emarsden/dash-mpd-rs

Rust library for parsing, serializing and downloading media content from a DASH MPD manifest.

dash-mpd-rs's Issues

`fetch` feature pulls in tokio

I took a look through the new async changes and that tokio is not a dev-dependency. Since this is a library crate, the rt-multi-thread feature is not required and backoff will take care of its dependency on tokio by itself. So, tokio can be safely moved to be a dev-dependency.

0.8 regression: namespace error : Namespace prefix xsi for schemaLocation on MPD is not defined

0.8 introduced a regression with our DASH example in GStreamer.

$ xmllint -format dash_stream/manifest.mpd
dash_stream/manifest.mpd:2: namespace error : Namespace prefix xsi for schemaLocation on MPD is not defined
:mpeg:dash:profile:isoff-on-demand:2011" mediaPresentationDuration="PT1M23.333S"

With 0.7 the XML was

<MPD xmlns="urn:mpeg:dash:schema:mpd:2011" type="dynamic" schemaLocation="urn:mpeg:dash:schema:mpd:2011 DASH-MPD.xsd" profiles="urn:mpeg:dash:profile:isoff-live:2011" availabilityStartTime="2023-04-26T13:35:15.546317603Z">

which changed to this with 0.8:

<MPD xmlns="urn:mpeg:dash:schema:mpd:2011" type="dynamic" xsi:schemaLocation="urn:mpeg:dash:schema:mpd:2011 DASH-MPD.xsd" profiles="urn:mpeg:dash:profile:isoff-live:2011" availabilityStartTime="2023-04-26T13:34:24.374770165Z">

The XML parser does not seem happy with prefixing schemaLocation with xsi:.

Scripting interface to implement muxers

It would be useful to implement a scripting interface to the muxing functionality required by this library. This would make it easier to implement a new muxer, and to change the arguments used for a muxing application (in particular, ffmpeg has many commandline arguments that control encoding which can be useful to change, for example depending on the container format).

We need muxing for:

• muxing an audio and a video stream together
• adding an audio stream to a specified container format
• adding a video stream to a specified container format
• concatenating multiple media containers for multi-Period manifests

This could be implemented using the platform's commandline interface, or a scripting platform such as rhai or Python-in-WASM.

Interface to MPD XML-rewriting filters/plugins

It can be useful to allow the user to read and modify the MPD manifest before downloading media segments from it. Use cases:

• print additional diagnostics concerning the manifest
• delete some Periods that the user is not interested in (based for example on their duration)
• delete Representations that use undesired codecs
• modify the BaseURL to include another CDN
• delete all subtitle languages and formats except for the one the user is interested in (as a complement to the prefer_language functionality)
• delete audio Representations whose language the user is not interested in
• drop an audio AdaptationSet if the user only is interested in video (though this functionality is already built in to the library)

As of 2023-10-22 we have implemented support for XSLT stylesheets via xsltproc, which allows XML rewriting. This is a standards-based and easy to implement solution to filtering/rewriting, but xsltproc only supports XSLT v1.0 (v3.0 is more flexible and for example includes functions for manipulating xs:duration attributes) and XSLT is not the most widely adopted language.

Possible alternatives:

• Saxon-HE, free Java software (MPL v2) which implements XSLT 3.0
• A generic filter interface implemented as a pipe
• A command API that takes two filename arguments
• A WebAssembly-based interface that could be implemented in any programming language that can generate WASM bytecode

I will be thinking about implementing some of these in the next few months.

with 0.14.8 I can no longer build on mac m1

I moved from 0.14.7 to 0.14.8 (only using "scte35" feature - only doing serialization of MPDs, not fetching via lib) and now I get protobuf build/link errors

  ld: multiple errors: archive member '/' not a mach-o file in '/Users/buzz/comcast-development/viper/clo/clo.main/target/debug/build/protobuf-src-ac10cb5cd0b29ebe/out/install/build/src/.libs/libprotoc.a'; archive member '/' not a mach-o file in '/Users/buzz/comcast-development/viper/clo/clo.main/target/debug/build/protobuf-src-ac10cb5cd0b29ebe/out/install/build/src/.libs/libprotobuf.a'; archive member '/' not a mach-o file in '/Users/buzz/comcast-development/viper/clo/clo.main/target/debug/build/protobuf-src-ac10cb5cd0b29ebe/out/install/build/src/.libs/libprotobuf.a'

Any idea what I can do to get around it?

I figured if I wasn't using the non-serialization features I wouldn't bring them in as dependencies but it could be a transient dependency.

Thanks as always for the lib!

Have bitstreamSwitching and BitstreamSwitching properties in SegmentTemplate

For 0.15 I noticed there's bitstreamSwitching and BitstreamSwitching properties in SegmentTemplate - that seems like a mistake?

reduce default dependencies

This library is super useful but requires more dependencies than required (for parsing only).

Main mpd parser only use serde, quick-xml and regex dependencies all other dependencies are used by fetch module. My proposal is that fetch module and it's dependencies should be kept under a feature and main mpd parser in the root.

anyhow can also be replaced with std::error::Error trait. For fetch feature anyhow can be optional dep.

Missing attributes for BaseURL

byteRange, availabilityTimeComplete and availabilityTimeOffset should be part of BaseURL attribute

The schema from 23009-1:2019 is the following

<!-- Base URL -->
<xs:complexType name="BaseURLType">
  <xs:simpleContent>
    <xs:extension base="xs:anyURI">
    <xs:attribute name="serviceLocation" type="xs:string"/>
    <xs:attribute name="byteRange" type="xs:string"/>
    <xs:attribute name="availabilityTimeOffset" type="xs:double"/>
    <xs:attribute name="availabilityTimeComplete" type="xs:boolean"/>
    <xs:anyAttribute namespace="##other" processContents="lax"/>
    </xs:extension>
  </xs:simpleContent>
</xs:complexType>

Allow async HTTP

dash_mpd currently uses blocking HTTP when downloading a stream which can cause problems in async applications. Is there an intent to support async HTTP?

f64 infinity should be serialized as INF instead of inf

Serde serialize f64::INFINITY to 'inf', in XML it should be INF
see http://www.datypic.com/sc/xsd/t-xsd_float.html

You will find bellow an example of MPD that I want to deserialize and serialize again. availabilityTimeOffset="INF" becomes availabilityTimeOffset="inf" after serialization and then is not accepted by Dash.js

<?xml version="1.0" encoding="UTF-8"?>
<MPD xmlns="urn:mpeg:dash:schema:mpd:2011" profiles="urn:mpeg:dash:profile:full:2011" type="dynamic" availabilityStartTime="1970-01-01T00:00:00Z" publishTime="1970-01-01T00:00:00Z" minimumUpdatePeriod="PT1S" minBufferTime="PT1S" timeShiftBufferDepth="PT1M" maxSegmentDuration="PT1S">
  <Period id="P0" start="PT0S">
    <AdaptationSet id="1" contentType="video" par="16:9" maxWidth="3840" maxHeight="2160" segmentAlignment="true">
      <SegmentTemplate media="$RepresentationID$/$Number$.m4s" initialization="$RepresentationID$/init.mp4" duration="12800" startNumber="0" timescale="12800" availabilityTimeOffset="INF"></SegmentTemplate>
      <Representation id="2160p" bandwidth="19242008" width="3840" height="2160" mimeType="video/mp4" codecs="hev1.1.6.L153.90" startWithSAP="1"></Representation>
    </AdaptationSet>
    <AdaptationSet id="2" lang="en" contentType="audio" segmentAlignment="true">
      <SegmentTemplate media="$RepresentationID$/$Number$.m4s" initialization="$RepresentationID$/init.mp4" duration="48000" startNumber="0" timescale="48000" availabilityTimeOffset="INF"></SegmentTemplate>
      <Representation id="audio" bandwidth="297168" audioSamplingRate="48000" mimeType="audio/mp4" codecs="mp4a.40.2" startWithSAP="1">
        <AudioChannelConfiguration schemeIdUri="urn:mpeg:dash:23003:3:audio_channel_configuration:2011" value="2"></AudioChannelConfiguration>
      </Representation>
    </AdaptationSet>
  </Period>
  <UTCTiming schemeIdUri="urn:mpeg:dash:utc:http-xsdate:2014" value="https://time.akamai.com/?iso&amp;ms"></UTCTiming>
</MPD>

I'm not sure how to fix that, one possibility is to overwrite the f64 serializer

#[serde(rename = "@availabilityTimeOffset", serialize_with = "f64_to_xml")]
 pub availabilityTimeOffset: Option<f64>,

Maybe quick-xml has a solution ?

MPD Serialization

Can MPD be serialized to String?

Support for dynamic MPD files (for live streaming)

Under the Limitations / unsupported features section of the README it lists dynamic MPD manifests. Is there plans to support dynamic MPD manifests or is this something you decided not to implement?

My specific use case is as follows:
I have a media ingest server written in Rust that takes in some streams from OBS, re-encodes in various formats, and then produces HLS and DASH manifests. Currently the HLS + DASH manifests are produced by FFmpeg, but I'd like to have some more control over the manifests outputted so I was hoping to build them myself and just use FFmpeg for the encoding.

I don't actually know much about the difference between static/dynamic MPDs, so I'm not sure what "supporting" them would mean for you 😄 so was just curious

Problems with the trust-dns feature

You may have noticed in the past few days a build issue when using the "trust-dns" feature, caused by hickory-dns/hickory-dns#1946. The build failure is now resolved, but HTTP requests don't seem to be working currently (indefinite hang) with this feature enabled. This probably affects all users of the reqwest crate build with the trust-dns feature. I'll close this issue when the problem is resolved.

maybe consider updating deserializer (parser) and adding serializer for datetime fields to get micro/nanos precision

Because the iso860 crate only includes precision down to millisecond, if an MPD is delivered with more precision date/time (as some of ours are), we lose that precision. It would be helpful if there was a way we could maintain that precision.

Add derived PartialEq to data structures

Any objective to adding derived PartialEq to the data structures?

Make reqwest features optional

I like to keep my binary size down and some features specified for reqwest I don't need. For example, compression and use of rustls over native-tls are features I don't need in my application and I prefer to turn those off to speed up CI and lower binary size.

Ideally, those features would be enabled by default (using default features) but opt-out for those that don't need it. Considering the next update will be a breaking release, I think now is the ideal time to introduce these changes.

Fetch audio and video without muxing

Is there a way to download video and audio without muxing them into one file? I would like to have them as separate files.

Add serde attribute rename="Location" to MPD.locations field

dash-mpd-rs/src/lib.rs

Line 1657 in 0e543c4

pub locations: Vec<Location>,

The location tag in the dash schema is "Location" but in the MPD struct the field name is "locations".

The field is missing

#[serde(rename = "Location", default)]

AdaptationSet.Role node isn't parsed

Role struct is exported but isn't used anywhere.

https://docs.rs/dash-mpd/0.7.2/dash_mpd/struct.AdaptationSet.html?search=Role

<?xml version="1.0" encoding="UTF-8"?>
<MPD xmlns="urn:mpeg:dash:schema:mpd:2011" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:schemaLocation="urn:mpeg:dash:schema:mpd:2011 DASH-MPD.xsd" profiles="urn:mpeg:dash:profile:isoff-on-demand:2011" minBufferTime="PT2S" type="static" mediaPresentationDuration="PT734.1666870117188S">
  <Period id="0">
    <AdaptationSet contentType="audio" lang="en" subsegmentAlignment="true" group="1" label="English">
      <Role schemeIdUri="urn:mpeg:dash:role:2011" value="main" />
      <Label>English</Label>
      <Representation id="0" bandwidth="134352" codecs="mp4a.40.2" mimeType="audio/mp4" audioSamplingRate="44100">
        <AudioChannelConfiguration schemeIdUri="urn:mpeg:dash:23003:3:audio_channel_configuration:2011" value="2" />
        <BaseURL>audio-en-audio-en-mp4a.mp4</BaseURL>
        <SegmentBase indexRange="822-2329" timescale="44100">
          <Initialization range="0-821" />
        </SegmentBase>
      </Representation>
    </AdaptationSet>
    <AdaptationSet contentType="audio" lang="en-low" subsegmentAlignment="true" group="2" label="English-low">
      <Role schemeIdUri="urn:mpeg:dash:role:2011" value="main" />
      <Label>English-low</Label>
      <Representation id="1" bandwidth="133675" codecs="mp4a.40.2" mimeType="audio/mp4" audioSamplingRate="44100">
        <AudioChannelConfiguration schemeIdUri="urn:mpeg:dash:23003:3:audio_channel_configuration:2011" value="2" />
        <BaseURL>audio-de-audio-de-mp4a.mp4</BaseURL>
        <SegmentBase indexRange="822-2329" timescale="44100">
          <Initialization range="0-821" />
        </SegmentBase>
      </Representation>
    </AdaptationSet>
    <AdaptationSet contentType="audio" lang="en-high" subsegmentAlignment="true" group="3" label="English-high">
      <Role schemeIdUri="urn:mpeg:dash:role:2011" value="main" />
      <Label>English-high</Label>
      <Representation id="2" bandwidth="133550" codecs="mp4a.40.2" mimeType="audio/mp4" audioSamplingRate="44100">
        <AudioChannelConfiguration schemeIdUri="urn:mpeg:dash:23003:3:audio_channel_configuration:2011" value="2" />
        <BaseURL>audio-fr-audio-fr-mp4a.mp4</BaseURL>
        <SegmentBase indexRange="822-2329" timescale="44100">
          <Initialization range="0-821" />
        </SegmentBase>
      </Representation>
    </AdaptationSet>
    <AdaptationSet contentType="video" maxWidth="1920" maxHeight="1080" frameRate="24/1" subsegmentAlignment="true" par="16:9" group="4">
      <Role schemeIdUri="urn:mpeg:dash:role:2011" value="main" />
      <Representation id="3" bandwidth="484364" codecs="avc1.640015" mimeType="video/mp4" sar="1:1" width="512" height="288">
        <BaseURL>video-H264-288-400k-video-avc1.mp4</BaseURL>
        <SegmentBase indexRange="865-2372" timescale="24">
          <Initialization range="0-864" />
        </SegmentBase>
      </Representation>
      <Representation id="4" bandwidth="947132" codecs="avc1.64001e" mimeType="video/mp4" sar="1:1" width="640" height="360">
        <BaseURL>video-H264-360-800k-video-avc1.mp4</BaseURL>
        <SegmentBase indexRange="866-2373" timescale="24">
          <Initialization range="0-865" />
        </SegmentBase>
      </Representation>
      <Representation id="5" bandwidth="1425808" codecs="avc1.64001e" mimeType="video/mp4" sar="640:639" width="852" height="480">
        <BaseURL>video-H264-480-1200k-video-avc1.mp4</BaseURL>
        <SegmentBase indexRange="886-2393" timescale="24">
          <Initialization range="0-885" />
        </SegmentBase>
      </Representation>
      <Representation id="6" bandwidth="2498547" codecs="avc1.64001f" mimeType="video/mp4" sar="1:1" width="1280" height="720">
        <BaseURL>video-H264-720-2100k-video-avc1.mp4</BaseURL>
        <SegmentBase indexRange="866-2373" timescale="24">
          <Initialization range="0-865" />
        </SegmentBase>
      </Representation>
      <Representation id="7" bandwidth="3573424" codecs="avc1.640028" mimeType="video/mp4" sar="1:1" width="1920" height="1080">
        <BaseURL>video-H264-1080-3000k-video-avc1.mp4</BaseURL>
        <SegmentBase indexRange="867-2374" timescale="24">
          <Initialization range="0-866" />
        </SegmentBase>
      </Representation>
    </AdaptationSet>
    <AdaptationSet contentType="text" mimeType="text/vtt" lang="de" group="5" label="German">
      <Label>German</Label>
      <Role schemeIdUri="urn:mpeg:dash:role:2011" value="subtitle" />
      <Representation bandwidth="1225" id="8">
        <BaseURL>subtitle-de.vtt</BaseURL>
      </Representation>
    </AdaptationSet>
    <AdaptationSet contentType="text" mimeType="text/vtt" lang="en" group="6" label="English">
      <Label>English</Label>
      <Role schemeIdUri="urn:mpeg:dash:role:2011" value="subtitle" />
      <Representation bandwidth="1119" id="9">
        <BaseURL>subtitle-en.vtt</BaseURL>
      </Representation>
    </AdaptationSet>
    <AdaptationSet contentType="text" mimeType="text/vtt" lang="fr" group="7" label="French">
      <Label>French</Label>
      <Role schemeIdUri="urn:mpeg:dash:role:2011" value="subtitle" />
      <Representation bandwidth="1165" id="10">
        <BaseURL>subtitle-fr.vtt</BaseURL>
      </Representation>
    </AdaptationSet>
  </Period>
</MPD>

Deserialization of ContentProtection?

Even though a ContentProtection struct exists, it is not deserialized as part of the AdaptationSet struct.
Is there any reason for this?

Furthermore, I'd be interested in deserializing ContentProtection with the following structure, which currently isn't possible:

      <ContentProtection schemeIdUri="urn:uuid:edef8ba9-79d6-4ace-a3c8-27dcd51d21ed">
        <cenc:pssh>Base64Data==</cenc:pssh>
      </ContentProtection>

Adding both the schemeIdUri and cenc:pssh fields to ContentProtection would be appreciated.