 minutes animation!

Have fun, and please post your images in the gallery 

bib

[cKleinhuis]

be also sure to set the oversample parameters correctly, if you use 2x oversampling, it will result in 4 times longer rendering,
if you use 4xoversampling you will end up with 16 times longer rendering

[Jameses]

and dont forget to turn subdivisions down too. a fractal that took 2.5 hours to render on non-adaptive settings took 18 hours to render on 2X oversampling with 3x3 subdivisions!

[David Makin]

With programs like UF and ChaosPro etc. where there are a very large number of different formulas then the render times can vary by a degree of around 1000*.

For example a standard single layer z^2+c Mandelbrot will probably take say a few minutes using a disk render at 8000*6000 pixels without any AA, but some fractal formulas are so much slower that an 8000*6000 pixel render even without AA could literally take weeks.

The way to estimate is keep an eye on the render time within UF, these give a pretty good idea of how slow or fast the renders will be when rendered large to disk.

[klixon]

How deep you zoom in also affects rendering time. The deeper you zoom, the more precision is needed for an accurate image (more decimal places are needed).
You can see this happen in the "statistics" window. The precision needed depends on the number of decimals. When you get to "arbitrary" precision, things start to really slow down... And there's no way to get around it, except for shallower zooms

[The Rev]

Has anyone else found that, when rendering animations in UF, that part of the way through the process, the job just fails?  I get a log that says something about something used by another process then just reports job finished without there being any AVI file.  (I would have included the actual log content, but I deleted it).

The Rev

[Sockratease]

Has anyone else found that, when rendering animations in UF, that part of the way through the process, the job just fails?  I get a log that says something about something used by another process then just reports job finished without there being any AVI file.  (I would have included the actual log content, but I deleted it).

The Rev

I don't use Ultrafractal, but as a rule - Never render an avi file from a fractal generator!

Only render image sequences and compile them in a video editor.

The memory overhead for movies is insane, and you will get failures trying to render an avi directly, unless the video is extremely short with no deep zooming.

This general rule holds true for 3D Animations in programs like Carrara and Blender too.

Image sequences are the way to go, and the way the Pros make movies.  One reason I absolutely Hate UF is that when told to render an image sequence, you don't get images until the whole job is done.  The images should render one at a time, then clear the memory used to make them before starting the next image.  But even in UF a render of an image sequence will pick up where it left off after a crash - however a movie file will not.  So right there you should have all the reason you need to avoid rendering movies directly.

There are many choices for compiling image sequences.  If you need some pointers to video editors, just ask!

The short list is : VirtualDub.  It's free and powerful! 

[The Rev]

Thanks alot for the info!  I'm still getting started in all this, so every time someone tells me something, I feel like it's another giant leap forward.

The Rev