Jellyfin 1080p Mini PC Transcode Benchmark

purpose and claim boundary

Test whether one declared exact configuration can sustain one repeatable 1080p H.264 to 720p H.264 Jellyfin hardware transcode without thermal, playback, or operational failure.

download the public machine-readable protocol JSON

reproducibility identity

protocol ID

protocol-jellyfin-1080p-h264-to-720p-h264-v1

protocol SHA-256

sha256:968366ca2b2265f0f41bf9ede1e0312b3ec3534d60797244ffc9426e817d0b74

workload ID

workload-jellyfin-1080p-transcode

application scope

Jellyfin Server 10.11.x · >=10.11.0 <10.12.0

deployment method

Jellyfin Docker container in a Debian KVM virtual machine

runtime chain

KVM virtual machine → Docker Engine on 64-bit Linux

A result belongs to this protocol revision only when its accepted manifest pins this exact hash. A matching title or version label is insufficient.

fixture gate

status

selected-hash-pinned

stable URL

https://repo.jellyfin.org/test-videos/SDR/AVC/Test%20Jellyfin%201080p%20AVC%203M.mp4

SHA-256

30f19d65b19f606f5352deb956843692cb6cee9a5bcfb43fce549f94e3bd29ce

fixture ID

fixture-jellyfin-1080p-avc-3m

exact bytes

11302222

license

CC BY-SA 4.0

accepted result possible

not in the current protocol state

The selected fixture is accepted only when its exact byte count and SHA-256 match before the run; no title-only substitution is allowed.

Test video courtesy of Gnattu, distributed by the Jellyfin Project under CC BY-SA 4.0.

controlled inputs

target

720p H.264

concurrent streams

1

warm-up

10 minutes

measurement window

30 minutes

repetitions

3

sample interval

2 seconds

Return to within 2 C of the recorded pre-run idle package temperature before the next repetition.

ordered test procedure

The order below is part of protocol version 1.0.0. Reordering, omitting, or silently changing a step requires a new protocol revision before a run can be accepted.

1. Identify the exact specimen and software chain. Record the configuration, lab asset, firmware, operating system, runtime, Jellyfin, FFmpeg, graphics driver, media runtime, storage, network, and instrument identities in manifest.json.
   required output: A complete secret-free manifest whose configuration and version fields match the intended run plan.
2. Verify the controlled media fixture. Download or select the declared fixture, record its byte count, and calculate SHA-256 before any repetition.
   required output: fixture.sha256 matches 30f19d65b19f606f5352deb956843692cb6cee9a5bcfb43fce549f94e3bd29ce and the file is exactly 11302222 bytes.
3. Normalize the environment. Disable undeclared background load, confirm wired links and storage topology, record ambient temperature, and return the machine to the declared clean pre-run state.
   required output: The manifest and instrument notes record the environment and confirm no undeclared workload is active.
4. Confirm hardware acceleration. Start Jellyfin, request the declared conversion, and inspect the playback and transcode diagnostics for active hardware acceleration instead of inferring it from the GPU model.
   required output: jellyfin-diagnostics.txt identifies the active hardware decode and encode path for the declared request.
5. Run the fixed warm-up. Run the declared 720p H.264 conversion for 10 minutes while instrumentation is active; do not include the warm-up interval in the principal measurement.
   required output: Timestamped diagnostics and metrics cover the warm-up and identify the measurement-window boundary.
6. Measure one controlled transcode. Continue one concurrent 1080p H.264 to 720p H.264 transcode for 30 minutes, sampling every required series at two-second intervals.
   required output: metrics.csv and Jellyfin diagnostics cover the entire measurement window with synchronized timestamps.
7. Cool down and repeat. Stop the workload, preserve every failed or excluded repetition, return package temperature to within 2 C of pre-run idle, and repeat until at least three completed repetitions exist.
   required output: At least three separately identified repetition records plus disclosed failed or excluded attempts.
8. Derive and cross-check the summary. Generate run-summary.json from raw series and logs, calculate the declared principal statistics and variance, and cross-check playback, thermal, application, and kernel failures.
   required output: A generated summary that agrees with raw artifacts and discloses variance, failures, exclusions, and limitations.
9. Apply the human acceptance gate. A human reviews the manifest, fixture identity, raw artifacts, generated summary, privacy redactions, acceptance checks, and exact claim scope.
   required output: An explicit human decision; automatic Grade-A promotion remains forbidden.

required manifest

• configuration identity and specimen ID
• BIOS or firmware version and relevant settings
• memory module and storage-device identities
• operating system, kernel, container or VM runtime
• Jellyfin and FFmpeg versions
• graphics driver and media runtime versions
• Jellyfin hardware-acceleration settings
• fixture URL and SHA-256
• ambient temperature and measurement instruments
• network topology
• command or request used to start the workload

raw series and artifacts

acceptance gate

At least 3 completed repetitions are required. A human must approve the record, and automatic promotion to Grade A is forbidden.

1. each repetition completes the measurement window
2. hardware acceleration is confirmed rather than inferred
3. transcode speed remains at or above 1.0x
4. the client reports no playback stall attributable to the server
5. no thermal shutdown, application crash, or kernel error occurs
6. raw series and logs agree with the human-readable summary
7. variance and every excluded run are disclosed

privacy and change control

• Real personal media is forbidden.
• Machine names must be pseudonymous.
• Secrets and private network addresses must be redacted.
• A material fixture, target, concurrency, duration, threshold, or instrumentation change creates a new protocol version.

accepted scoped results