如何手动发布CMSampleBuffer

Question

这段代码导致内存泄漏和应用程序崩溃：

    var outputSamples = [Float]()

    assetReader.startReading()
    while assetReader.status == .reading {
        let trackOutput = assetReader.outputs.first!

        if let sampleBuffer = trackOutput.copyNextSampleBuffer(),
            let blockBuffer = CMSampleBufferGetDataBuffer(sampleBuffer) {
            let blockBufferLength = CMBlockBufferGetDataLength(blockBuffer)
            let sampleLength = CMSampleBufferGetNumSamples(sampleBuffer) * channelCount(from: assetReader)
            var data = Data(capacity: blockBufferLength)
            data.withUnsafeMutableBytes { (blockSamples: UnsafeMutablePointer<Int16>) in
                CMBlockBufferCopyDataBytes(blockBuffer, atOffset: 0, dataLength: blockBufferLength, destination: blockSamples)
                CMSampleBufferInvalidate(sampleBuffer)

                let processedSamples = process(blockSamples,
                                               ofLength: sampleLength,
                                               from: assetReader,
                                               downsampledTo: targetSampleCount)
                outputSamples += processedSamples
            }
        }
    }
    var paddedSamples = [Float](repeating: silenceDbThreshold, count: targetSampleCount)
    paddedSamples.replaceSubrange(0..<min(targetSampleCount, outputSamples.count), with: outputSamples)

这要归功于copyNextSampleBuffer()和创建规则。

反过来，我们不能在Swift中使用CFRelease()。链接到目标C唯一规则的原因是我所无法理解的。

有没有一种在Swift中手动发布CMSampleBuffer的方法？

Answer 1

这并不是一个真正的解决方案，因为手动释放内存似乎是不可能的，并且在读取不安全的可变字节时，结合使用when循环和assetReader将不会释放内存。

这个问题通过一个解决办法得到了解决:在将音频文件公开给while循环之前，将其转换为CAF格式。

缺点:热秒时间越长，音频文件越长--时间越长。

优点:它只使用极少量的内存，这首先是问题所在。

灵感来自：https://stackoverflow.com/users/2907715/carpsen90 从音频文件中提取仪表电平的答案

Answer 2

最近，我使用一个自动发布池解决了一个类似的问题。

尝试包装在自动释放池中使用sampleBuffer的区域。就像这样：

var outputSamples = [Float]()

assetReader.startReading()
while assetReader.status == .reading {
    let trackOutput = assetReader.outputs.first!

    autoreleasepool {
        if let sampleBuffer = trackOutput.copyNextSampleBuffer(),
            let blockBuffer = CMSampleBufferGetDataBuffer(sampleBuffer) {
            let blockBufferLength = CMBlockBufferGetDataLength(blockBuffer)
            let sampleLength = CMSampleBufferGetNumSamples(sampleBuffer) * channelCount(from: assetReader)
            var data = Data(capacity: blockBufferLength)
            data.withUnsafeMutableBytes { (blockSamples: UnsafeMutablePointer<Int16>) in
                CMBlockBufferCopyDataBytes(blockBuffer, atOffset: 0, dataLength: blockBufferLength, destination: blockSamples)
                CMSampleBufferInvalidate(sampleBuffer)

                let processedSamples = process(blockSamples,
                                               ofLength: sampleLength,
                                               from: assetReader,
                                               downsampledTo: targetSampleCount)
                outputSamples += processedSamples
            }
        }
    }
}
var paddedSamples = [Float](repeating: silenceDbThreshold, count: targetSampleCount)
paddedSamples.replaceSubrange(0..<min(targetSampleCount, outputSamples.count), with: outputSamples)

如果我正确理解，一旦它超出了autoreleasepool的范围，sampleBuffer就会被释放。