Jet-Stream, the independent streaming specialist, with 32 years experience in streaming innovations, benchmarked multiple popular vod and live streaming services for encoding and packaging compliance.

We benchmarked the streams you actually watch. Most of them are not what they claim to be.

Jet-Stream benchmarked popular VOD and live streaming services on one simple question: does the stream that reaches your screen comply with the HLS specification it claims to follow?

Not the marketing page, not the vendor datasheet, but the actual playlists and segments, captured exactly as a real viewer receives them.

Why encoding and packaging compliance matters

Every streaming service says it delivers “adaptive, high-quality HLS”. Almost none of them are ever really checked. What is checked? Real end user metrics, checking CDN performance. Yet encoding and packaging are where streaming quality is decided. A stream that violates the spec in small ways still plays, most of the time, on most devices. That is exactly the problem: the failures are invisible in the demo and visible in the wild, or end up at your storage and CDN bill.

Non-compliant packaging causes buffering that nobody can explain, playback that breaks on one TV brand but not another, subtitles that drift, ads that stall, and accessibility features that silently disappear. Non compliant encoding and packaging inflates CDN bills through poorly built ladders and redundant renditions.

Compliance is not a checkbox. It is the foundation of interoperability, resilience and independence. Compliance is fundamental for digital sovereignty: you want to have the ability to migrate without encountering incompatibility issues.

How we did it

We wanted a method that anyone could repeat and nobody could argue with.

  • Real streams, real conditions. Every stream was captured as a normal subscriber or visitor receives it, including behind-login services. No test feeds, no vendor-supplied samples.
  • One run, one moment. All services were captured in the same period, so results reflect the same generation of tooling.
  • Objective criteria, public sources. Every check references documented requirements: Apple’s HLS authoring specification, the HLS RFC, and published packaging best practices. No opinions, only measurable deviations.
  • Full evidence trail. Source playlists and segments are stored per service. Every finding can be traced back to a file. The benchmark can easily be reproduced by anyone.
  • Jet-Stream tested as an ordinary participant. Our own platform went through exactly the same checks as everyone else, with no special treatment.

The benchmark covered multiple streaming chains: public and commercial broadcasters, regional media, news and radio, sports platforms, global subscription services, parliamentary broadcasting, and the encoding, packaging and player vendors behind them.

What we found

We are not publishing scores or naming names. This benchmark was not built to embarrass anyone. It was built to establish a factual baseline for an industry that has been running on assumptions for far too long.

What we can say:

  • [General pattern 1, e.g. “The majority of streams deviate from the specification on at least one point that affects playback compatibility.”]
  • [General pattern 2, e.g. “Bitrate ladders are frequently built for the encoder’s convenience, not for the viewer’s network.”]
  • [General pattern 3, e.g. “Compliance is not correlated with budget. Some of the largest services made basic mistakes; some of the smallest were nearly flawless.”]

The pattern is clear enough: the industry has quietly accepted “it plays” as the definition of quality. It should not. We have found quite some fundamental mistakes in how encoders are configured and in how packagers are designed.

Five issues stood out most

  1. Configuration without understanding. In many chains, encoders and packagers are clearly set up by people who know how to make a stream play, but not why the specification says what it says. Defaults are accepted, parameters are copied from another project. If the picture appears, the job is considered done. The specification exists precisely because “it plays” is not the same as “it is correct”.
  2. Fundamental architectural flaws. Some solutions are designed in a way that only works if everyone else in the chain is perfect, which may work in a closed broadcasting network, but not in the real world with challenging sources, network conditions and a mix of vendors in a pipeline. The encoder must be exact, the CDN must be forgiving, the player must compensate. When something breaks, the finger points to the next party. We saw architectures that make no allowance for the normal imperfections of real-world delivery, and then treat those imperfections as someone else’s failure, while the rest of the industry warns for such fundamental architectural mistakes. We saw architectures where content is not packaged on the fly but is prepackaged, with non compliant packaging, which means that all content needs to be re-packaged for compliance, instead of tuning the just in time packager.
  3. Nobody owns the output. In a modern streaming chain, encoding, packaging, delivery and playback are often provided by different vendors. Each one tests its own piece. You would expect the integrator to capture, analyze and benchmark the final stream as the viewer receives it and checks it end to end on compliance.
  4. Waste nobody notices. Because no one validates the actual output, inefficiencies persist for years: renditions that carry far more data than they need to, ladders that steer viewers to a worse picture than their connection allows, and re-processing that quietly breaks caching. None of this shows up as an outage. It shows up as cost, energy and a mediocre viewing experience that everyone has learned to accept.
  5. A polished front, an unexamined back. The quality of a service’s website, a RUM CDN performance analysis, a player skin or sales deck said nothing about the compliance of its stream. Some of the most professional-looking services delivered some of the least professional output, and vice versa.

What this means for broadcasters and platforms

If you operate a streaming service, you probably do not know how compliant your stream actually is. Your vendor will tell you it is fine. Players will hide most of the problems. Your viewers will experience them and blame their internet connection. You are accepting high storage and traffic bills, because you assume that the price per GB is right, but in the meantime a lot of unnecessary GB’s are generated, effectively doubling the price per GB. Who’s taking your money on their bad encoding setup?

Three things you can do today:

  1. Ask for evidence, not assurances. Request a compliance report on your live production stream, not on a lab sample.
  2. Test with more than one player. If your stream only works in your vendor’s player, you do not have a stream. You have a dependency.
  3. Treat packaging as infrastructure. Packager compliance deserves the same scrutiny as your CDN contract and your security audit.
  4. Audit the actual output streams. Save manifests and segments and analyze them on compliance and correctness.

Why we did this

Jet-Stream has built streaming infrastructure since 1994. We have seen the industry move from proprietary formats to open standards, and we have seen those standards eroded by convenience and by vendors who benefit from lock-in. Open, compliant streams are not a technical nicety. They are what keeps European media independent of any single player, platform or cloud.

Get the benchmark results

Reach out to us for a confidential call in which we will share the benchmark results. We are also happy to share the methodology, criteria and a benchmark instruction set with any broadcaster, platform or vendor who wants to benchmark their own chain, and we are happy to run a compliance check on your production streams.

Go back to see all blogposts