ASSETWriterInput for making Video from UIImages on Iphone Issues

Question

I try the following 2 methods of appending UIImages pixelbuffer to ASSETWriterInput. Everything looks good except there's No data in the video file. What's wrong?

1 Adaptor class

AVAssetWriterInputPixelBufferAdaptor * avAdaptor = [AVAssetWriterInputPixelBufferAdaptor assetWriterInputPixelBufferAdaptorWithAssetWriterInput:writerInput sourcePixelBufferAttributes:NULL];

[avAdaptor appendPixelBufferixelBuffer withPresentationTime:CMTimeMake(1, 10)];

2 Making the

// Create sample buffer.
CMSampleBufferRef sampleBuffer = NULL;
result = CMSampleBufferCreateForImageBuffer(kCFAllocatorDef ault, pixelBuffer, true, NULL, NULL, videoInfo, &timing, &sampleBuffer);

// Ship out the frame.
NSParameterAssert(CMSampleBufferDataIsReady(sample Buffer));
NSParameterAssert([writerInput isReadyForMoreMediaData]);
BOOL success = [writerInput appendSampleBuffer:sampleBuffer];

Answer 1

I found that for some reason I needed to append the buffer more than once. The timing in this example from a test app I made might not be proper, but since it works it should give you a good idea.

+ (void)writeImageAsMovie:(UIImage*)image toPath:(NSString*)path size:(CGSize)size duration:(int)duration 
{
    NSError *error = nil;
    AVAssetWriter *videoWriter = [[AVAssetWriter alloc] initWithURL:
                                  [NSURL fileURLWithPath:path] fileType:AVFileTypeQuickTimeMovie
                                                              error:&error];
    NSParameterAssert(videoWriter);

    NSDictionary *videoSettings = [NSDictionary dictionaryWithObjectsAndKeys:
                                   AVVideoCodecH264, AVVideoCodecKey,
                                   [NSNumber numberWithInt:size.width], AVVideoWidthKey,
                                   [NSNumber numberWithInt:size.height], AVVideoHeightKey,
                                   nil];
    AVAssetWriterInput* writerInput = [[AVAssetWriterInput
                                        assetWriterInputWithMediaType:AVMediaTypeVideo
                                        outputSettings:videoSettings] retain];

    AVAssetWriterInputPixelBufferAdaptor *adaptor = [AVAssetWriterInputPixelBufferAdaptor
                                                     assetWriterInputPixelBufferAdaptorWithAssetWriterInput:writerInput
                                                     sourcePixelBufferAttributes:nil];
    NSParameterAssert(writerInput);
    NSParameterAssert([videoWriter canAddInput:writerInput]);
    [videoWriter addInput:writerInput];

    //Start a session:
    [videoWriter startWriting];
    [videoWriter startSessionAtSourceTime:kCMTimeZero];

    //Write samples:
    CVPixelBufferRef buffer = [Utils pixelBufferFromCGImage:image.CGImage size:size];
    [adaptor appendPixelBuffer:buffer withPresentationTime:kCMTimeZero];
    [adaptor appendPixelBuffer:buffer withPresentationTime:CMTimeMake(duration-1, 2)];

    //Finish the session:
    [writerInput markAsFinished];
    [videoWriter endSessionAtSourceTime:CMTimeMake(duration, 2)];
    [videoWriter finishWriting];
}

This method is not required, but is used here as an example of a pixel buffer source:

+ (CVPixelBufferRef) pixelBufferFromCGImage:(CGImageRef)image size:(CGSize)size
{
    NSDictionary *options = [NSDictionary dictionaryWithObjectsAndKeys:
                             [NSNumber numberWithBool:YES], kCVPixelBufferCGImageCompatibilityKey,
                             [NSNumber numberWithBool:YES], kCVPixelBufferCGBitmapContextCompatibilityKey,
                             nil];
    CVPixelBufferRef pxbuffer = NULL;
    CVReturn status = CVPixelBufferCreate(kCFAllocatorDefault, size.width,
                          size.height, kCVPixelFormatType_32ARGB, (CFDictionaryRef) options, 
                          &pxbuffer);
    status=status;//Added to make the stupid compiler not show a stupid warning.
    NSParameterAssert(status == kCVReturnSuccess && pxbuffer != NULL);

    CVPixelBufferLockBaseAddress(pxbuffer, 0);
    void *pxdata = CVPixelBufferGetBaseAddress(pxbuffer);
    NSParameterAssert(pxdata != NULL);

    CGColorSpaceRef rgbColorSpace = CGColorSpaceCreateDeviceRGB();
    CGContextRef context = CGBitmapContextCreate(pxdata, size.width,
                                                 size.height, 8, 4*size.width, rgbColorSpace, 
                                                 kCGImageAlphaNoneSkipFirst);
    NSParameterAssert(context);

    //CGContextTranslateCTM(context, 0, CGImageGetHeight(image));
    //CGContextScaleCTM(context, 1.0, -1.0);//Flip vertically to account for different origin

    CGContextDrawImage(context, CGRectMake(0, 0, CGImageGetWidth(image), 
                                         CGImageGetHeight(image)), image);
    CGColorSpaceRelease(rgbColorSpace);
    CGContextRelease(context);

    CVPixelBufferUnlockBaseAddress(pxbuffer, 0);

    return pxbuffer;
}

Answer 2

I've had a bit of a problem with this code. It geave me a skewed image as a result.

Changing:

CGContextRef context = CGBitmapContextCreate(pxdata,
                                             size.width,
                                             size.height,
                                             8,
                                             4 * size.width,
                                             rgbColorSpace,
                                             kCGImageAlphaNoneSkipFirst);

To:

CGContextRef context = CGBitmapContextCreate(pxdata,
                                             size.width,
                                             size.height,
                                             8,
                                             CVPixelBufferGetBytesPerRow(pxbuffer),
                                             rgbColorSpace,
                                             kCGImageAlphaNoneSkipFirst);

helped.

Answer 3

Hold on, though the answer given by @Peter DeWeese is the direction to follow, the code has two huge issues: firstly, you need to wait while the system is ready to append a new media and secondly, you've got a great memory leak as you need to release your buffer after it was appended to the video writer.

This is true in your very case, but even more in a general case, where you want to loop in multiple frame as follows:

NSInteger i = 0;

for (; i<n; i++) {

     image = allImages[i];

     CVPixelBufferRef buffer = [self pixelBufferFromCGImage:image.CGImage cropFrame:frame];

    // wait for more media data is ready
    while (adaptor.assetWriterInput.readyForMoreMediaData == FALSE) {
        NSDate *maxDate = [NSDate dateWithTimeIntervalSinceNow:0.1];
        [[NSRunLoop currentRunLoop] runUntilDate:maxDate];
    }
    NSLog(@"Panorama: appending frame %ld out of %ld", (long)i, (long)n);

    // Append data
    [adaptor appendPixelBuffer:buffer withPresentationTime:CMTimeMake(i, freq)];//kCMTimeZero];


    // release the buffer
    CVPixelBufferRelease(buffer);

}