01 · ffmpeg-cli-reference
Audio Joiner
Ordered tones, joined with an attenuated boundary.
- Output
- 2.25 s
- First tone
- 440 Hz
- Second tone
- 660 Hz
- Boundary RMS
- 0.004157 FS
Conversion Lab
Inspect the evidence, not just the claim.
Reproducible PDF benchmarks and audio reference findings, published with synthetic inputs, measured outputs, source provenance, hashes, and explicit limitations.
Audio v1 reference corpus
Four workflows, measured against deterministic signals.
These results come from recorded FFmpeg CLI reference commands over CC0 synthetic PCM fixtures. They test duration, frequency, join-boundary attenuation, echo-tail energy, and sample alignment. They do not establish performance or fidelity for every source file or device.
- Generated
- September 3, 2026
- Corpus
- audio-v1
- Source commit
- e11eef48f52df732e0ef32d69f7ea10635b150bc
02 · ffmpeg-cli-reference
Audio Speed Changer
A 0.75x tempo reference with pitch preservation.
- Input
- 2.4000 s
- Output
- 3.1934 s
- Dominant tone
- 523.2 Hz
03 · ffmpeg-cli-reference
Slowed and Reverb
Tape-style slowdown with a measurable echo tail.
- Input
- 2.0000 s
- Output
- 3.0625 s
- Dominant tone
- 264 Hz
- Echo-tail RMS
- 0.031474 FS
04 · ffmpeg-cli-edit-reference
Voice Recorder edit
A deterministic trim of a synthetic voice-like signal.
- Input
- 2.2000 s
- Output
- 1.8 s
- Alignment RMS
- 0 FS
FFmpeg CLI reference
Deterministic commands and signals establish the published reference outcomes above.
Browser release gate
Chromium, Firefox, and Playwright WebKit run the shipped workflow as a separate exact-candidate gate. That evidence does not certify Safari or archive the built artifact.
User-device behavior
Hardware, memory, source media, browser versions, and permissions vary. Users should review important exports on their own device.
Review status and limitations
Automated evidence contracts passed when this corpus was generated. Independent technical review and native-language review remain pending. The corpus is synthetic, narrow, and designed to expose specific semantic outcomes. It does not demonstrate arbitrary-source fidelity, cross-device equivalence, browser speed, microphone capture, or room acoustics.
PDF benchmark corpus
Shipped converter functions on seeded synthetic PDFs.
FileMorf's local PDF tools use pdf-lib and JSZip. The benchmark harness calls the shipped converter functions against a fixed corpus so releases remain comparable. Reference timings were produced in a server-class JavaScript runtime, not on a representative phone or laptop.
Last measured July 28, 2026 · build 4ef6888
Merging
| Benchmark | Input | Median | p95 | Output |
|---|---|---|---|---|
Merge 2 PDFs mergePdfs() over 2 three-page documents. | 2 PDFs × 3 pages | 2.4 ms | 5.7 ms | 9.95 KB |
Merge 10 PDFs mergePdfs() over 10 three-page documents. | 10 PDFs × 3 pages | 6.3 ms | 9.1 ms | 45.94 KB |
Merge 50 PDFs mergePdfs() over 50 three-page documents. | 50 PDFs × 3 pages | 30.4 ms | 32.7 ms | 226.95 KB |
Splitting & extraction
| Benchmark | Input | Median | p95 | Output |
|---|---|---|---|---|
Split a 20-page PDF into single pages splitPdf() with no range — one output document per page. | 1 PDF × 20 pages | 5.9 ms | 8.1 ms | 45.16 KB · 20 files |
Extract pages 1–5 from a 100-page PDF extractPages() with the range "1-5". | 1 PDF × 100 pages | 7 ms | 7.3 ms | 8.65 KB |
Page tools
| Benchmark | Input | Median | p95 | Output |
|---|---|---|---|---|
Number every page of a 100-page PDF addPageNumbers() with the "page n of total" format. | 1 PDF × 100 pages | 32.3 ms | 36.9 ms | 177.4 KB |
Watermark every page of a 100-page PDF watermarkPdf() drawing diagonal "DRAFT" text on each page. | 1 PDF × 100 pages | 29.9 ms | 31.5 ms | 183.16 KB |
Packaging
| Benchmark | Input | Median | p95 | Output |
|---|---|---|---|---|
Package 50 PDFs into a ZIP JSZip generateAsync with the same defaults the site's batch download uses. | 50 PDFs × 3 pages | 5.5 ms | 10.1 ms | 277.63 KB |
PDF methodology
- Each benchmark runs 1 warm-up pass, then 7 timed runs. The table reports the median and 95th percentile.
- The code under test is
src/lib/converters/pdf.ts, with pdf-lib 1.17.1 and JSZip 3.10.1. - The corpus contains seeded pseudo-random text, no customer or third-party files. Every file's SHA-256 is in the manifest.
- Reference environment: Node.js v23.7.0 on Apple M2 Pro (arm64, 10 cores).
PDF limitations
- User devices run the same PDF functions, but these timings come from reference hardware and should not be read as a promise for a particular browser or device.
- The corpus is text-only. Scanned, image-heavy, encrypted, unusually large, or malformed documents can behave differently.
- Canvas image conversion and interactive browser timings are outside this PDF timing table. Audio reference findings and the separate browser release gate are documented above.
Download and inspect
The measurements, fixtures, manifests, and generator are public. Use the hashes in the manifests to verify local copies before comparison.
Audio evidence
- evidence.json (complete versioned manifest)
- audio-v1-findings.csv (2.96 KB, SHA-256 in the manifest)
- audio-v1-duration.svg (3.3 KB, SHA-256 in the manifest)
- audio-v1-frequency.svg (3.29 KB, SHA-256 in the manifest)
- audio-v1-signal-checks.svg (2.96 KB, SHA-256 in the manifest)
- Reference generator snapshot
- CC0 fixture license
Show all WAV fixtures
- join-a-440hz.wav (172.31 KB)
- join-b-660hz.wav (215.38 KB)
- joined-reference.wav (387.64 KB)
- speed-metronome-523hz.wav (413.48 KB)
- speed-075-keep-pitch-reference.wav (550.16 KB)
- slowed-impulse-330hz.wav (344.57 KB)
- slowed-080-hall-reference.wav (527.61 KB)
- synthetic-voice-like-source.wav (379.03 KB)
- voice-edit-trim-reference.wav (310.12 KB)
PDF benchmark data
- results.json (schema
filemorf-lab-results/1, CC BY 4.0) - corpus-manifest.json (page count, bytes, seed, and SHA-256)
Found a methodology issue or want a correction? Email support@filemorf.com. The harness and source are available at github.com/jddelia/filemorf-lab. Product context is on the about page.