How to get text in a CATextLayer to be clear - text

I've made a CALayer with an added CATextLayer and the text comes out blurry. In the docs, they talk about "sub-pixel antialiasing", but that doesn't mean much to me. Anyone have a code snippet that makes a CATextLayer with a bit of text that is clear?
Here's the text from Apple's documentation:
Note: CATextLayer disables sub-pixel antialiasing when rendering text.
Text can only be drawn using sub-pixel antialiasing when it is
composited into an existing opaque background at the same time that
it's rasterized. There is no way to draw subpixel-antialiased text by
itself, whether into an image or a layer, separately in advance of
having the background pixels to weave the text pixels into. Setting
the opacity property of the layer to YES does not change the rendering
mode.
The second sentence implies that one can get good looking text if one composites it into an existing opaque background at the same time that it's rasterized. That's great, but how do I composite it and how do you give it an opaque background and how do you rasterize it?
The code they use in their example of a Kiosk Menu is as such: (It's OS X, not iOS, but I assume it works!)
NSInteger i;
for (i=0;i<[names count];i++) {
CATextLayer *menuItemLayer=[CATextLayer layer];
menuItemLayer.string=[self.names objectAtIndex:i];
menuItemLayer.font=#"Lucida-Grande";
menuItemLayer.fontSize=fontSize;
menuItemLayer.foregroundColor=whiteColor;
[menuItemLayer addConstraint:[CAConstraint
constraintWithAttribute:kCAConstraintMaxY
relativeTo:#"superlayer"
attribute:kCAConstraintMaxY
offset:-(i*height+spacing+initialOffset)]];
[menuItemLayer addConstraint:[CAConstraint
constraintWithAttribute:kCAConstraintMidX
relativeTo:#"superlayer"
attribute:kCAConstraintMidX]];
[self.menuLayer addSublayer:menuItemLayer];
} // end of for loop
Thanks!
EDIT: Adding the code that I actually used that resulted in blurry text. It's from a related question I posted about adding a UILabel rather than a CATextLayer but getting a black box instead. http://stackoverflow.com/questions/3818676/adding-a-uilabels-layer-to-a-calayer-and-it-just-shows-black-box
CATextLayer* upperOperator = [[CATextLayer alloc] init];
CGColorSpaceRef space = CGColorSpaceCreateDeviceRGB();
CGFloat components1[4] = {1.0, 1.0, 1.0, 1.0};
CGColorRef almostWhite = CGColorCreate(space,components1);
CGFloat components2[4] = {0.0, 0.0, 0.0, 1.0};
CGColorRef almostBlack = CGColorCreate(space,components2);
CGColorSpaceRelease(space);
upperOperator.string = [NSString stringWithFormat:#"13"];
upperOperator.bounds = CGRectMake(0, 0, 100, 50);
upperOperator.foregroundColor = almostBlack;
upperOperator.backgroundColor = almostWhite;
upperOperator.position = CGPointMake(50.0, 25.0);
upperOperator.font = #"Helvetica-Bold";
upperOperator.fontSize = 48.0f;
upperOperator.borderColor = [UIColor redColor].CGColor;
upperOperator.borderWidth = 1;
upperOperator.alignmentMode = kCAAlignmentCenter;
[card addSublayer:upperOperator];
[upperOperator release];
CGColorRelease(almostWhite);
CGColorRelease(almostBlack);
EDIT 2: See my answer below for how this got solved. sbg.

Short answer — You need to set the contents scaling:
textLayer.contentsScale = [[UIScreen mainScreen] scale];
A while ago I learned that when you have custom drawing code, you have to check for the retina display and scale your graphics accordingly. UIKit takes care of most of this, including font scaling.
Not so with CATextLayer.
My blurriness came from having a .zPosition that was not zero, that is, I had a transform applied to my parent layer. By setting this to zero, the blurriness went away, and was replaced by serious pixelation.
After searching high and low, I found that you can set .contentsScale for a CATextLayer and you can set it to [[UIScreen mainScreen] scale] to match the screen resolution. (I assume this works for non-retina, but I haven't checked - too tired)
After including this for my CATextLayer the text became crisp. Note - it's not necessary for the parent layer.
And the blurriness? It comes back when you're rotating in 3D, but you don't notice it because the text starts out clear and while it's in motion, you can't tell.
Problem solved!

Swift
Set the text layer to use the same scale as the screen.
textLayer.contentsScale = UIScreen.main.scale
Before:
After:

Before setting shouldRasterize, you should:
set the rasterizationScale of the base layer you are going to rasterize
set the contentsScale property of any CATextLayers and possibly other types of layers(it never hurts to do it)
If you don't do #1, then the retina version of sub layers will look blurry, even for normal CALayers.
- (void)viewDidLoad {
[super viewDidLoad];
CALayer *mainLayer = [[self view] layer];
[mainLayer setRasterizationScale:[[UIScreen mainScreen] scale]];
CATextLayer *messageLayer = [CATextLayer layer];
[messageLayer setForegroundColor:[[UIColor blackColor] CGColor]];
[messageLayer setContentsScale:[[UIScreen mainScreen] scale]];
[messageLayer setFrame:CGRectMake(50, 170, 250, 50)];
[messageLayer setString:(id)#"asdfasd"];
[mainLayer addSublayer:messageLayer];
[mainLayer setShouldRasterize:YES];
}

First off I wanted to point out that you've tagged your question with iOS, but constraint managers are only available on OSX, so I'm not sure how you're getting this to work unless you've been able to link against it in the simulator somehow. On the device, this functionality is not available.
Next, I'll just point out that I create CATextLayers often and never have the blurring problem you're referring to so I know it can be done. In a nutshell this blurring occurs because you are not positioning your layer on the whole pixel. Remember that when you set the position of a layer, you use a float values for the x and y. If those values have numbers after the decimal, the layer will not be positioned on the whole pixel and will therefore give this blurring effect--the degree of which depending upon the actual values. To test this, just create a CATextLayer and explicitly add it to the layer tree ensuring that your position parameter is on a whole pixel. For example:
CATextLayer *textLayer = [CATextLayer layer];
[textLayer setBounds:CGRectMake(0.0f, 0.0f, 200.0f, 30.0f)];
[textLayer setPosition:CGPointMake(200.0f, 100.0f)];
[textLayer setString:#"Hello World!"];
[[self menuLayer] addSublayer:textLayer];
If your text is still blurry, then there is something else wrong. Blurred text on your text layer is an artifact of incorrectly written code and not an intended capability of the layer. When adding your layers to a parent layer, you can just coerce the x and y values to the nearest whole pixel and it should solve your blurring problem.

You should do 2 things, the first was mentioned above:
Extend CATextLayer and set the opaque and contentsScale properties to properly support retina display, then render with anti aliasing enabled for text.
+ (TextActionLayer*) layer
{
TextActionLayer *layer = [[TextActionLayer alloc] init];
layer.opaque = TRUE;
CGFloat scale = [[UIScreen mainScreen] scale];
layer.contentsScale = scale;
return [layer autorelease];
}
// Render after enabling with anti aliasing for text
- (void)drawInContext:(CGContextRef)ctx
{
CGRect bounds = self.bounds;
CGContextSetFillColorWithColor(ctx, self.backgroundColor);
CGContextFillRect(ctx, bounds);
CGContextSetShouldSmoothFonts(ctx, TRUE);
[super drawInContext:ctx];
}

If you came searching here for a similar issue for a CATextLayer in OSX, after much wall head banging, I got the sharp clear text by doing:
text_layer.contentsScale = self.window!.backingScaleFactor
(I also set the views background layer contentsScale to be the same).

This is faster and easier: you just need to set contentsScale
CATextLayer *label = [[CATextLayer alloc] init];
[label setFontSize:15];
[label setContentsScale:[[UIScreen mainScreen] scale]];
[label setFrame:CGRectMake(0, 0, 50, 50)];
[label setString:#"test"];
[label setAlignmentMode:kCAAlignmentCenter];
[label setBackgroundColor:[[UIColor clearColor] CGColor]];
[label setForegroundColor:[[UIColor blackColor] CGColor]];
[self addSublayer:label];

Related

Sprite effect on SKShapeNode

I am trying to get make an effect with sprite kit on Mac on SKShapeNodes ( a normal Node would be fine too ) to like something like this:
http://firewall.com.pl/wp-content/uploads/2013/05/mailstore-cloud-edition-en.png
a circular node surrounded by a rather glow effect, not fully opaque. Any ideas? I also want the "glow" to get different colors. The only idea right now would be an round, white image as png with "fade out" intensity on the edges until it is completely transparent. Then cover that with an blend factor. But I wonder, if there is an simpler way to do this.
If any of you have an good idea, I would be very grateful.
Regards
Thomas
That looks like more of a lens flare. but here is something that may get you started: try this code then fill with black, of course you will have to adjust the glow and circle to fit your needs.
-(void) CreateGlowingCircles {
SKShapeNode *ball = [[SKShapeNode alloc] init];
CGMutablePathRef myPath = CGPathCreateMutable();
CGPathAddArc(myPath, NULL, 0,0, 20, 0, M_PI*2, YES);
ball.path = myPath;
ball.lineWidth =0.1;
ball.glowWidth = 15.5;//adjust for more glow effect
//add fill and stock for the black dot inside the glow
ball.position = CGPointMake(200, 200);
[self addChild:ball];
}
This is a piece of code from a example project i did...
from there you could add a filter! I hope this gets you started. for the lines that connect the black spots you could figure out some advanced physics magic. look up SKPhysicsJointFixed etc.. good luck!

UICollectionView layout, cells spaced far out

I have a UICollectionView that displays images. My images are size 114x133 but for some reason the actual UICollectionVewCell is bigger than this (I can see this because I set the cell background to red). Because of this the view can display only 2 images and there is a huge gap between the two. I am implementing the following
- (CGSize)collectionView:(UICollectionView *)collectionView layout:(UICollectionViewLayout*)collectionViewLayout sizeForItemAtIndexPath:(NSIndexPath *)indexPath {
UIImage* img = [UIImage imageNamed:#"thumbnail_frame.png"];
return img.size;
}
- (UIEdgeInsets)collectionView:
(UICollectionView *)collectionView layout:(UICollectionViewLayout*)collectionViewLayout insetForSectionAtIndex:(NSInteger)section {
return UIEdgeInsetsMake(1, 1, 1, 1);
}
I am creating the cell's contentview programmatically (i.e init imageView etc).
I face the same problem in a different view where I am creating the whole view programmatically. There I do:
UICollectionViewFlowLayout *flowLayout = [[UICollectionViewFlowLayout alloc] init];
flowLayout.minimumInteritemSpacing=10;
flowLayout.minimumLineSpacing= 30;
flowLayout.sectionInset = UIEdgeInsetsMake(5, 5, 5, 5);
if(collectionView != nil) {
[collectionView removeFromSuperview];
}
CGRect frm= CGRectMake(45, 80, 250, 230);
collectionView = [[UICollectionView alloc] initWithFrame:frm collectionViewLayout:flowLayout];//(0, 0, self.view.frame.size.width, self.view.frame.size.height) collectionViewLayout:flowLayout];
collectionView.backgroundColor = [UIColor colorWithRed:0 green:0 blue:0 alpha:0.5];
collectionView.clipsToBounds = YES;
collectionView.layer.cornerRadius = 10.0;
collectionView.dataSource = self;
collectionView.delegate = self;
[collectionView registerClass:[YourPicCollectionViewCell class] forCellWithReuseIdentifier:#"YourPicCell"];
Again there is either a big gap between two images, or if I make the width smaller, there are 3 images but the right-most one is half-visible.
I tinkered with spacing but it doesn't seem to work. What could be the problem?
I think instead of resizing the cell.. you should try resizing the collection view. Make sure that the collection view has just enough width to hold 2 items in a row. This is because if the collection view has more width than the cells, then then will stick to the extreme right and left. If it's too short, then only one cell will be displayed at the centre (Assuming that you are trying to keep 2 cells in a row).
The reason that one of your cell was half visible is that the collection view had more width than its super view.
So doing the following would help you get what you want..
Frame your collection view such that its less than its super view and should be equal to (numberOfCells X width) + ((numberOfCells-1) X cellGap).
If it's a custom collection view cell try adding constraints to the imageView such that the image sticks to the cells edges.

UItextview will disregard line height when NSFontAttributeName is set

Well, Ive searched in several places and although some people allegedly have found fixes it doesn't seem to apply to my case.
I'm trying to procedurally set the line height of a few UItextviews like this :
UITextView *lab = [LocalTexts objectAtIndex:j];
NSMutableParagraphStyle *paragraphStyle = [[NSMutableParagraphStyle alloc] init];
paragraphStyle.lineHeightMultiple = 50.0f;
paragraphStyle.maximumLineHeight = 50.0f;
paragraphStyle.minimumLineHeight = 50.0f;
NSString *string = lab.text;
NSDictionary *ats = #{
NSFontAttributeName : [UIFont fontWithName:#"DIN Medium" size:16.0f],
NSParagraphStyleAttributeName : paragraphStyle,
};
lab.attributedText = [[NSAttributedString alloc] initWithString:string attributes:ats];
Strange thing is that if I disable the NSFontAttributeName assignment, the line height will work, also, if I set the Paragraph style to have a certain paragraph height, that always works too, so the NSParagraphStyleAttribute IS NOT being fully ignored. I dont know if it is a bug or I'm actually doing something wrong.
I tried implementing it as pure CORE TEXT, but it is a bit too complex for the current scope of the project.
Hope someone can point me in the right direction. Thanks.
This is a known bug in NSHTMLWriter which is used by UITextView to convert your attributed string to HTML: http://www.cocoanetics.com/2012/12/radar-uitextview-ignores-minimummaximum-line-height-in-attributed-string/
You can use UITextView replacement we have in DTCoreText to render this text correctly: https://github.com/Cocoanetics/DTCoreText

Merging a previosly rotated by gesture UIImageView with another one. WYS is not WYG

I'm getting crazy while trying to merge two UIImageView.
The situation:
a background UIImageView
(userPhotoImageView)
an overlayed
UIImageView (productPhotoImageView)
that can be streched, pinched and
rotated
I call my function on the UIImages but I can take coords and new stretched size from the UIImageView containing them (they are synthesized in my class)
- (UIImage *)mergeImage:(UIImage *)bottomImg withImage:(UIImage *)topImg;
Maybe simplest way would be rendering the layer of the top UIImageView in a the new CGContextRef like this:
[bottomImg drawInRect:CGRectMake(0, 0, bottomImg.size.width, bottomImg.size.height)];
[productPhotoImageView.layer renderInContext:ctx];
But in this way I loose the rotation effect previosly applied by gestures.
A second way would be trying to apply AffineTransformation to UIImage to reproduce GestureEffects and then draw it in the context like this:
UIImage * scaledTopImg = [topImg imageByScalingProportionallyToSize:productPhotoView.frame.size];
UIImage * rotatedScaledTopImg = [scaledTopImg imageRotatedByDegrees:ANGLE];
[rotatedScaledTopImg drawAtPoint:CGPointMake(productPhotoView.frame.origin.x, productPhotoView.frame.origin.y)];
The problem of this second approach is that I'm not able to exactly get the final rotation degrees (the ANGLE parameter that should be filled in the code above) amount since the user started to interact, this because the RotationGesture is reset to 0 after applying so the next callback is a delta from the current rotation.
For sure the most easy way would be the first one, freezing the two UIImageViews as they are displayed in that very moment but actually I still didn't find anyway to do it.
Ok basically there is another workaround for all this crazy merging stuff, but it's definitively not an elegant solution. For avoid handling any kind of AffineTransformation just capture the ImageScreen and then crop it.
CGImageRef screenImage = UIGetScreenImage();
CGRect fullRect = [[UIScreen mainScreen] applicationFrame];
CGImageRef saveCGImage = CGImageCreateWithImageInRect(screenImage, fullRect);
CGRect cropRect = CGRectMake(x,y,width,height);
CGImageRef saveCGImage = CGImageCreateWithImageInRect(screenImage, cropRect);
Told you it wasnt elegant, but for someone it could be useful.
Great that was helpful and so too much workaroundy ;)
so since i see it's pretty hard to find around some code examples to merge pics after a manipulation here goes mine, hope it can be helpful:
- (UIImage *)mergeImage:(UIImage *)bottomImg withImage:(UIImage *)topImg {
UIImage * scaledTopImg = [topImg imageByScalingProportionallyToSize:productPhotoView.frame.size];
UIGraphicsBeginImageContext(scaledTopImg.size);
CGContextRef ctx = UIGraphicsGetCurrentContext();
CGContextTranslateCTM(ctx, scaledTopImg.size.width * 0.5f, scaledTopImg.size.height * 0.5f);
CGFloat angle = atan2(productPhotoView.transform.b, productPhotoView.transform.a);
CGContextRotateCTM(ctx, angle);
[scaledTopImg drawInRect:CGRectMake(- scaledTopImg.size.width * 0.5f, -(scaledTopImg.size.height * 0.5f), scaledTopImg.size.width, scaledTopImg.size.height)];
UIImage *newImage = UIGraphicsGetImageFromCurrentImageContext();
UIGraphicsEndImageContext();
UIGraphicsBeginImageContext(bottomImg.size);
[bottomImg drawInRect:CGRectMake(0, 0, bottomImg.size.width, bottomImg.size.height)];
[newImage drawInRect:CGRectMake(productPhotoView.frame.origin.x, productPhotoView.frame.origin.y, newImage.size.width, newImage.size.height)];
UIImage *newImage2 = UIGraphicsGetImageFromCurrentImageContext();
UIGraphicsEndImageContext();
return newImage2;
}
Since I got a background image it was easier first create a context just to apply tranformation to the overlayview, than save it and write on top of the bottom layer.
Take care of the CTM translation before rotating the overlayview or it will rotate around an axis placed at {0,0} while I want to rotate my image around its center.
Regards
I have the same problem.
I only have problem with rotation.
I use gesture recognizers for move, scale and rotation.
At the beginning when I save the image tha scale was right, the position was right but never apply rotation to the saved image.
Now it works
if you want to know the rotation angle I get it with:
CGFloat angle = atan2(overlay.transform.b, overlay.transform.a);
And rotation with:
CGContextRotateCTM(context, angle);
If you have a different approach please let me know.

Resize CATextLayer to fit text on iOS

All my research so far seems to indicate it is not possible to do this accurately. The only two options available to me at the outset were:
a) Using a Layout manager for the CATextLayer - not available on iOS as of 4.0
b) Use sizeWithFont:constrainedToSize:lineBreakMode: and adjust the frame of the CATextLayer according to the size returned here.
Option (b), being the simplest approach, should work. After all, it works perfectly with UILabels. But when I applied the same frame calculation to CATextLayer, the frame was always turning out to be a bit bigger than expected or needed.
As it turns out, the line-spacing in CATextLayers and UILabels (for the same font and size) is different. As a result, sizeWithFont (whose line-spacing calculations would match with that of UILabels) does not return the expected size for CATextLayers.
This is further proven by printing the same text using a UILabel, as against a CATextLayer and comparing the results. The text in the first line overlaps perfectly (it being the same font), but the line-spacing in CATextLayer is just a little shorter than in UILabel. (Sorry I can't upload a screenshot right now as the ones I already have contain confidential data, and I presently don't have the time to make a sample project to get clean screenshots. I'll upload them later for posterity, when I have the time)
This is a weird difference, but I thought it would be possible to adjust the spacing in the CATextLayer by specifying the appropriate attribute for the NSAttributedString I use there, but that does not seem to be the case. Looking into CFStringAttributes.h I can't find a single attribute that could be related to line-spacing.
Bottomline:
So it seems like it's not possible to use CATextLayer on iOS in a scenario where the layer is required to fit to its text. Am I right on this or am I missing something?
P.S:
The reason I wanted to use CATextLayer and NSAttributedString's is because the string to be displayed is to be colored differently at different points. I guess I'd just have to go back to drawing the strings by hand as always....of course there's always the option of hacking the results from sizeWithFont to get the proper line-height.
Abusing the 'code' tags a little to make the post more readable.
I'm not able to tag the post with 'CATextLayer' - surprisingly no such tags exist at the moment. If someone with enough reputation bumps into this post, please tag it accordingly.
Try this:
- (CGFloat)boundingHeightForWidth:(CGFloat)inWidth withAttributedString:(NSAttributedString *)attributedString {
CTFramesetterRef framesetter = CTFramesetterCreateWithAttributedString( (CFMutableAttributedStringRef) attributedString);
CGSize suggestedSize = CTFramesetterSuggestFrameSizeWithConstraints(framesetter, CFRangeMake(0, 0), NULL, CGSizeMake(inWidth, CGFLOAT_MAX), NULL);
CFRelease(framesetter);
return suggestedSize.height;
}
You'll have to convert your NSString to NSAttributedString. In-case of CATextLayer, you can use following CATextLayer subclass method:
- (NSAttributedString *)attributedString {
// If string is an attributed string
if ([self.string isKindOfClass:[NSAttributedString class]]) {
return self.string;
}
// Collect required parameters, and construct an attributed string
NSString *string = self.string;
CGColorRef color = self.foregroundColor;
CTFontRef theFont = self.font;
CTTextAlignment alignment;
if ([self.alignmentMode isEqualToString:kCAAlignmentLeft]) {
alignment = kCTLeftTextAlignment;
} else if ([self.alignmentMode isEqualToString:kCAAlignmentRight]) {
alignment = kCTRightTextAlignment;
} else if ([self.alignmentMode isEqualToString:kCAAlignmentCenter]) {
alignment = kCTCenterTextAlignment;
} else if ([self.alignmentMode isEqualToString:kCAAlignmentJustified]) {
alignment = kCTJustifiedTextAlignment;
} else if ([self.alignmentMode isEqualToString:kCAAlignmentNatural]) {
alignment = kCTNaturalTextAlignment;
}
// Process the information to get an attributed string
CFMutableAttributedStringRef attrString = CFAttributedStringCreateMutable(kCFAllocatorDefault, 0);
if (string != nil)
CFAttributedStringReplaceString (attrString, CFRangeMake(0, 0), (CFStringRef)string);
CFAttributedStringSetAttribute(attrString, CFRangeMake(0, CFAttributedStringGetLength(attrString)), kCTForegroundColorAttributeName, color);
CFAttributedStringSetAttribute(attrString, CFRangeMake(0, CFAttributedStringGetLength(attrString)), kCTFontAttributeName, theFont);
CTParagraphStyleSetting settings[] = {kCTParagraphStyleSpecifierAlignment, sizeof(alignment), &alignment};
CTParagraphStyleRef paragraphStyle = CTParagraphStyleCreate(settings, sizeof(settings) / sizeof(settings[0]));
CFAttributedStringSetAttribute(attrString, CFRangeMake(0, CFAttributedStringGetLength(attrString)), kCTParagraphStyleAttributeName, paragraphStyle);
CFRelease(paragraphStyle);
NSMutableAttributedString *ret = (NSMutableAttributedString *)attrString;
return [ret autorelease];
}
HTH.
I have a much easier solution, that may or may not work for you.
If you aren't doing anything special with the CATextLayer that you can't do a UILabel, instead make a CALayer and add the layer of the UILabel to the CALayer
UILabel*label = [[UILabel alloc]init];
//Do Stuff to label
CALayer *layer = [CALayer layer];
//Set Size/Position
[layer addSublayer:label.layer];
//Do more stuff to layer
With LabelKit you don't need CATextLayer anymore. No more wrong line spacing and wider characters, all is drawn in the same way as UILabel does, while still animated.
This page gave me enough to create a simple centered horizontally CATextLayer : http://lists.apple.com/archives/quartz-dev/2008/Aug/msg00016.html
- (void)drawInContext:(CGContextRef)ctx {
CGFloat height, fontSize;
height = self.bounds.size.height;
fontSize = self.fontSize;
CGContextSaveGState(ctx);
CGContextTranslateCTM(ctx, 0.0, (fontSize-height)/2.0 * -1.0);
[super drawInContext:ctx];
CGContextRestoreGState(ctx);
}

Resources