SDR to HDR: the FFmpeg Alternative That Scales

Why this matters

FFmpeg's zscale + tonemap path applies a global curve: shadows crush or highlights clip depending on the scene. It is fast and free, and the results show it.

A trained inverse-tone-mapping model analyzes each scene's luminance histogram and rebuilds dynamic range per scene, which is why the output holds highlight and shadow detail.

How it works

If a one-line ffmpeg command covers your quality bar, keep it. The API is for when you need per-scene quality at volume without GPU farms.

Submit a job with an input URL (S3, GCS or HTTPS), poll the job URL, and download the rendered output. No GPU, no queues, no ffmpeg builds to babysit.

Code

import requests, time

API = "https://api.mlslabs.io/v1/sdr2hdr/jobs"
headers = {"X-API-Key": "YOUR_API_KEY"}
payload = {"input": "s3://bucket/video.mp4", "preset": "cinema"}

resp = requests.post(API, json=payload, headers=headers)
job = resp.json()
while job["status"] not in ("succeeded", "failed"):
    time.sleep(3)
    job = requests.get(job["url"], headers=headers).json()

print("Output:", job["output_url"])

Pro tips

• Test both on the same 3 clips before deciding — look at specular highlights and skin tones.
• ffmpeg is great for burn-in previews; use the API for the actual deliverable.
• License-wise, HDR conversion has no codec patent baggage — the API handles encoding separately.

Try the SDR2HDR collection — the honest A/B is 10 minutes.