If you've ever tried to open a broadcast caption file in a video editor and it looked like scrambled coordinates, you've met SCC. Here's what it is, why it exists, and when converting to SRT is the right call.
| Aspect | SCC (Scenarist Closed Captions) | SRT (SubRip) |
|---|---|---|
| Origin | DVD/ATSC broadcast authoring (Sonic Solutions Scenarist) | SubRip DVD-ripping tool, late 1990s; adopted by YouTube, VLC, OTT |
| Timecode base | Video frame count (HH:MM:SS:FF or HH:MM:SS;FF) | Wall-clock milliseconds (HH:MM:SS,mmm) |
| DF/NDF | Yes — DF (29.97 drop) and NDF (30 non-drop) are explicit | No — milliseconds don't model 29.97 drop, conversions can drift |
| Character encoding | Hex bytes (8-bit chars, often with an offset like 0x80/0x90) | Plain UTF-8 text |
| File shape | Compact, binary-ish, ~1–2 KB/min | Plain text, ~2–4 KB/min |
| Stack it lives in | Broadcast playout servers, MPEG-2 mux, FCC captioning pipelines | Web players (HTML5 <track>), YouTube, Vimeo, NLEs, social platforms |
| Editing by hand | Painful — you read hex | Trivial — it's text |
| Standards body | ATSC A/53, CEA-608/708 lineage | De-facto; no formal spec |
If you've only ever worked in NDF (non-drop-frame, 30 fps integer math), DF feels like a tax. But NTSC video is 29.97 frames per second — not 30. After one hour of 29.97 footage, naive 30 fps timecode is 3.6 seconds ahead of real time. Drop-frame timecode fixes that by skipping frame numbers 00 and 01 at the start of every minute except minutes divisible by 10 (which is why you see 00:01:00;02 jump straight to 00:01:01;02 on a DF counter).
Pricing snapshot for context. Numbers below are what we could verify on 2026-09-25; treat anything you can't reconcile against a current login screen as a planning estimate.
| Vendor | Free tier | Cheapest paid | Billing model | Source |
|---|---|---|---|---|
| Sonix | 30 min trial, no card | $10/hr pay-as-you-go; Core $25/mo (~$22.9/mo annual) | Per transcribed hour; SRT/VTT included | Official pricing page |
| HappyScribe | 10 min trial; free export bears a watermark | Lite: €6/mo (Capterra) / $8.50/mo annual (itqlick) / $9/mo (eesel) | Monthly import minutes; PAYG by minute | Aggregators (official page 403) |
| TurboScribe | 3 files × 30 min/day (≈90 min), signup required | Unlimited $10/mo annual / $20/mo monthly | Subscription + free daily allowance | Site search snippets + aggregators |
| VEED | Free with watermark, 10-min single-export cap, ~120 credits | Lite ≈ $19–24/mo; annual ≈ $12/mo | Subscription + AI credits | Third-party reviews (official page timeout) |
Pricing checked September 2026. Sonix = official direct read. HappyScribe, TurboScribe, VEED = aggregated/third-party; flagged where applicable.
Sonix's strength is transcription accuracy on long-form audio, and SRT/VTT export is built into every paid tier. The catch is the billing unit: it's per transcribed hour, not per file. For a podcaster dumping 40 one-hour episodes a month, the math is clean. For someone who just wants to convert a single 90-minute SCC archive to SRT, the $10/hr PAYG rate is fine, but you're paying for the engine, not the format conversion.
Best for: heavy monthly volume, teams that want timestamps and diarization alongside subtitles. Not ideal for: one-off conversion of a broadcast archive.
HappyScribe sits in the "I have a tight budget and a small queue" zone — every aggregator we checked put their entry tier somewhere in the €6–$9 range once you annualize. The free trial is genuinely short (10 minutes), and the basic free export path carries a watermark, which is the usual trade. Aggregator data is messy because EUR/USD conversions, stale listings, and promo codes don't agree; if you're about to subscribe, log in and confirm the current price against your currency before you commit.
Best for: European users who want a low monthly ceiling and don't mind paying per minute on overflow. Not ideal for: anyone who needs a no-signup, no-watermark free path.
TurboScribe is the daily-allowance play: 90 free minutes, no card, but you do need an account. Above that, an Unlimited subscription at $10/mo (annual) is one of the cheapest seats on this list. There's no per-file fee — if you've seen "$4.99 per file + $1 per 30 min" attached to TurboScribe somewhere, that's actually Videotobe, not TurboScribe; the conflation shows up often in comparison blogs.
Best for: light daily users who'd burn through 90 free minutes on weekdays and only occasionally hit the cap. Not ideal for: someone who wants to drop one 4-hour archive a month without a subscription.
VEED is a video editor first, subtitle tool second. The free tier exists, but every export bears a watermark and a 10-minute single-clip ceiling — which makes "free conversion" feel more like a demo than a deliverable. Paid tiers lift the watermark and the resolution cap, and add AI credit pools for translation and generation. Pricing on aggregator reviews ranges roughly $19–$24 monthly / $12 monthly on annual — likely reflecting in-year promo movement.
Best for: creators who need subtitles as part of a broader video edit. Not ideal for: a clean, watermark-free SCC-to-SRT pass.
Scenarist_SCC V1.0 and a 00: timestamp line; the semicolon separator signals NDF, the colon separator signals DF.HH:MM:SS,mmm --> HH:MM:SS,mmm, UTF-8 plain text.Yes — for U.S. broadcast (ATSC) delivery and for FCC captioning compliance on cable, SCC and its CEA-608 lineage remain the on-air standard. SRT is for everything upstream of the broadcast encoder.
.scc to .srt?No. The format is entirely different — binary hex payload vs plain text. Renaming will produce a broken file that no player will parse.
SRT carries only millisecond time. It doesn't model DF/NDF directly — that's the converter's job. A correct converter preserves real wall-clock time regardless of source frame rate.
MCC is MacCaption's authoring cousin of SCC (same CEA-608 logic, slightly different header). CEA-708 is the HD successor and lives in a different container (typically .mcc or embedded in MPEG user data). If you're working in HD, the question is usually "do I need 708 or 608-within-708," not "SCC vs SRT."
Two usual causes: (1) the converter treated a DF SCC as NDF or vice versa; (2) it rounded frame-to-ms incorrectly at hour boundaries. A correct converter chooses DF/NDF explicitly and computes fractional ms at 29.97, not 30.00.
For a one-off or batch job, yes — browser-side converters that decode the SCC locally and hand you a plain SRT do exist and don't add platform watermarks to your cues. For ongoing heavy transcription (not just format conversion), the SaaS tools above are the realistic choice.