Objective c : fastest method to overlay a video

Question

After recording a video, I want to overlay it with a dynamic uiview which is updated related to the current video frame timestamp. In fact I am trying to do the same thing as Vidometer application. Following the Apple example, I am able to extract the video frame as a buffer and I overlay it with the UIImage of my UIView.

The steps are :

1. Extract video frame:

   CMSampleBufferRef sampleBuffer = [assetReaderVideoOutput copyNextSampleBuffer];
   if (cancelled) {
       completedOrFailed = YES;
       [assetWriterVideoInput markAsFinished];
   }
   CVPixelBufferRef pixelBuffer = CMSampleBufferGetImageBuffer(sampleBuffer);
   

2. Update my UIView's subviews following the frame timestamp:

   mergeTime = CMTimeGetSeconds(CMSampleBufferGetPresentationTimeStamp(sampleBuffer));
   [self updateWidget:mergeTime*1000];
   

3. Get the UIImage of my UIView:

   mSurcoucheImage = [self imageFromSurcouche];
   

   With

   -(UIImage*)imageFromSurcouche{
   CGSize mSize;
   mSize = CGSizeMake(self.mSurcouche.bounds.size.width, self.mSurcouche.bounds.size.height);
   UIGraphicsBeginImageContextWithOptions(mSize, NO, 0.0);
   if (videoOrientation == UIInterfaceOrientationLandscapeRight) {
       CGContextTranslateCTM(UIGraphicsGetCurrentContext(), mSize.width, mSize.height);
       CGContextRotateCTM(UIGraphicsGetCurrentContext(), M_PI);
   }
   [self.mSurcouche.layer renderInContext:UIGraphicsGetCurrentContext()];
   UIImage* image = UIGraphicsGetImageFromCurrentImageContext();
   UIGraphicsEndImageContext();
   return image;
   }
   

4. Apply a filter to keep the alpha of my UIImage:

   CVPixelBufferLockBaseAddress(pixelBuffer, 0);
                           inputImage = [CIImage imageWithCVPixelBuffer:pixelBuffer options:options];
   filter = [CIFilter filterWithName:@"CISourceOverCompositing"];
   [filter setValue:maskImage forKey:kCIInputImageKey];
   [filter setValue:inputImage forKey:kCIInputBackgroundImageKey];
   outputImage = [filter outputImage];
   CVPixelBufferUnlockBaseAddress(pixelBuffer, 0);
   

   With

   colorSpace = CGColorSpaceCreateDeviceRGB();
   options = [NSDictionary dictionaryWithObject:(__bridge id)colorSpace forKey:kCIImageColorSpace];
   

5. Render my UIImage to the video frame buffer

   [ciContext render:outputImage toCVPixelBuffer:pixelBuffer bounds:[outputImage extent] colorSpace:CGColorSpaceCreateDeviceRGB()];
   

   With

   eaglContext = [[EAGLContext alloc] initWithAPI:kEAGLRenderingAPIOpenGLES2];
   ciContext = [CIContext contextWithEAGLContext:eaglContext options:@{kCIContextWorkingColorSpace : [NSNull null]}];
   if (eaglContext != [EAGLContext currentContext])
      [EAGLContext setCurrentContext:eaglContext];
   

I am trying to reach the same performance as Vidometer application which have an overlay duration less than the video duration, but I am far with this method.

Question 1: Is this method the best performing?

Question 2: A also see another method which use AVMutableComposition, but I don't think that I can synchronize my UIView with the video frame timestamps. Can I?