Skip to content

Tag: media services

Using Azure Media Services to get metadata from a media file

Azure Media Services is a great tool for encoding all types of media.  One of it’s major advantages is that it accepts a bunch of different input types (AMS supported file types).  So you can almost agnostically give it a video file, and get an mp4 as output (amongst a myriad of other things it can do).  However, with the current version, you cant get an information about a video file until AFTER it has been transcoded.

I wanted to get information about the input video file BEFORE it is transcoded….

Sure, there are packages that can do this (MediaInfo although you can’t get audio channel info, FFProbe, TaglibSharp, etc.), but most if not all require the file to be written to disk.  That is a problem if you are looking at a byte array from blob storage, or want to get that information from a stream uploaded from a web client without writing it to disk.

So I applied a little hack to use AMS to get the audio and video metadata from a video file, and I need it quickly, so I don’t want to encode the entire video.

First, you need a JSON preset to perform the simplest (see fastest) of AMS tasks, generate a single thumbnail.

{
   "Version":1.0,
   "Codecs":[
      {
         "Type":"PngImage",
         "PngLayers":[
            {
               "Type":"PngLayer",
               "Width":640,
               "Height":360
            }
         ],
         "Start":"{Best}"
      }
   ],
   "Outputs":[
      {
         "FileName":"{Basename}_{Index}{Extension}",
         "Format":{
            "Type":"PngFormat"
         }
      }
   ]
}

This will create a single thumbnail, and it will ask AMS to generate the “best” one, meaning it uses it’s brain to figure out the most relevant thumbnail, so you don’t get a blank one because the first few seconds of the video are black.

IJob metaDataJob = CreateJob(context, $"{inputAsset.Asset.Name}-Metadatajob");                    
IMediaProcessor processor = GetLatestMediaProcessorByName("Media Encoder Standard", inputAsset.GetMediaContext());

ITask task = job.Tasks.AddNew(taskName,
        processor,
        EncodingPreset.GetBaseThumnailOnlyEncodingPreset(), // this is the JSON from above,
        TaskOptions.None);
task.InputAssets.Add(inputAsset);

task.OutputAssets.AddNew($"Thumbnail_{inputAsset.Name}", AssetCreationOptions.None);

await metaDataJob.SubmitAsync();

// Check job execution and wait for job to finish.
  Task progressJobTask = job.GetExecutionProgressTask(CancellationToken.None);
 progressJobTask.Wait();

// Get a refreshed job reference after waiting on a thread.
metaDataJob = GetJob(metaDataJob.Id, context);

// Check for errors
if (metaDataJob.State == JobState.Error)
{                    
  return;
}

I have found that regardless of the video size or video length, this process takes an average of 40 seconds (if you have a media encoder unit available, and you should if you utilize this technique: Auto Scaling Media Reserved Units in AMS .  And, so that not all is lost, you likely will have a use for this thumbnail if you are doing this sort of thing in the first place.

Once the job is complete, you can see that the output for the job is a single thumbnail, and an _metadata.xml file

image

Now we must assign a Sas locator to it so we can download that xml file:

IAsset outputAsset = job.OutputMediaAssets[0];
ILocator outputLocator = outputAsset.Locators.Where(l => l.Type == LocatorType.Sas).FirstOrDefault() ??
outputAsset.GetOrCreateLocator(LocatorType.Sas, AccessPermissions.Read | AccessPermissions.List | AccessPermissions.Write, AssetManager.CalculateExpirationDate(message));               
IAssetFile outputAssetXmlFile = outputAsset.AssetFiles.Where(file => file.Name.EndsWith(".xml")).First(); //gets guid_metadata.xml
Uri xmlUri = outputAssetXmlFile.GetSasUri(outputLocator);

By the way, GetOrCreateLocator is an extension to help reuse locators since AMS limits you to only 5:

public static ILocator GetOrCreateLocator(this IAsset asset, LocatorType locatorType, AccessPermissions permissions, TimeSpan duration, DateTime? startTime = null, TimeSpan? expirationThreshold = null)
{
  MediaContextBase context = asset.GetMediaContext();

  ILocator assetLocator = context.Locators.Where(l => l.AssetId == asset.Id && l.Type == locatorType).OrderByDescending(l => l.ExpirationDateTime).ToList().Where(l => (l.AccessPolicy.Permissions & permissions) == permissions).FirstOrDefault();

  if (assetLocator == null)
  {
    // If there is no locator in the asset matching the type and permissions, then a new locator is created.
    assetLocator = context.Locators.Create(locatorType, asset, permissions, duration, startTime);
  }
  else if (assetLocator.ExpirationDateTime <= DateTime.UtcNow.Add(expirationThreshold ?? DefaultExpirationTimeThreshold))
  {
    // If there is a locator in the asset matching the type and permissions but it is expired (or near expiration), then the locator is updated.
    assetLocator.Update(startTime, DateTime.UtcNow.Add(duration));
  }

  return assetLocator;
}

Ok, so now we have an xml file containing the metadata, and it looks like this:

<?xml version="1.0"?>
<AssetFiles xmlns:xsd="http://www.w3.org/2001/XMLSchema" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xmlns="http://schemas.microsoft.com/windowsazure/mediaservices/2014/07/mediaencoder/inputmetadata">
  <AssetFile Name="Taylor_Swift_-_Blank_Space_mp4_ItemId(1040).mp4" Size="62935612" Duration="PT4M32.463S" NumberOfStreams="2" FormatNames="mov,mp4,m4a,3gp,3g2,mj2" FormatVerboseName="QuickTime / MOV" StartTime="PT0S" OverallBitRate="1847">
    <VideoTracks>
      <VideoTrack Id="1" Codec="h264" CodecLongName="H.264 / AVC / MPEG-4 AVC / MPEG-4 part 10" TimeBase="1/90000" NumberOfFrames="6531" StartTime="PT0S" Duration="PT4M32.355S" FourCC="avc1" Profile="High" Level="4.0" PixelFormat="yuv420p" Width="1920" Height="1080" DisplayAspectRatioNumerator="16" DisplayAspectRatioDenominator="9" SampleAspectRatioNumerator="1" SampleAspectRatioDenominator="1" FrameRate="23.980" Bitrate="1716" HasBFrames="1">
        <Disposition Default="1" Dub="0" Original="0" Comment="0" Lyrics="0" Karaoke="0" Forced="0" HearingImpaired="0" VisualImpaired="0" CleanEffects="0" AttachedPic="0"/>
        <Metadata key="language" value="und"/>
        <Metadata key="handler_name" value="VideoHandler"/>
      </VideoTrack>
    </VideoTracks>
    <AudioTracks>
      <AudioTrack Id="2" Codec="aac" CodecLongName="AAC (Advanced Audio Coding)" TimeBase="1/44100" NumberOfFrames="11734" StartTime="PT0S" Duration="PT4M32.463S" SampleFormat="fltp" ChannelLayout="stereo" Channels="2" SamplingRate="44100" Bitrate="125" BitsPerSample="0">
        <Disposition Default="1" Dub="0" Original="0" Comment="0" Lyrics="0" Karaoke="0" Forced="0" HearingImpaired="0" VisualImpaired="0" CleanEffects="0" AttachedPic="0"/>
        <Metadata key="language" value="und"/>
        <Metadata key="handler_name" value="SoundHandler"/>
      </AudioTrack>
    </AudioTracks>
    <Metadata key="major_brand" value="isom"/>
    <Metadata key="minor_version" value="512"/>
    <Metadata key="compatible_brands" value="isomiso2avc1mp41"/>
    <Metadata key="encoder" value="Lavf56.40.101"/>
  </AssetFile>
</AssetFiles>

That has the information I am looking for!  We can consume it now:

XDocument doc = XDocument.Load(xmlUri.ToString());

Dictionary<XName, string> videoAttributes = doc.Descendants().First(element => element.Name.LocalName == "VideoTrack").Attributes().ToDictionary(attribute => attribute.Name, attribute => attribute.Value);

int width = Convert.ToInt32(videoAttributes["Width"]);

int height = Convert.ToInt32(videoAttributes["Height"]);

int bitrate = Convert.ToInt32(videoAttributes["Bitrate"]);

Dictionary<XName, string> audioAttributes = doc.Descendants().First(element => element.Name.LocalName == "AudioTrack").Attributes().ToDictionary(attribute => attribute.Name, attribute => attribute.Value);

int channels = Convert.ToInt32(audioAttributes["Channels"]);

or pull anything else out of that file that you want.

In my post about creating a custom bitrate ladder (see Creating a custom bitrate ladder from AMS), we need that information BEFORE we transcode so we can set the ceiling of our bitrate ladder.

I spoke with David Bristol from Microsoft about this (his blog has a bunch of great AMS related information David Bristol’s Media blog), and he agrees that something like the output of Media Info or FFMPEG’s ffprobe.exe would be great to run on an uploaded asset and is suggesting it to the AMS team, so hopefully we will see this kind of functionality in the future.  40 seconds isn’t great, so I hope we can improve on that, but this might get you to the dance for now.

Leave a Comment

Creating a custom bitrate ladder from Azure Media Services Transcoding

When submitting a transcoding job to Azure Media Services with Media Encoder Standard, the documentation will tell you to use one of the provided presets like this:

string configuration = File.ReadAllText(@"c:\supportFiles\preset.json"); // Create a task

ITask task = job.Tasks.AddNew("Media Encoder Standard encoding task", processor, configuration, TaskOptions.None);

//https://docs.microsoft.com/en-us/azure/media-services/media-services-mes-presets-overview

or by Adaptive Streaming by adding a task like this:

ITask task = job.Tasks.AddNew("My encoding task", processor, "Adaptive Streaming", TaskOptions.None);

In the first example, you are creating multi-bitrate mp4s all the way up to 1080, or even 4k if that is the preset you selected.  In the latter example, what this is doing under the covers is great; You are telling AMS to create the bitrate ladder on the fly based on the input, and to let Microsoft work its magic.  But there are limitations to using Adaptive Streaming from C#, one being that you can’t add thumbnails in the same job, for example.

So what if you want a little more control?  I’ve created a fluent interface for creating your own presets and creating a bitrate ladder that doesn’t “up-encode” based on the quality of the original video.

First, we need to define an EncodingPreset class that will eventually be converted to JSON in valid MES preset format:

public class EncodingPreset
     {
         /// <inheritdoc />
         private EncodingPreset()
         {
             Codecs = new List<Codec>();
             Outputs = new List<Output>();
         }

        public double Version { get; set; }
         public List<Codec> Codecs { get; set; }
         public List<Output> Outputs { get; set; }

        public static EncodingPreset GetBaseEncodingPreset()
         {
             var preset = new EncodingPreset
                          {
                              Version = 1.0d
                          };

            preset.Codecs.Add(Codec.GetH264Codec());
             preset.Outputs.Add(Output.GetMp4Output());

            return preset;
         }

        public EncodingPreset AddNormalAudio()
         {
             Codec codec = Codecs.FirstOrDefault(c => c.Type == "AACAudio");
             if (codec == null)
             {
                 Codec audioCodec = Codec.GetNormalAudioCodec();

                Codecs.Add(audioCodec);
             }

            return this;
         }

        public EncodingPreset AddHDAudio()
         {
             Codec codec = Codecs.FirstOrDefault(c => c.Type == "AACAudio");
             if (codec == null)
             {
                 Codec audioCodec = Codec.GetHDAudioCodec();
                 Codecs.Add(audioCodec);
             }

            return this;
         }

        public EncodingPreset AddBitrateLadder(int width, int height, int bitrate)
         {
             IList<ResolutionInfo> orderedLadder = BitrateLadder.OrderedLadder; //lowest to highest resolution
             int originalPixels = width * height;
             var bitrateTolerance = .05;

            var layersToGenerate = new List<ResolutionInfo>
                                    {
                                        new ResolutionInfo // add the original
                                        {
                                            Width = width,
                                            Height = height,
                                            Bitrate = bitrate
                                        }
                                    };
             foreach (ResolutionInfo step in orderedLadder)
             {
                 if (step.Pixels <= originalPixels)
                 {
                     int min = Math.Min(step.Bitrate, bitrate);
                     layersToGenerate.Add(new ResolutionInfo
                                          {
                                              Width = step.Width,
                                              Height = step.Height,
                                              Bitrate = min
                                          });
                 }
             }

            // make the bitrates distinct - not sure i like this
             List<ResolutionInfo> orderedLayersToGenerate = layersToGenerate.OrderBy(info => info.Pixels).ThenBy(info => info.Bitrate).ToList();
             for (var i = 0; i < orderedLayersToGenerate.Count - 1; i++)
             {
                 foreach (ResolutionInfo layerToGenerate in orderedLayersToGenerate.Where(layerToGenerate => orderedLayersToGenerate.Any(info => info.Bitrate == layerToGenerate.Bitrate && info.Pixels != layerToGenerate.Pixels)))
                 {
                     layerToGenerate.Bitrate = layerToGenerate.Bitrate - 1;
                 }
             }

            foreach (ResolutionInfo layerToGenerate in orderedLayersToGenerate.Where(layerToGenerate => !HasExistingStepWithinTolerance(layerToGenerate.Width, layerToGenerate.Height, layerToGenerate.Bitrate, bitrateTolerance)))
             {
                 AddVideoLayer(layerToGenerate.Width, layerToGenerate.Height, layerToGenerate.Bitrate);
             }

            return this;
         }

        private bool HasExistingStepWithinTolerance(int width, int height, int min, double bitrateTolerance)
         {
             Codec codec = Codecs.FirstOrDefault(c => c.Type == "H264Video");
             if (codec == null)
             {
                 return false;
             }
             return codec.H264Layers.Any(layer => layer.Width == width && layer.Height == height && Math.Abs((layer.Bitrate - min) / (double) layer.Bitrate) <= bitrateTolerance);
         }

        public EncodingPreset AddVideoLayer(int width, int height, int bitrate)
         {
             H264Layer h264Layer = H264Layer.GetVideoLayer(width, height, bitrate);
             Codec codec = Codecs.FirstOrDefault(c => c.Type == "H264Video");
             if (codec == null)
             {
                 codec = Codec.GetH264Codec();
                 Codecs.Add(codec);
             }

            if (!codec.H264Layers.Any(layer => layer.Width == width && layer.Height == height && layer.Bitrate == bitrate))
             {
                 codec.H264Layers.Add(h264Layer);
             }

            return this;
         }

        public EncodingPreset AddPngThumbnails()
         {
             Codec codec = Codecs.FirstOrDefault(c => c.Type == "PngImage");
             if (codec == null)
             {
                 PngLayer pngLayer = PngLayer.Get640x360Thumbnail();

                Codec thumbnailCodec = Codec.GetPngThumbnailCodec();
                 thumbnailCodec.Start = "00:00:01";
                 thumbnailCodec.Step = "00:00:01";
                 thumbnailCodec.Range = "00:00:58";
                 thumbnailCodec.Type = "PngImage";
                 thumbnailCodec.PngLayers.Add(pngLayer);

                Codecs.Add(thumbnailCodec);

                Outputs.Add(Output.GetPngThumbnailOutput());
             }

            return this;
         }
     }
}

With supporting classes for the collections and other classes:

 

public class Codec
     {
         private Codec()
         {
         }

        public string KeyFrameInterval { get; set; }
         public List<H264Layer> H264Layers { get; set; }
         public string Type { get; set; }
         public List<PngLayer> PngLayers { get; set; }
         public string Start { get; set; }
         public string Step { get; set; }
         public string Range { get; set; }
         public string Profile { get; set; }
         public int? Channels { get; set; }
         public int? SamplingRate { get; set; }
         public int? Bitrate { get; set; }
         public string Condition { get; set; }

        public static Codec GetH264Codec()
         {
             return new Codec
                    {
                        Type = "H264Video",
                        KeyFrameInterval = "00:00:02",
                        H264Layers = new List<H264Layer>()
                    };
         }

        public static Codec GetNormalAudioCodec()
         {
             return new Codec
                    {
                        Type = "AACAudio",
                        Profile = "AACLC",
                        Channels = 2,
                        SamplingRate = 48000,
                        Bitrate = 128,
                        Condition = "InsertSilenceIfNoAudio"
                    };
         }

        public static Codec GetHDAudioCodec()
         {
             return new Codec
                    {
                        Type = "AACAudio",
                        Profile = "AACLC",
                        Channels = 6,
                        SamplingRate = 48000,
                        Bitrate = 384,
                        Condition = "InsertSilenceIfNoAudio"
                    };
         }

        public static Codec GetPngThumbnailCodec()
         {
             return new Codec
                    {
                        Type = "PngImage",
                        Start = "00:00:01",
                        Step = "00:00:01",
                        Range = "00:00:58",
                        PngLayers = new List<PngLayer>()
                    };
         }
     }

public class Output
  {
      private Output()
      {
      }

     public string FileName { get; set; }
      public Format Format { get; set; }

     public static Output GetMp4Output()
      {
          return new Output
                 {
                     Format = new Format
                              {
                                  Type = "MP4Format"
                              },
                     FileName = "{Basename}_{Width}x{Height}_{VideoBitrate}{Extension}"
                 };
      }

     public static Output GetPngThumbnailOutput()
      {
          return new Output
                 {
                     Format = new Format
                              {
                                  Type = "PngFormat"
                              },
                     FileName = "{Basename}_{Index}{Extension}"
                 };
      }
  }

public class H264Layer
    {
        private H264Layer()
        {
        }

       public string Profile { get; set; }
        public string Level { get; set; }
        public int Bitrate { get; set; }
        public int MaxBitrate { get; set; }
        public string BufferWindow { get; set; }
        public int Width { get; set; }
        public int Height { get; set; }
        public int BFrames { get; set; }
        public int ReferenceFrames { get; set; }
        public bool AdaptiveBFrame { get; set; }
        public string Type { get; set; }
        public string FrameRate { get; set; }

       public static H264Layer GetVideoLayer(int width, int height, int bitrate)
        {
            return new H264Layer
                   {
                       Profile = "Auto",
                       Level = "auto",
                       Bitrate = bitrate,
                       MaxBitrate = bitrate,
                       BufferWindow = "00:00:05",
                       Width = width,
                       Height = height,
                       BFrames = 3,
                       ReferenceFrames = 3,
                       AdaptiveBFrame = true,
                       Type = "H264Layer",
                       FrameRate = "0/1"
                   };
        }
    }

public class PngLayer
    {
        private PngLayer()
        {
        }

       public string Type { get; set; }
        public int Width { get; set; }
        public int Height { get; set; }

       public static PngLayer Get640x360Thumbnail()
        {
            return new PngLayer
                   {
                       Height = 360,
                       Width = 640,
                       Type = "PngLayer"
                   };
        }
    }

public class Format
    {
        public string Type { get; set; }
    }

a class to hold our original video information to compare to our ideal ladder:

public class ResolutionInfo
    {
        public int Width { get; set; }
        public int Height { get; set; }
        public int Bitrate { get; set; }

       public long Pixels
        {
            get
            {
                return Width * Height;
            }
        }
    }

and an extension method to convert to json properly for this case:

public static class EncodingPresetExtensions
     {
         public static string ToJson(this EncodingPreset preset)
         {
             return JsonConvert.SerializeObject(preset,
                                                new JsonSerializerSettings
                                                {
                                                    NullValueHandling = NullValueHandling.Ignore
                                                });
         }
     }

and finally our ideal bitrate ladder:

public static class BitrateLadder
    {
        private static readonly IList<ResolutionInfo> Ladder = new List<ResolutionInfo>();

       static BitrateLadder()
        {
            Ladder.Add(
                       new ResolutionInfo
                       {
                           Bitrate = 20000,
                           Width = 4096,
                           Height = 2304
                       });
            Ladder.Add(
                       new ResolutionInfo
                       {
                           Bitrate = 18000,
                           Width = 3840,
                           Height = 2160
                       });
            Ladder.Add(
                       new ResolutionInfo
                       {
                           Bitrate = 16000,
                           Width = 3840,
                           Height = 2160
                       });
            Ladder.Add(
                       new ResolutionInfo
                       {
                           Bitrate = 14000,
                           Width = 3840,
                           Height = 2160
                       });
            Ladder.Add(
                       new ResolutionInfo
                       {
                           Bitrate = 12000,
                           Width = 2560,
                           Height = 1440
                       });
            Ladder.Add(
                       new ResolutionInfo
                       {
                           Bitrate = 10000,
                           Width = 2560,
                           Height = 1440
                       });
            Ladder.Add(
                       new ResolutionInfo
                       {
                           Bitrate = 8000,
                           Width = 2560,
                           Height = 1440
                       });
            Ladder.Add(
                       new ResolutionInfo
                       {
                           Bitrate = 6000,
                           Width = 1920,
                           Height = 1080
                       });
            Ladder.Add(
                       new ResolutionInfo
                       {
                           Bitrate = 4700,
                           Width = 1920,
                           Height = 1080
                       });
            Ladder.Add(
                       new ResolutionInfo
                       {
                           Bitrate = 3400,
                           Width = 1280,
                           Height = 720
                       });
            Ladder.Add(
                       new ResolutionInfo
                       {
                           Bitrate = 1500,
                           Width = 960,
                           Height = 540
                       });
            Ladder.Add(
                       new ResolutionInfo
                       {
                           Bitrate = 1000,
                           Width = 640,
                           Height = 360
                       });
        }

       /// <inheritdoc />
        public static IList<ResolutionInfo> OrderedLadder
        {
            get
            {
                return Ladder.OrderBy(pair => pair.Pixels).ThenBy(info => info.Bitrate).ToList();
            }
        }
    }

Note that I have set some defaults in these classes for my particular use case.

So let’s talk about the AddBitrateLadder function:

It takes in the width, height, and bitrate from the origin media file so as not to wastefully, “up-encode” it.  Then, it creates a ladder making the “top” layer the original specs, and steps down from there using our ideal bitrate ladder as a guide.  I should also note that AMS keys off the bitrate, so you can not have 2 different resolutions with the same bitrate, and that is why there is code in that method to merely subtract 1 from each bitrate to make them unique if the original video quality is too low to fit into our specified ladder.  Lastly, it includes a tolerance so that you don’t create 2 layers that are virtually identical.

So now I can use this to generate me a custom bitrate ladder with normal audio and thumbnails, for example:

EncodingPreset.GetBaseEncodingPreset()
.AddNormalAudio()
.AddPngThumbnails()
.AddBitrateLadder(playlistItem.AVFile.Width, playlistItem.AVFile.Height, playlistItem.AVFile.Bitrate);

or HD Audio with no thumbnails:

EncodingPreset.GetBaseEncodingPreset()
.AddHDAudio()
.AddBitrateLadder(playlistItem.AVFile.Width, playlistItem.AVFile.Height, playlistItem.AVFile.Bitrate);]

etc. etc.

And it’s totally testable:

[TestMethod]
       public void CalcLayers1920x1080at266()
       {
           List<H264Layer> layers = CalcLayers(1920, 1080, 266);
           Assert.AreEqual(4, layers.Count);

           H264Layer layer1 = layers[0];
           Assert.AreEqual(263, layer1.Bitrate);
           Assert.AreEqual(360, layer1.Height);
           Assert.AreEqual(640, layer1.Width);

           H264Layer layer2 = layers[1];
           Assert.AreEqual(264, layer2.Bitrate);
           Assert.AreEqual(540, layer2.Height);
           Assert.AreEqual(960, layer2.Width);

           H264Layer layer3 = layers[2];
           Assert.AreEqual(265, layer3.Bitrate);
           Assert.AreEqual(720, layer3.Height);
           Assert.AreEqual(1280, layer3.Width);

           H264Layer layer4 = layers[3];
           Assert.AreEqual(266, layer4.Bitrate);
           Assert.AreEqual(1080, layer4.Height);
           Assert.AreEqual(1920, layer4.Width);
       }

 private static List<H264Layer> CalcLayers(int width, int height, int bitrate)
       {
           EncodingPreset preset1 = EncodingPreset.GetBaseEncodingPreset()
                                                  .AddNormalAudio()
                                                  .AddPngThumbnails()
                                                  .AddBitrateLadder(width, height, bitrate);
           return preset1.Codecs.Where(codec => codec.Type == "H264Video")
                         .SelectMany(codec => codec.H264Layers)
                         .ToList();
       }

Then, when it is time to submit my job, I can:

ITask task = job.Tasks.AddNew(“My encoding task”, processor, myPreset.ToJson(), TaskOptions.None);

Boom! Now we have the power of Adaptive Streaming with the benefit of more control over the ideal ladder, as well as other functions of AMS.

Leave a Comment

Auto scaling Media Reserved Units in Azure Media Services

When you spin up an Azure Media Services instance in Azure, you are prompted with a choice:  How many Media Reserved Units do you want?  and what horsepower do you want behind them?

Well, that exactly does that mean?

Reserving a Unit means that when you submit a job to Media Services, you wont go in a public queue in order for your submitted job to start.  This is important, because if the public queue is busy, it could take quite a while for your job to get picked up.  If you have all the time in the world for your job to complete, this isn’t a big deal, but if you are like me with a customer waiting on the job, speed is a priority.  You can choose from 1-10 reserved units (you can request  more via a support request), and they come at a cost.  Also, when you reserve a unit, it has be a specific speed (S1, S2, or S3).

image

So if you want to have 10 reserved units at all times, and you want S3 so the job completes the fastest that Azure offers, that is 80 cents an hour, and that can add up over time.  I should also note that you can NOT reserve zero S2 or S3 units.  If you want to be in the public pool, it has to be S1.  Therefore, you are 4 cents an hour at the very least if you want to have an immediate response time of your jobs by reserving one S1.  I should also note that if you made a support request to get more than 10 units, when you change the speed of those reserved units, the MaxReservableUnits gets reset to 10, and your support request is essentially lost.  I have spoken with Azure support on this, and while they don’t call it a bug, it is something they are addressing in a future release of AMS.

So, the solution I came up with was to auto scale our units with C#.

When a message is sent to my worker role to work with Azure Media Services, I reserve (currently reserved units + 1) S3 units, and when it is done I decrement one S3 unit.  When I hit 0 units, I set the speed back to S1 (because remember you can only have zero units if you are set to S1)

internal static async Task ReserveMediaEncodingUnit(MediaContextBase context, int amount)
      {
          if (ConfigurationProvider.AutoScaleMRU())
          {
              IEncodingReservedUnit encodingReservedUnit = context.EncodingReservedUnits.FirstOrDefault(); //there is always only one of these (https://github.com/Azure/azure-sdk-for-media-services/blob/dev/test/net/Scenario/EncodingReservedUnitDataTests.cs)
              if (encodingReservedUnit != null)
              {
                   encodingReservedUnit.CurrentReservedUnits = Math.Min(amount,
                                                                       ConfigurationProvider.MaxMRUProvisioned() == 0
                                                                           ? encodingReservedUnit.MaxReservableUnits
                                                                           : ConfigurationProvider.MaxMRUProvisioned());
                   encodingReservedUnit.ReservedUnitType = ReservedUnitType.Premium;
                   await encodingReservedUnit.UpdateAsync();
               }
           }
       }

ConfigurationProvider.MaxMRUProvisioned() is a setting I have that is equal to 10.  I did that because I initially put in the service request to get more than 10, only to find out it gets reset back to 10 if you change the speed.  If Microsoft changes this behavior, I can set my setting to 0 and user their variable MaxReservedUnits, without any code changes.

Deallocating units:

 

 internal static async Task DeallocateMediaEncodingUnit(MediaContextBase context, int amount)
       {
           if (ConfigurationProvider.AutoScaleMRU())
           {
               IEncodingReservedUnit encodingReservedUnit = context.EncodingReservedUnits.FirstOrDefault(); //there is always only one of these (https://github.com/Azure/azure-sdk-for-media-services/blob/dev/test/net/Scenario/EncodingReservedUnitDataTests.cs)

              if (encodingReservedUnit != null)
               {
                   encodingReservedUnit.CurrentReservedUnits = Math.Max(0, amount);
                   encodingReservedUnit.ReservedUnitType = encodingReservedUnit.CurrentReservedUnits == 0
                                                               ? ReservedUnitType.Basic
                                                               : ReservedUnitType.Premium;

                  await encodingReservedUnit.UpdateAsync();
               }
           }
       }

If I hit 0 units I can reset:

 

    private static async Task ResetMediaEncodingUnits(MediaContextBase context)
       {
           if (ConfigurationProvider.AutoScaleMRU())
           {
               IEncodingReservedUnit encodingReservedUnit = context.EncodingReservedUnits.FirstOrDefault(); //there is always only one of these (https://github.com/Azure/azure-sdk-for-media-services/blob/dev/test/net/Scenario/EncodingReservedUnitDataTests.cs)

              if (encodingReservedUnit != null)
               {
                   encodingReservedUnit.CurrentReservedUnits = 0;
                   encodingReservedUnit.ReservedUnitType = ReservedUnitType.Basic;
                   await encodingReservedUnit.UpdateAsync();
               }
           }
       }

So now, when my users aren’t transcoding anything, and my AMS instance is sitting idle, I will incur no cost.  And, when they submit a job, I allocate a unit to avoid going to the public pool and the job gets submitted right away and completed with premium speeds.  I can’t guarantee this hack will work forever; when speaking with MS they told me this code has prompted them to think about how reserved units work in AMS, and may change this behavior in the future.

Happy transcoding!

3 Comments