CVPixelBufferCreate创建出来的buffer每行字节数比指定的宽度要长32个字节

发表于 | 更新于 | 分类于

CVPixelBufferCreate创建出来的buffer每行字节数比指定的宽度要长一32个字节

做一个功能，用一组图片组装一个视频。
最后生成的视频图片是乱码，找了好久原因，发现是CVPixelBufferCreate的理解不够，错误的使用了自己计算的每行像素字节数。

把图片转换成字节的函数

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
static func pixeleBufferFromUIImage(image: UIImage) -> CVPixelBuffer? {
        let width = 600 //Int(image.size.width)
        let height = 600 //Int(image.size.height)
        let options : Dictionary<String, AnyObject>=
                [kCVPixelFormatCGImageCompatibility as String: NSNumber(bool : true),
                 kCVPixelFormatCGBitmapContextCompatibility as String : NSNumber(bool : true)]
        //print("kCVPixelFormatType_32ARGB \(kCVPixelFormatType_32ARGB) _ \(CGImageAlphaInfo.NoneSkipFirst.rawValue)")
        
        var pxBuffer : CVPixelBuffer? = nil
        let status = CVPixelBufferCreate(kCFAllocatorDefault, width, height, kCVPixelFormatType_32ARGB, options, &pxBuffer)
        guard (status == kCVReturnSuccess && pxBuffer != nil ) else {
            return nil
        }
        
        CVPixelBufferLockBaseAddress(pxBuffer!, 0)
        let pxData = CVPixelBufferGetBaseAddress(pxBuffer!)
        
        let colorSpace = CGColorSpaceCreateDeviceRGB()
        print("CVPixelBufferGetBytesPerRow(pxBuffer!)=\(CVPixelBufferGetBytesPerRow(pxBuffer!))")  //结果为CVPixelBufferGetBytesPerRow(pxBuffer!)=2432，不是简单的600 ＊ 4 ＝ 2400
        //把CVPixelBufferGetBytesPerRow(pxBuffer!)换成 （width＊ 4）则最后输出到视频里是乱码
        let context = CGBitmapContextCreate(pxData, width, height, 8, CVPixelBufferGetBytesPerRow(pxBuffer!), colorSpace, CGImageAlphaInfo.NoneSkipFirst.rawValue)
        CGContextClearRect(context, CGRect(x: 0, y: 0, width: width, height: height))
        CGContextDrawImage(context!, CGRect(x: 0, y: 0, width: width, height: height), image.CGImage)
        CVPixelBufferUnlockBaseAddress(pxBuffer!, 0)

		//一个陷阱，计算指定BytesPerRow参数时，因为计算出的值比实际的小。这里你从context中获取出来的图片是正确的。
        return pxBuffer!
    }

把buffer写入video的例子代码。

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
static public func images2Video(lstImages :[UIImage], outputUrl : NSURL) throws -> Bool {
        let videoWriter = try AVAssetWriter(URL: outputUrl, fileType: AVFileTypeMPEG4)
        let outputSetting = [AVVideoCodecKey : AVVideoCodecH264,
                             AVVideoWidthKey : NSNumber(int: 600),
                             AVVideoHeightKey : NSNumber(int: 600)
        ]

        if(!videoWriter.canApplyOutputSettings(outputSetting, forMediaType: AVMediaTypeVideo)){
            print("not support output setting.")
            return false
        }
        
        let videoWriteInput = AVAssetWriterInput(mediaType: AVMediaTypeVideo, outputSettings: outputSetting)
        let sourcePixelBufferAttributesDictionary = [
            kCVPixelBufferPixelFormatTypeKey as String : NSNumber(unsignedInt : kCVPixelFormatType_32ARGB),
            kCVPixelBufferWidthKey as String : NSNumber(int : 600),
            kCVPixelBufferHeightKey as String : NSNumber(int : 600),
        ]
        
        let adapter = AVAssetWriterInputPixelBufferAdaptor(assetWriterInput: videoWriteInput, sourcePixelBufferAttributes: sourcePixelBufferAttributesDictionary)
        videoWriter.addInput(videoWriteInput)
        videoWriter.startWriting()
        videoWriter.startSessionAtSourceTime(kCMTimeZero)
        
        var idxFrame = 0
        var buffer : CVPixelBuffer?
        while idxFrame < lstImages.count {
            if(videoWriteInput.readyForMoreMediaData){
                print("ready for add next pic")
                //每张图片展示0.5s
                let frameTime = CMTime(seconds: 0.5, preferredTimescale: 600)
                let lastTime = CMTime(seconds: Int64(idxFrame * 0.5), timescale: 600)
                var presentTime = CMTimeAdd(lastTime, frameTime)
                if(idxFrame == 0){
                    presentTime = CMTime(value: 0, timescale: 600)
                }
                
                let buffer = pixeleBufferFromUIImage(lstImages[idxFrame])
                ／／这里是直接使用buffer，所以会根据buffer的属性来获取数据，如果上面的函数使用了错误的参数，这里写入video就会是乱码。
                let res = adapter.appendPixelBuffer(buffer!, withPresentationTime: presentTime)
                idxFrame += 1
            }
        }
        videoWriteInput.markAsFinished()
        videoWriter.finishWritingWithCompletionHandler { 
            if(videoWriter.status == .Completed) {
                print("success")
            }else {
                print("error=\(videoWriter.error)")
            }
        }
        return true
    }

参考代码
RRVideoKit
UIImageToVideo