How do I find out which direction a user is panning with UIPanGestureRecognizer? - ios

Question

So I am using UIPanGestureRecognizer in my project which I added to a view. I would like to know when a user either goes up, down,left or right. I am using the left and right feature to scrub through video. The up and down gesture is still to be determined. I have used the following code but I can't seem to figure it out. Thanks for the help!
#IBAction func panVideo(_ recognizer: UIPanGestureRecognizer) {
let vel = recognizer.velocity(in: self.videoView)
if vel.x > 0 {
// user dragged towards the right
print("right")
}
else {
// user dragged towards the left
print("left")
}
}
EDIT: Using Slider
if let duration = avPlayer?.currentItem?.duration {
let totalSeconds = CMTimeGetSeconds(duration)
let value = Float64(scrubberSlider.value) * totalSeconds
let seekTime = CMTime(value: Int64(value), timescale: 1)
avPlayer?.seek(to: seekTime, completionHandler: { (completedSeek) in
//perhaps do something later here
})
}

Answer 1

Joe's answer is close, but it won't take into account direct vertical or horizontal pans. (I'd comment on his answer except the formatting won't take.) Try this:
let vel = recognizer.velocity(in: self.videoView)
if vel.x > 0 {
// user dragged towards the right
print("right")
}
else if vel.x < 0 {
// user dragged towards the left
print("left")
}
if vel.y > 0 {
// user dragged towards the down
print("down")
}
else vel.y < 0 {
// user dragged towards the up
print("up")
In essence, you are getting the CGPoint of the gesture (x,y) and determining the velocity of the movement. You have an alternative to this - taking the starting and ending point:
var startingPoint = CGPoint.zero
#IBAction func panVideo(_ recognizer: UIPanGestureRecognizer) {
if recognizer.state == .began {
startingPoint = recognizer.location(in: self.videoView)
}
if recognizer.state == .ended {
let endingPoint = recognizer.location(in: self.videoView)
[ do the same comparing as above]
}
}
The advantage of the second option is you aren't doing unnecessary calculations during the pan. The disadvantage is that there are certain scenarios (like animating view movements) that are not conducive to it.
EDIT: I'm adding a bit more verbiage after reading your comment. It sounds to me that you may not be fully understanding what a pan gesture really is.
Like most (all?) gestures, it has a beginning, an in-between, and and end.
It is a two-dimensional drag with two components, both x and y.
There are actually SEVEN possible states, but FOUR of them (cancelled, failed, possible, recognized) do not happen with a pan gesture, leaving THREE states (began, changed, ended) that trigger.
I threw out one example - moving a view with a pan gesture - earlier. Now I'll try a second one - tracing an outline of, say, the Statue of Liberty in an image.
Here you want all THREE states, in order to know when to being tracing, when the path changes, and when it ends. And restricting this to the change state, I think you can see where both the X and the Y coordinate changes.
So yes, a logging of "left, up, left, up, left" is quite possible.I would think that if you traced a completely vertical line across the entire screen you might expect all "up" or "down" values in your log, but the odds of any human being panning that perfect is unlikely, so sure, a few "left" or "rights" may happen.
My tweak to Joe's code was to eliminate those moments of perfection. If vel.x == 0 you would have "left", and where bel.y == 0 you would have "down".
Again, if you simply want to know what the "result" of the pan is, use .began and .ended and ignore .changed - do not use recognizer.velocity but recognizer.state.
The "if" statements both of us gave you are really frameworks. If you understand both state and the two-dimensional nature of things, and you need to use .changed, then adapt those "if" statements - maybe compare the velocity of X to Y and take the greater, or eliminate those changes where the change in X or Y was under a threshold.

Answer 2

Try this code: tested in Swift 3.
Updated Answer: Below code will give you a starting and end location of your view when touch began.
if recognizer.state == .began {
let vel = recognizer.velocity(in: view) // view is your UIView
if vel.x > 0 {
print("right")
} else {
print("left")
}
}
if recognizer.state == .ended {
let vel = recognizer.velocity(in: view)
if vel.y > 0 {
print("down")
} else {
print("up")
}
}
Note : Your answer actually hidden in your code ?
#IBAction func panVideo(_ recognizer: UIPanGestureRecognizer) {
let vel = recognizer.velocity(in: self.videoView)
if vel.x > 0 {
// user dragged towards the right
print("right")
}
else {
// user dragged towards the left
print("left")
}
if vel.y > 0 {
// user dragged towards the down
print("down")
}
else {
// user dragged towards the up
print("up")
}
}
hope this helps...

Answer 3

Okay, now I'm getting the correct mental picture. You want scrub control. This is something very different, and I would recommend a UISlider over working with gestures - highly recommend it. For starters, they have the pan gesture already built in! Here's what I think apps like YouTube, QuickTime, etc. do.
(1) Let's take a specific example of having a video that is 1:53:22 in length, or (1*60*60)+(53*60)+22 = 6802 seconds in length.
(2) Add a "scrubber" subview to your main screen. You'll probably want a UISlider, two UILabels (one to each side of the slider), and anything else you think for a polished look.
(3) The UISLider will have a minimumValue of 0 seconds and a maximumValue of 6802 seconds. Of course, you'll want that max value to be calculated on each change of source.
(4) A question you'll want to answer for your app is whether to go the route of iTunes (where this scrubber view is always visible) or YouTube (where it is overly visible when the user or mouse cursor hovers over an area). For the former, you just need to position this scrub view in a position on the screen. For the latter though, you may wish to use a pan gesture - but only for visibility. Hold that thought.
(5a) You need two, maybe three more things on you UISlider. First is an automatic value update. Again it will depend on the visibility of the entire scrub view. You want to update, once a second, both the left hand UILabel and the UISLider value if it's always visible. For a disappearing one you probably can get away with only updating it once a second when it's visible.
(5b) The second thing you need to do with the UISlider is track changes the user makes to it (the "scrubbing") while it's visible. The event you are looking for is UIControl.valueChanged(). It will trigger anytime the user works with the slider, giving you the new seconds value to "scrub" the video to.
(5c) The third thing you might want to do with the UISlider is customize it a few ways - change the thumb image and the slider itself. My app changes the thumb image. These can only be done in code, there are no IB properties available.
Back to #4. All of the above doesn't need a pan gesture, unless you want the scrub view to appear only when needed.
If you have a mental picture of what I've described above, all you want to know is if a pan gesture has happened. No regards for direction. You might wish to have some regards for screen area - do want this scrub view to appear when a user pans over an area where the scrub view will not appear?
Wire up a CALayer (or the entire video view) with the pan gesture. Then code for a state of UIGestureRecognizer.began. Make the scrub view visible by changing it's alpha state from 0 to 1, or "sliding" it into view by changing it's origin or height. Add a UIView.animate(withDuration:) to it for a good effect.
Now, all that's left is setting the scrub view back to it's natural state. You'll need to code the reverse of whatever you did, and attach it to a timer set for however many seconds you want it visible.
TL;DR;
My app uses 4 UISliders that change various things (height, width, saturation, grill thickness) of of a photo effect that uses CoreImage. Performance is very tight, about 5/100 of a second to grab the new values of all 4 sliders and update the image.
These sliders are always visible today, but my next update (about 2 weeks away) will feature a "sliding control board" - think a keyboard with sliders and other controls on it. (There's limitations on the alpha value for a custom keyboard that forced me to write my own, but that's a separate discussion.)
So I know a "sliding scrub view" is possible. What I don't know for you is if you set the alpha value to a view to zero, will it detect pan gestures? I don't know, thus a CALayer may be needed.