Creating animated gifs with less jitter?

Question

I'm looking to improve the quality of exported animations (for which I'm limited to animated gifs for compatibility reasons), and the default export looks a bit unnatural/jittery. I'm wondering if there's a way to improve this. Potential ideas:

1. Simple motion blur
2. Temporal antialiasing

step = 10;
genPlot[i_] := 
  Graphics[{Circle[{0, 0}, 1], Rotate[Line[{{-1, 0}, {1, 0}}], i]}];
images = Table[genPlot[angle], {angle, 0, Pi - Pi/step, Pi/step}];
Export["circle.gif", images, "AnimationRepetitions" -> Infinity, 
 "DisplayDurations" -> 1/50]

Animation frames can be rendered at any time granularity, but gif frame delays are multiple of 10ms, so 25/50 FPS is possible, but 30/60 is not. Most displays nowadays seem to have refresh rate of 60 Hertz, so using 1/50 delay you would expect a "freeze-frame" every 5 frames. On other hand, this video claims to be 50fps and looks smooth on my 60 Hz screen, so perhaps frame rate mismatch is not a significant contributor to jitter

Edit

simple blending makes it a bit better, but the line looks gray instead of black and the sector boundary looks clearly visible. Also 50x oversampling would be pretty expensive to compute for more general animations (ie, like this one)

radius = 50;
images2 = 
  Table[genPlot[angle], {angle, 0, Pi - Pi/(step*radius), 
    Pi/(step*radius)}];
images3 = Blend /@ Partition[images2, radius];
Export["circle2.gif", images3, "AnimationRepetitions" -> Infinity, 
 "DisplayDurations" -> 1/50]

Answer 1

I ended up using Blend combined with a helper function below that creates weights from Sin curve to average several frames together.

IE, weights to smear each image over 5 frames, where each frame is a subsampled 3 times, the weights from genWeights[5, 3] look like this

(* generate sin plot centered at offset, 
 with given width . offset - width/2 .. offset + width/2 *)
makePlot[offset_, width_] := (
   (Cos[(2 [Pi] (-offset + x))/width] + 1)/2
   );
(* generates weights for line spanning "spread" frames, each frame 

using "subsample" frames *)
genWeights[spread_, subsample_] := (
   Assert[OddQ[spread]];
   Assert[OddQ[subsample]];
width = (spreadsubsample - 1) + 2;
   middleFrame = (width - 2)/2;
   expr = makePlot[middleFrame, width];
   expr /. x -> # & /@ Range[0, spreadsubsample - 1] // N
   );
spread = 5;
subsample = 3;
vals = genWeights[spread, subsample];
xvals = Transpose[{Range[Length[vals]], vals}]
ListPlot[Partition[xvals, subsample], InterpolationOrder -> 0, 
 Filling -> Axis, FillingStyle -> Thickness[.05], 
 PlotStyle -> PointSize[Large], 
 PlotLabel -> 
  StringForm["spread=, subsample=", spread, subsample]]