ruanjx / videolab Goto Github PK

添加图片 11 长的时候，内存暴涨到 1 个 G

edit video with linear speed is easy, but now, I want to realize curve or ramp speed such as accord from Bezier path. Any help will appreciated!

delete pls

I have a RenderComposition with a video RenderLayer. I want to mix an audio file with this video.
I tried to create another RenderLayer with an AVAsset created from an AAC file and add it to my RenderComposition. After that, I could hear two audio streams from my video and audio files together, but the image disappeared. I got a black screen instead of my video.
Also, the length of the resulting video equals the length of the AAC file, even though I set a shorter CMTimeRange.

So, how do I accomplish this? Is this supported by VideoLab? If not, will it be supported and how can I work around this problem now?

can i hire u for video editing app

Contact me pleaee live:.cid.83f046ffe2932e9d

This is a general callout for input / advice / support on adding better Transitions to the Framework.

I've reviewed the demo project and see the limited options for Transitions available. I would like to expand on this and implement full support for Transitions but am having trouble finding the best approach for this. A lack of technical documentation also makes this more challenging.

My initial thought was to add a TransitionLayer which would hold any RenderLayers which you'd like the Transition applied on. But this felt too restrictive. Ex: What if you want different Transitions for each cut within your composition? I think this is a deadend idea but happy to hear other ways it might work at this level.

My next idea / approach was to add the Transition as a property directly to RenderLayer. The Transition should be applied / rendered when the current RenderLayer's content finishes playback. This felt like the best approach but curious if I'm overlooking any downsides.

The biggest issue is that even adding these classes, I haven't been able to figure out how to correctly apply the Filter effects to the composition. The LayerCompositor class seemed like the obvious place but, again without any technical docs, it's all remaining a bit cryptic for me at the moment.

I would love if I could get some input from @ruanjx here or anyone else who's used this extensively.

在逆向分享那里有写到使用 IOSurface 生成纹理性能更优，这个是为什么呢？为什么不使用CVMetalTextureCache呢？

How could I use it in macOS development ?

背景：
我看到了RenderLayer可以设置transform，来进行移动、旋转、缩放。

let renderLayer1 = RenderLayer(timeRange: timeRange, source: source)
var transform = Transform(center: center, rotation: 0, scale: 0.5)
renderLayer1.transform = transform

期望：
新增裁剪画面区域，比如：

et renderLayer1 = RenderLayer(timeRange: timeRange, source: source)
var transform = Transform(center: center, rotation: 0, scale: 0.5)
renderLayer1.transform = transform

/// ------
renderLayer1.cropFrame = CGRect(x: 0, y: 0, width: 300, height: 300) //< 新增裁剪方法

Error:

Fatal error: Could not create render pipeline state for vertex:oneInputVertex, fragment:lookupFragment, error:Error Domain=CompilerError Code=2 "reading from a rendertarget is not supported" UserInfo={NSLocalizedDescription=reading from a rendertarget is not supported}

What is wrong?

当有旋转效果时,会出现边缘锯齿效果

 func transition2Demo() -> VideoLab {
        // 1.1 LayerGroup1
        var timeRange = CMTimeRange(start: CMTime.zero, duration: CMTime(seconds: 5, preferredTimescale: 600))
        let layerGroup1 = RenderLayerGroup(timeRange: timeRange)

        // Add sub-renderLayer1
        var image = UIImage(named: "image1.JPG")
        var imageSource = ImageSource(cgImage: image?.cgImage)
        imageSource.selectedTimeRange = CMTimeRange(start: CMTime.zero, duration: timeRange.duration)
        timeRange = imageSource.selectedTimeRange
        let renderLayer1 = RenderLayer(timeRange: timeRange, source: imageSource)
        
        var center = CGPoint(x: 0.5, y: 0.5)
        // 添加旋转
	let rotation = GLKMathDegreesToRadians(15)
	var transform = Transform(center: center, rotation: rotation, scale: 0.15)
        renderLayer1.transform = transform
        
        // Add sub-renderLayer2
        var url = Bundle.main.url(forResource: "video1", withExtension: "MOV")
        var asset = AVAsset(url: url!)
        var source = AVAssetSource(asset: asset)
        source.selectedTimeRange = CMTimeRange(start: CMTime.zero, duration: timeRange.duration)
        timeRange = source.selectedTimeRange
        let renderLayer2 = RenderLayer(timeRange: timeRange, source: source)
        
        center = CGPoint(x: 0.25, y: 0.25)
        transform = Transform(center: center, rotation: rotation, scale: 0.5)
        renderLayer2.transform = transform
        
       .........
        
        return videoLab
    }

Hey, love this library!

How would you go about adding multiple text layers (for example captions) to the composition?

The composition can only take a single animationLayer value, so no place to add an array.

My thoughts are to create a base CALayer, then add all of the subsequent text animation layers to that layer, adjust their animation (from and duration values), and then just assign that base layer to composition.

It hasn't worked for me yet, but is that the right approach at all?

Thanks!

Hi @ruanjx ,

First of all thank you for your effort, this is awesome framework that make work with AVFoundation easier.

• I read documents
• I read open or closed issues

Secondly I have a question about muting original audio of AVAsset and add custom audio to RenderComposition? I already tried to mute videoLayer with audioConfiguration like this

var audioConfiguration = AudioConfiguration()
let volumeRampTimeRange = source.selectedTimeRange
let volumeRamp1 = VolumeRamp(startVolume: 0.0, endVolume: 0.0, timeRange: volumeRampTimeRange)
audioConfiguration.volumeRamps = [volumeRamp1]
videoLayer.audioConfiguration = audioConfiguration

and then added new renderLayer for custom audio (mp3 file) named audioLayer and passed it to render composition like this

let composition = RenderComposition()
composition.layers = [videoLayer, audioLayer]

videoLayer sound plays 1 second and then its being muted, audioLayer audio works well but no video on composition entirely it becomes black screen.
Actually I couldn't see how to manage audio files in your documentation.
Any helps will be appreciated .

// 1. Layer 1
var url = Bundle.main.url(forResource: "video1", withExtension: "MOV")
var asset = AVAsset(url: url!)
var source = AVAssetSource(asset: asset)
source.selectedTimeRange = CMTimeRange(start: CMTime.zero, duration: asset.duration)
var timeRange = source.selectedTimeRange
let renderLayer1 = RenderLayer(timeRange: timeRange, source: source)
RenderLayer同样适用音频资源么？有没有demo展示一下？

I know how to create video from some images using AVFoundation. But exporting the video using AVFoundation takes time. I want to make this export time faster. So I want to use metal for exporting a video from UIImage. Is there any way to achieve this using this framework?

demo里，播放完视频后不释放内存，并且退到列表页也不释放内存。

VideoLab/MetalRendering.swift:113: Fatal error: Could not create render pipeline state for vertex:blendOperationVertex, fragment:blendOperationFragment, error:Error Domain=CompilerError Code=2 "reading from a rendertarget is not supported" UserInfo={NSLocalizedDescription=reading from a rendertarget is not supported}
2021-08-18 13:05:13.551297+0300 dubme-app[32964:963572] VideoLab/MetalRendering.swift:113: Fatal error: Could not create render pipeline state for vertex:blendOperationVertex, fragment:blendOperationFragment, error:Error Domain=CompilerError Code=2 "reading from a rendertarget is not supported" UserInfo={NSLocalizedDescription=reading from a rendertarget is not supported}

@ruanjx

                let tempFile = TemporaryMediaFile(withData: video.asset)
                if let asset = tempFile.avAsset {
                    let resource = AVAssetSource(asset: asset)
                    resource.selectedTimeRange = CMTimeRange(start: CMTime.zero, duration: asset.duration)
                    var timeRange = resource.selectedTimeRange
                    lastVideo = resource

                    let renderLayer1 = RenderLayer(timeRange: timeRange, source: resource)
                    
                    // 2. Composition
                    let composition = RenderComposition()
                    composition.renderSize = CGSize(width: 1280, height: 720)
                    composition.layers = [renderLayer1]

                    // 3. VideoLab
                    let videoLab = VideoLab(renderComposition: composition)

                    // 4. Make playerItem
                    let playerItem = videoLab.makePlayerItem()
                    
//                    self.renderLayers.append(renderLayer1)
                }
                ```

edit video with linear speed is easy, but now, I want to realize curve or Ramped Slow Motion such as accord from Bezier path. Any help will appreciated!

my app is crashing , when I click on save video while video is playing. And save functionality works fine when video is not playing or video ends playing , I want to save video meanwhile video playing on screen

我需要只渲染视频的一个部分到屏幕上，这可能吗？谢谢

Is it possible to make the video aspect fill and fit to the render size?

你好，我从相册取三个视频出来，url类似这样的：file:///var/mobile/Media/DCIM/105APPLE/IMG_1999.MOV。
AVURLAsset *urlAsset = (AVURLAsset *)asset;
NSURL *url = urlAsset.URL;
let asset = AVAsset(url: url)
let source = AVAssetSource(asset: asset);
后面走的代码跟simpleDemo()的是一模一样的，最终三个视频只播放了一个，VideoLab是不支持相册资源么？你们的演示示例只有加载项目里的资源。
想请教一下，这个问题出现的原因？