I did a short turn-table animation with my F302, but when I looked at the result in VLC I noticed the animation got a lot brighter than what I saw in blender. Quicktime seems to have the same problem, whereas windows mediaplayer displays the animation too dark…
Please see the attachment for the first frame in each of the programs.
Does anyone know what’s causing this and how this can be solved? Thanks!
What codec did you use?
The output was set to “Avi JPG”, the colour scheme to RGBA.
EDIT: Rendered a couple of frames using “Avi coded” instead of “Avi jpg”, and now the colours are allright again. However, it’s 66 MB for 24 frames now, whereas I first had 6 MB for 100 frames… 
You should remember that MediaPlayer uses the video adjustments from your graphics card. So if you have tweaked those settings, say to watch DVDs or something, those tweaks will be applied to every video you view in Media Player.
Hmm, didn’t know that, but never used WMP before, in fact I only used it to check this particular video 
What codec / stuff do other people use, where the colours stay true but the file size isn’t too huge?
As Atom had said, ‘people’ will tweak their displays etc to suit their tastes and different implementations of codecs are notorious at screwing it up at playback anyway, so the ‘best’ you can do is work to a spec or standard so you know it’s ‘correct’ or just do your own thing, what looks good to you.
But first always render out your final projects to image sequences, Raw AVI or lossless codec. Preferably image sequences rather than final codec straight off. It allows you more flexibility and choice to get it ‘right’.
The reason the difference in your file sizes is the compression you are using.
Then encode the image sequences to a codec to suit your final delivery, if that delivery format changes from say DVD to HD then encode to suit HD, different colourmetrics to DVD. Use something like Virtualdub or AVIdemux to do that.
Without getting too bogged down in it, there are two concerns when encoding apart from compression and those are colourmetrics and luma range. If you don’t get those correct for the delivery format, players will interpret them incorrectly and you video will playback too dark or too light compared to how you saw it.
Good advice all around, especially the recommendation about original rendering format(s). I would go farther and say to avoid Raw AVI & lossless codec choices, and use only image sequence as an initial output format.
The main reason is because any interruption in a rendering session to any video format will require re-rendering the entire file to recover, whereas with image sequence, you can pick up the rendering at the last-rendered frame number. Raw AVI is also a very inefficient format in terms of file size – comparable to rendering individual frame to uncompressed BMP. Lossless codecs don’t really exist afaik – some are much more efficient at compression and affect image quality much less, but in the end, they all remove some data from the original rendering because that’s how video compression works, and that’s what codecs are for.
The most popular image sequence format seems to to PNG, because it uses a lossless but efficient compression method. This is fundamentally different from how codecs compress video data, btw. PNG images show all the data generated by the rendering, but can often compare or exceed file size savings compared to some codecs, depending on the codec settings, but with no image quality loss.
Blender can take image sequences and output movie/video formats to suit many options, as can utilities like Virtual Dub and AVIdemux, as mentioned. Rendering to PNG means that when you go this route, you will always start with the highest-quality images possible, and you can generated videos for whatever needs may crop up without having to re-render the images themselves.
I agree with chipmasque, forget about loseless avi. A short animation of 5 seconds can take Gigabytes of space. Most media players don’t buffer fast enough, or the computer will simply not be able to handle it, and will lag quite often. The result you view will be less good when you would have chosen a compression.
However, there are certain compressions that do compression with no quality loss. The main reason 80% jpeg compression is standard is because below that the first artifacts of compression start to show.
Compression, in the end, is an art of it’s own. I like H.264 compression, but usually it comes down to what you intend to do.
The suggestion was in the context of an intermediate file for encoding out to h264 or whatever after. Not as final delivery and playback.
I doubt this has much of anything to do with the codec being used. It’s probably to do with how the color is handled by both the decoder and the video sink. You need to use a media player that lets you choose between full-range color, 601, and 709 color ranges. (both at the codec level and video sink level)
That plus some media players just don’t handle color correctly at all.
WMP’s default direct show display filters are jacked up. Don’t trust what you see on it unless you install and setup stuff like ffshow + Haali so that the color is correct.
Blender is rendering at full color range.
If I had to guess it looks like VLC is clamping the display to 601 and Quicktime is probably clamping to 709.
