hello hanyi if you dont mind can I ask what setting you use for your gif size and colouring? My gifs end up looking very grainy and with tumblrs size limit it's been such a trouble posting gifs here. thank you
hello! I usually use the full 540px in width (268px if I want the set to have two gifs in a row, but I rarely do this anyway), and the height usually varies depending on the post (e.g. the layout I have in mind, number of gifs in total in the set), from between 270px and 480 or 500px.
I sometimes struggle with the limit too but I generally try to keep them around 8mb (10mb max). bigger gifs hit the limit easier, but reducing the number of frames should help counter that! I’ve heard somewhere before that even with the 10mb limit, gifs above a certain size (3mb?? 4mb??) get automatically reduced when they are uploaded... I don’t know how true this is, sometimes my gifs look of a lower quality when uploaded, but sometimes they look fine ρ(´-＿-｀●)
using 1080p videos is the very big first step in avoiding graininess. I usually load the video through vapoursynth to crop, resize, and sharpen first instead of importing it directly into photoshop. I personally feel that this helps reduce graininess??? although when I tested both ways to reply this ask, the differences were really, really small. maybe I picked the wrong scene to demonstrate this... 《ﾟДﾟ》maybe vapoursynth is more useful when you’re working with lower quality videos and/ or want to use the denoising filter as well
resulting gif with no photoshop sharpening/ filtering/ colouring
in comparison: video imported straight to photoshop for cropping and resizing
after importing the vapoursynth clip into photoshop, I do my usual smart sharpening and minor denoising...
if your particular scene is too grainy, you can always increase the denoising, but I usually keep it around 10% because higher levels will interfere with my gif sharpness (T⌓T)
for clip imported directly into photoshop (notice the higher sharpening radius to compensate and reach a similar degree of sharpening as the vapoursynth clip)
then, colouring! colouring is probably an important factor in helping reduce graininess, or at least “patches” of colour. I don’t use pre-set photoshop filters so for every gif I make, depending on the lighting and end result I want, the colouring layers vary a lot. the settings below are specific only to this particular gif .+(´^ω^`)+.
I usually start with brightness/ contrast
then curves - I usually do an S curve looking thing, but this really depends on the lighting of the scene
then levels. I didn’t touch levels for this particular one
but take note to not overly brighten a dark gif or overly darken a bright gif or this would worsen the graininess
vibrance because I love stronger colours
my next step after vibrance is usually selective colour, but before that, I use hue/ saturation and turn up the saturation to determine what are the colours that I should be working on in selective colour. from here, I can determine that to recolour his face, I should be recolouring the reds, magentas, and whites, and to recolour the background, cyans and blues
and then selective colouring! one of my favourite colouring steps hehe. for this particular scene I ended up recolouring every single colour except yellow and green (i.e. I made changes to red, magenta, cyan, blue, white, neutral, black) which were mostly absent in this scene
way back before I had this habit of checking the colours first, I used to just assume which colours I should change, so this is what would’ve happened had I just recoloured the reds on selective colouring and hue/ saturation and left out the magentas, whites, and neutrals... lots of uneven red patches أ‿أ
anyway, after selective colouring, my last colouring step would be to make final adjustments on hue/ saturation. so here, I changed the magenta hue to something slightly redder (to remove that magenta tinge at his cheek), and reduced the saturation and lightness of the cyans and blues to remove the blue light in the background
finally, when saving as web:
adaptive + pattern: I usually use pattern because diffusion sometimes gives these huge, nasty, and quite obvious patches, like below
adaptive + diffusion: see his cheek... it looks like there are three distinct diagonal strips of colour
selective + pattern: also works too, depending on the colours you want to bring out
and back to my earlier point on vapoursynth vs non-vapoursynth, when I compared between these two final gifs... I realised that the difference is really small. I personally prefer vapoursynth because smart sharpening on photoshop tends to leave this white outline thingamajig which becomes more obvious and makes the image look less realistic the higher your radius goes (;Д;)
not sure if this helps or if I answered your question otl... it’s really a fair bit of trial and error and tinkering around with your settings and layers! (◕ㅅ◕✿)
izzy806
pixelskylines
jacobvanloon
areudirty
andulkaphoto
ellievsbear
dededos
tangentwithtanner
