Add a bit more explanation

Examples/Mixing/Sources/Mixing/main.swift

@@ -131,10 +131,13 @@ let onConnection: LiveOnConnection = { pub, sub in
         // Here we are pulling samples off of our audio and video buses that have been generated by the composer
         // There is a GPU barrier to download the textuure from the GPU introduced here.  This must be used
         // even if using the Composer class because the composer makes no assumptions about where you want 
-        // the texture to live after it's finished with it.
+        // the texture to live after it's finished with it. 
         //
         // We must also filter the assets coming from the bus so that we only get the samples we want rather than
-        // all of them.
+        // all of them. You'll notice as well that when we compose the rtmp publisher at the end of the
+        // composition here we use a standard composition operator for video (>>>) and for audio we use a
+        // mapping composition operator (|>>).  The mapping operator will take a list of samples and map them to the
+        // publisher, returning a list of the result type.
         rtmpPublisher = (audioBus <<| assetFilter("composer") >>> FFmpegAudioEncoder(.aac, bitrate: 96000) |>> txn,
             pictureBus <<| assetFilter("composer") >>> GPUBarrierDownload(compute) >>> FFmpegVideoEncoder(.avc,
                 bitrate: 3_000_000, frameDuration: manifest.video.frameDuration) >>> txn)

Examples/Transcoding/Sources/Transcoding/main.swift

@@ -14,7 +14,9 @@
    limitations under the License.
 */
 
-// This example is going to use the Transcode class to transcode a file and output it to an RTMP endpoint.
+// This example is going to use the transcoder functions to transcode a file and output it to an RTMP endpoint.
+// The transcoder functions return compositions containing the appropriate operations to transcode a CodedMediaSample
+// into one (or more) different CodedMediaSamples of the same media type (audio, video)
 
 import SwiftVideo
 import Foundation
@@ -44,13 +46,19 @@ let onEnded: LiveOnEnded = {
 // RTMP connection established, here we create the encoders and compose the outputs.
 let onConnection: LiveOnConnection = { pub, sub in
     if let pub = pub,
-       let txn = pub as? Terminal<CodedMediaSample> {
+       let publisher = pub as? Terminal<CodedMediaSample> {
         do {
             let src = try FileSource(clock, url: inputFile, assetId: "file", workspaceId: "sandbox")
+
+            // Here we are composing transformations: taking the encoded media from the file source, filtering based
+            // on the media type (audio or video), and transcoding the samples. You'll notice that when we compose
+            // the rtmp publisher at the end of the composition here we use a standard composition operator for 
+            // video (>>>) and for audio we use a mapping composition operator (|>>).  The mapping operator 
+            // will take a list of samples and map them to the publisher, returning a list of the result type.
             videoTranscoder = try codedBus <<| mediaTypeFilter(.video) >>> makeVideoTranscoder(.avc,
-                bitrate: 3_000_000, keyframeInterval: TimePoint(2000, 1000), newAssetId: "new") >>> txn
+                bitrate: 3_000_000, keyframeInterval: TimePoint(2000, 1000), newAssetId: "new") >>> publisher
             audioTranscoder = try codedBus <<| mediaTypeFilter(.audio) >>>
-                    makeAudioTranscoder(.aac, bitrate: 96_000, sampleRate: 48000, newAssetId: "new") |>> txn
+                    makeAudioTranscoder(.aac, bitrate: 96_000, sampleRate: 48000, newAssetId: "new") |>> publisher
             fileSource = src >>> codedBus
         } catch {
             print("Exception loading file \(error)")

README.md

@@ -6,13 +6,19 @@ Video streaming and processing framework for Linux, macOS, and iOS/iPadOS (maybe
 
 This is very much a WIP.  Support is currently better for Linux than macOS and even less-so for iOS devices.  Work needs to be done on the Podfile for better Xcode interop (swift package manager doesn't do such a great job there).
 
+
+### Getting Started
+
+For now, check the Examples directory for some hints about how to use this framework.  I promise I will be creating
+documentation to clarify how it all fits together and how you can do useful, interesting things with this framework.
+
 ### Current Features
 
 - RTMP Client and Server
 - "Flavor" Client and Server (toy protocol, see flavor.md)
 - OpenCL Support
 - Metal Support
-- Audio Decode/Encode (via FFmpeg)
+- Audio Decode/Encode (via FFmpeg and CoreAudio)
 - Video Decode/Encode (via FFmpeg and CoreVideo)
 - Camera capture (macOS/iOS)
 - Text rendering (via FreeType)
@@ -21,11 +27,4 @@ This is very much a WIP.  Support is currently better for Linux than macOS and e
 - Audio Resampler (via FFmpeg+SOX)
 
 
-### Short ToDo
-
-- Audio Capture and Playback (macOS/iOS)
-- Audio Encode/Decode (via CoreAudio for macOS/iOS)
-- More Metal colorspace options, use Metal as default on macOS/iOS/tvOS
-
-
 FFmpeg support is thanks to https://github.com/sunlubo/SwiftFFmpeg