Hi there,
I am using the writer to save my render window to an avi file. I set the frame rate to 25 fps (the same as the video in my render window) but when I go to open the avi file that has been created, it plays and something like 4x the speed. Does anyone know what is going wrong here? Thanks.
You are recording lesser frames per second, because you are compressing the files. There a few tricks you can do here, slow down your entire patch when recording, aka non real time rendering, or use a software screen recorder.
Hi there,
I am using the writer to save my render window to an avi file. I set the frame rate to 25 fps (the same as the video in my render window) but when I go to open the avi file that has been created, it plays and something like 4x the speed. Does anyone know what is going wrong here? Thanks.
How do I perform non-real time recordin?
The node Writer (EX9.Texture NRT), U’ll get a image sequence that can be converted in a video file.
that seems quite cumbersome. Is that a bug?
typically the hdd is the botleneck, when saving large amounts of images – your fps correspond to the speed of your hdd and image conversion (which might be higher when converting the images to an avi). so, it’s rather a feature than a bug ;)
if you use the NRT writer you can see how fast your machine saves the images, which can be quite frustrating.
so:
_ screen recorders like fraps are good for fast savings
_ NonRealTime rendering is best for high quality (save double image size and reduce later)
p.s. this is only my interpretation
Also try the NRTwriter Screenshot module on my user page, its much faster than the normal writer node, I’m sure there must be something that could be done to speed that up!
well, since the sharedmem nodes came out i always wondered if it’s not possible to pump the RAM full of NRT output (should be quite fast) until you’re finished (should be quite short, e.g. if you’re recording loops etc only a couple secs in realtime), then purge the RAM content as an image sequence to the HD. with the amounts of RAM available nowadays that may be a feasible approach…
alternatively (even better) the RAM could be used as a slowly filling buffer of sorts, constantly writing down to HD as fast as possible, so that the HD bottleneck only becomes noticeable when/if your RAM is full…
Or use very fast HDs (Tried with a RAID5, saves HD in near realtime) or SSDs.
Another Possibility is to record the DVI Output on a second Computer with a Framegrabber card (http://www.unigraf.fi/?page=64 or http://www.ems-imaging.com/catalog/modules.php?name=catalog&file=product_info&cPath=39&products_id=116&osCsid=08c412584c6cb2b27779eb6e368626c4). But that’s the most expensive way to do it.
But in almost all cases I experienced, the Writer NRT worked quite well. And I prefer an image sequence over a compressed avi File anytime, because of quality and alpha channel.