16 Free OBS ffmpeg and Silero Tools for Playtest VOD Concat and Whisper Batch Prep - 2026

Your #playtest channel has nineteen Replay Buffer MKV files from Tuesday. You already read local Whisper VOD triage. You ran Whisper on clip seven and got garbage timestamps. ffmpeg -f concat printed Non-monotonous DTS and you blamed the model.

July 2026 facilitators need a bookmark stack for capture → probe → normalize → concat → VAD smoke → Whisper batch—not another generic AI tools mega-list. This Listicles & Resource Roundups article curates sixteen free tools that map to gates O1–O6 in your first OBS concat evening pipeline.

Non-repetition note: 18 playtest feedback tools owns survey + issue capture; 15 Free Local Whisper resources is the resource-page anchor list. This URL owns concat + batch prep keywords. 14 LUFS tools owns trailer loudness—not VOD merge.

Time to read: ~40 minutes. Minimal stack tonight: Tools 1, 2, 3, 5, 6, 12 (~75 minutes).

Why this matters now (June–July 2026)

1. Replay Buffer volume — Hotkey saves produce many MKVs, not one session MP4.
2. Whisper batch scripts — Assume one merged timeline; concat is the missing toolchain layer.
3. DTS / sample-rate chaos — Mixed 44100/48000 fragments are normal after OBS restarts.
4. Consent defaults — Local merge beats uploading twelve files to cloud ASR.
5. Help queue depth — Zero-duration audio and planned MKV concat help assume you know which tool does which step.

Direct answer: OBS (1) → ffprobe (2) → ffmpeg (3) → optional Silero VAD (5) → faster-whisper (6) → jq receipt (11). Log concat_ok before any ASR spend.

Who this is for

Beginner path (first playtest night)

1. Install Tools 1, 2, 3 (OBS, ffprobe, ffmpeg).
2. Follow concat beginner pipeline once.
3. Add Tool 6 (faster-whisper) next evening.
4. File Tool 11 (jq) receipt with concat_ok: true.
5. Only then run Whisper triage blog batch steps.

Developer path (fest season)

1. Pin ffmpeg build in README (Tool 3).
2. Automate Tool 2 CSV export in CI for facilitator laptops.
3. Tool 5 VAD gate on session_merged.wav.
4. Tool 14 pre-commit checks concat_ok in receipt JSON.
5. Attach merged transcript + receipt to BUILD_RECEIPT.

Selection criteria

Each tool is free for the described workflow, supports evidence-friendly outputs (logs, CSV, JSON), and fits capture → probe → merge → ASR prep → triage.

playtest_vod_triage_receipt_v1.json (concat fields)

{
  "schema": "playtest_vod_triage_receipt_v1",
  "build_id": "nextfest-oct-2026-rc4",
  "fragment_count": 12,
  "concat_ok": true,
  "merged_audio": "playtest-vod/inbox/2026-05-25_session-rc4/session_merged.wav",
  "tools_used": ["OBS", "ffprobe", "ffmpeg", "Silero VAD", "faster-whisper"],
  "normalize_profile": "48000_stereo_pcm_s16le"
}

Full schema lives in Whisper triage blog; this listicle documents which free tools populate each field.

Evidence folder layout

playtest-vod/
  obs-replay-profile.md
  inbox/2026-05-25_session-rc4/
    ffprobe_table.csv
    norm/
    session_merged.wav
    playtest_vod_triage_receipt_v1.json
release-evidence/06-playtest-vod/
  concat_log.txt
  transcript/session_merged.txt

Pair with Thursday row review when build_id must match receipts.

Stack diagram

OBS Replay Buffer (Tool 1)
        │
        v
   ffprobe table (Tool 2)
        │
        v
 ffmpeg normalize (Tool 3)
        │
        v
 ffmpeg concat → session_merged.wav
        │
        v
 Silero VAD smoke (Tool 5) ── optional
        │
        v
 faster-whisper batch (Tool 6)
        │
        v
 GitHub Issues triage (Tool 13)

The 16 tools

Tool 1 — OBS Studio (Replay Buffer)

Free capture hub for last N seconds hotkey saves—default facilitator capture in community playtest resources.

Beginner: Enable Replay Buffer 120–180 s; set numbered output folder per session.
Dev: Export profile JSON to obs-replay-profile.md.
Mistake: Desktop audio-only track—see zero-duration audio help.
Link: OBS Studio

Workflow detail: Lock MKV, Replay Buffer seconds, and audio track matrix before fest week. Never change buffer length mid-session without noting it in concat_log.txt. Pair with playtest isolation so build_id overlay matches receipt.

Tool 2 — ffprobe (ffmpeg suite)

Read-only stream inspector—O1 gate (audio exists, duration greater than 0).

ffprobe -hide_banner -show_streams -select_streams a:0 clip.mkv

Beginner: Build ffprobe_table.csv before merge.
Dev: Fail script if any row missing audio_stream.
Link: ffprobe documentation

Workflow detail: Log sample_rate, channels, codec_name, duration per fragment. Mixed 44100/48000 is not failure—Tool 3 normalizes. No audio stream is failure—exclude fragment or re-capture.

Tool 3 — ffmpeg (normalize + concat)

Swiss-army CLI for 48 kHz stereo WAV segments and concat demuxer merge.

ffmpeg -y -i clip.mkv -vn -ac 2 -ar 48000 -c:a pcm_s16le norm_001.wav
ffmpeg -y -f concat -safe 0 -i concat_list.txt -c:a pcm_s16le session_merged.wav

Beginner: Copy commands from concat pipeline blog.
Dev: Pin ffmpeg 6.x+ in facilitator README.
Mistake: Raw MKV concat without normalize—Non-monotonous DTS.
Link: ffmpeg concat

Workflow detail: Always normalize per fragment before concat list. Re-encode concat if -c copy fails. Archive concat_log.txt with ffmpeg version string.

Tool 4 — VLC media player (free)

Quick human verification of each MKV before batch processing—catch black video / frozen frames facilitators misremember.

Beginner: Skim 5 s per clip; mark “menu idle” vs “combat” in spreadsheet.
Dev: Not for automation—feeds O5 listen smoke alongside Tool 8.
Link: VideoLAN VLC

Workflow detail: Use VLC when Discord report says “clip broken” but ffprobe shows audio—often wrong file attached. Rename after VLC check to fix 001_ ordering mistakes.

Tool 5 — Silero VAD (open source)

Voice activity detection—optional O5b gate proving merged WAV contains speech spans before GPU hours on Whisper.

Beginner: Run sample script on session_merged.wav; if zero speech spans over 10 min, skip ASR and tag “silent menu capture.”
Dev: Integrate silero-vad pip package in scripts/vad_smoke.py.
Link: Silero VAD GitHub

Workflow detail: VAD does not replace ffprobe—silent video with noise floor can fool naive VAD. Combine with facilitator notes. Pairs with local CUDA fallback help when ASR runs but text empty.

Tool 6 — faster-whisper (local ASR)

CTranslate2 Whisper implementation—batch transcribe after concat_ok.

Beginner: small or medium model on laptop GPU; one merged file per session.
Dev: device=cuda, compute_type=float16; batch size tuned to VRAM.
Mistake: Per-clip ASR without offsets when concat was skipped.
Link: faster-whisper GitHub

Workflow detail: Extract 16 kHz mono from session_merged.wav in one ffmpeg pass—do not resample each fragment thrice. Log model, language, duration_processed in receipt extension fields.

Tool 7 — OpenAI Whisper API (cloud fallback)

Paid API lane when local GPU unavailable or consent allows cloud—chunk merged WAV per 413 help.

Beginner: Use only after concat; never upload twelve MKVs separately without merge plan.
Dev: Exponential backoff; merge verbose_json segments with offset math.
Link: OpenAI speech-to-text docs

Workflow detail: Cloud cost scales with unmerged carelessness—concat first, chunk second. Cross-link 15-free API chunking resource.

Tool 8 — Audacity (free DAW)

O5 listen smoke—headphone check for chipmunk, gaps, or duplicate segments after concat.

Beginner: Open session_merged.wav; view waveform for obvious flatlines.
Dev: Export 30 s sample for Discord proof when disputing facilitator capture quality.
Link: Audacity

Workflow detail: Not for precision LUFS here—that is 14 LUFS listicle. Audacity answers “does merged audio sound like one session?”

Tool 9 — ShareX (Windows capture alternative)

Free hotkey capture when OBS Replay Buffer unavailable on facilitator laptop—export MP4/MKV into same inbox/ rules.

Beginner: Match numbered naming convention from concat pipeline.
Dev: Document codec in ffprobe_table.csv—ShareX may differ from OBS defaults.
Link: ShareX

Workflow detail: Do not mix ShareX and OBS fragments in one concat without noting codec in log—normalize step still saves you.

Tool 10 — MediaInfo (GUI or CLI)

Second opinion on container metadata when ffprobe output confuses beginners.

Beginner: Paste MediaInfo text into session folder README.md.
Dev: Redundant with ffprobe for automation—pick one in scripts.
Link: MediaInfo

Workflow detail: Use when facilitator insists “OBS broken” but MediaInfo shows AAC audio—education moment before re-install spirals.

Tool 11 — jq (JSON CLI)

Validate and patch playtest_vod_triage_receipt_v1.json in shell scripts.

jq -e '.concat_ok == true' playtest_vod_triage_receipt_v1.json

Beginner: Install once; copy one-liner into Friday checklist.
Dev: Gate Whisper job in CI on concat_ok.
Link: jq manual

Workflow detail: Pair with validate-packet culture—receipts are machine-readable promises.

Tool 12 — Python 3.10+ (stdlib + scripts)

Glue ffprobe CSV, concat lists, and receipt writing—pathlib, subprocess, json.

Beginner: Run provided sketch from concat pipeline blog.
Dev: Package playtest_merge module in repo scripts/.
Link: python.org

Workflow detail: Pin minor version in facilitator README—3.10 vs 3.12 rarely matters for subprocess glue.

Tool 13 — GitHub Issues (free tier)

Triage board for transcript summaries—labels playtest, build_id, hour-one.

Beginner: One issue per merged session summary.
Dev: Template requires concat_ok screenshot in issue body.
Link: GitHub Issues docs

Workflow detail: Link issue URL in Thursday row review notes when triage blocks promotion.

Tool 14 — Visual Studio Code (free editor)

Edit concat_list.txt, receipts, and facilitator README with folder workspace.

Beginner: Open playtest-vod/ as multi-root workspace.
Dev: Tasks.json runs ffmpeg normalize task.
Link: VS Code

Workflow detail: .vscode/settings.json excludes norm/ from search—thousands of WAVs slow indexer.

Tool 15 — mkvtoolnix / mkvmerge (optional)

Remux MKV without re-encode when video must be archived but audio alone goes to Whisper.

Beginner: Skip until you need video evidence for crash reports.
Dev: mkvextract audio lane when video corrupt but audio stream valid.
Link: MKVToolNix

Workflow detail: Whisper path only needs audio—Tool 3 -vn is default. mkvtoolnix helps “send video to engineer, audio to ASR.”

Tool 16 — GamineAI concat + triage guides (internal anchors)

Bookmark procedure, not a downloadable binary:

• OBS concat beginner pipeline
• Local Whisper triage
• 15 Free Local Whisper ffmpeg resources
• Planned MKV concat help

Pro tip: Tool 16 rows are your studio SOP—external tools fail when steps are skipped, not when OBS updates.

Minimal vs full stack

Tool pairing matrix

Windows vs Linux facilitators

Document OS in receipt notes for Asia-EU handoff.

Facilitator one-pager (paste in Discord pin)

Playtest capture stack 2026:
1. OBS Replay Buffer → numbered MKVs per build_id
2. ffprobe table (all clips)
3. ffmpeg → session_merged.wav
4. concat_ok in receipt JSON
5. faster-whisper → summary issue
Full list: /blog/16-free-obs-ffmpeg-silero-playtest-vod-concat-whisper-batch-tools-2026

Compare to 18 playtest tools listicle

Use both bookmarks—different search intents.

Install order (first laptop, 45 minutes)

Do not install faster-whisper before ffmpeg—beginners transcribe unmerged clips and blame Whisper.

Receipt columns facilitators forget

Add columns to GitHub Issue template (Tool 13) same day you adopt Tool 11.

GPU vs CPU decision (Tool 6)

Receipt notes should record device class—explains slow Tuesday without reopening ASR debate.

When to escalate to Help articles

Listicle bookmarks tools; help pages own failure fixes.

Multi-facilitator studio policy (producer)

1. One obs-replay-profile.md in repo—no per-person mystery settings.
2. Shared playtest-vod/inbox/ naming rules in Tool 16 links.
3. Weekly rotation who runs Tool 3 concat (not always engineer).
4. Transcripts public to team, raw VODs private per consent README.
5. Multi-channel facilitator contract when backlog ships—extends Tools 1–13 across GX + Steam + itch.

Search keywords this listicle targets

• free OBS Replay Buffer merge tools 2026
• ffmpeg concat playtest VOD
• Silero VAD before Whisper
• playtest transcript batch prep indie
• concat_ok receipt json

Use naturally in Issues and README—not keyword stuffing in Discord pins.

Key takeaways

• Sixteen free tools cover OBS capture → ffprobe → ffmpeg merge → VAD → Whisper → triage.
• Run concat before Whisper; set concat_ok in playtest_vod_triage_receipt_v1.json.
• Normalize fragments to 48 kHz WAV before concat demuxer.
• Concat beginner pipeline is the hands-on companion to this bookmark list.
• 18 playtest tools + this list = capture + merge + feedback ops.
• Silero VAD (Tool 5) saves GPU when merged audio is menu-silent.
• faster-whisper (Tool 6) batch runs on one session_merged.wav.
• jq (Tool 11) gates automation on concat_ok.
• Cloud Whisper (Tool 7) is fallback—merge first.
• Pairs Help #5 MKV concat fix when published.

FAQ

Is this the same as the Local Whisper resource page?

No. 15-free local Whisper resource is curated links for resource SEO. This blog listicle targets concat + batch prep search intent with O1–O6 mapping.

How many tools do I need day one?

Four: OBS, ffprobe, ffmpeg, jq—add Whisper tomorrow.

Can I skip Silero?

Yes for week one. Add when you waste GPU on silent menu merges.

Does OBS Replay Buffer replace full session record?

No—Replay is for highlights. Full record optional for crash forensics (Tool 15).

What about Mac facilitators?

OBS + ffmpeg + ffprobe work; ShareX is Windows—use OBS export paths.

Where does BUILD_RECEIPT fit?

Same build_id on session folder, receipt, and Issues—BUILD_RECEIPT tutorial.

Next backlog article?

When ffmpeg concat fails decision tree (planned)—read after Tools 3 and 7 here.