Back Page RF Capture Decoding Guide
Forward Page Post Processing
Sub-Page Jitterbugs TBC-Video-Export Releases Page
This is the part ware you take your luma & chroma data to the ld-chroma-decoder
which combines it and then decodes it to a RGB or YUV stream with FFmpeg
using the merge planes function, resulting in a final colour video stream that is encoded to FFV1 or to any other FFmpeg supported codec and container output format you wish to use.
You can add your own encoding profiles in the tbc-video-export.json
file, which when present will be the profile options available to the export tool.
Key Secondary Guide Read Pre-Export LD-Analyse Guide Here
Colour-Under formats VHS, Umatic, BetaMax, Video8, Hi8
etc are a .tbc
/_chroma.tbc
file (Y+C) set.
CVBS formats Composite, LaserDisc, SMPTE 1" Type C
are 1 single .tbc
file.
Always backup & keep your .log
& .json
files even if you delete your TBC files after you are happy with your export these still contain key metadata.
Tip
The output name and directory will the the same as the input unless manually defined by the user.
Linux & MacOS
tbc-video-export Input-Media.tbc
Windows
tbc-video-export.exe Input-Media.tbc
The export tool takes filename.tbc
/ filename_chroma.tbc
and metadata from the .json
and create video files automatically for PAL or NTSC it can also timecode based off VITC data if present.
The default output format is:
Aspect Ratio 4:3
760x488 NTSC
or 928x576 PAL
Interlaced Top Field First
Video Codec: FFV1 10-bit 4:2:2
(70-100 Mb/s)Audio Codec: FLAC
Ideal starting point for post processing footage without VBI data, but will require de-interlacing for online use.
Defining input location and output location is useful if your data is not inside the same folder as the export tool.
For Linux/MacOS based systems this is like:
/home/harry/desktop/vhs-decode/Input.tbc
For Windows this looks like
"C:\Users\harry\Desktop\Decode\Input.tbc"
--skip
, -s
Skips number of frames -s 25
for example skips 1 second of PAL video.
--length
, -l
Defines length to export so -l 1500
is 1 minute of PAL video at 25fps.
On tape Audio can be Mono
, Stereo
, Linear Timecode
You have 4 possible channels, 2 Linear
& 2 HiFi FM
This is format and tape to tape dependent as Sony 8mm can have PCM digital for example.
LaserDisc has 2 Analog channels which one can be populated with AC-3
or DTS
digital this can be 4.0 or 5.1 surround.
Your options are quite simple:
[!WARNING] This works only for clock synchronised data this means Video8/Hi8/Betamax NTSC/LaserDisc captures can work with AAA directly, but VHS/Betamax PAL etc will require a clockgen setup.
After decoding your Video RF and or HiFi FM audio you will want to sync your PCM audio data with your video file
With a standard configuration you just need to alter the following:
-i baseband.wave
- Input PCM Audio File--json capture.tbc.json
- Input TBC JSON from decodingbaseband_aligned.flac
- Output name at the end (also s16 will do 16-bit if s24 is replaced)--stream-sample-rate-hz
- Input rate of your audio file i.e 46875 for clockgen or 48000 for hifi-decode or 44100 for ld-decode.--rf-video-sample-rate-hz
capture rate of your base file this normally will always be 40000000 or 40mspsLinux & MacOS[!TIP]
Below is premade commands for usage:
"ffmpeg" -i baseband.wav -filter 'channelmap=map=FL-FL|FR-FR' -f s24le -ac 2 - | mono VhsDecodeAutoAudioAlign.exe stream-align --sample-size-bytes 6 --stream-sample-rate-hz 46875 --json capture.tbc.json --rf-video-sample-rate-hz 40000000 | ffmpeg -f s24le -ar 46875 -ac 2 -i - -af aresample=48000 -sample_fmt s32 baseband_aligned.flac
"ffmpeg" -i baseband.wav -filter_complex "[0:a]channelsplit=channel_layout=2.1:channels=FL[FL]" -map "[FL]" -f s24le -ac 1 - | mono VhsDecodeAutoAudioAlign.exe stream-align --sample-size-bytes 3 --stream-sample-rate-hz 46875 --json capture.tbc.json --rf-video-sample-rate-hz 40000000 | ffmpeg -f s24le -ar 46875 -ac 1 -i - -af aresample=48000 -sample_fmt s32 baseband_aligned.flac
ffmpeg.exe -i baseband.wav -filter_complex "channelmap=map=FL-FL|FR-FR" -f s24le -ac 2 - | VhsDecodeAutoAudioAlign.exe stream-align --sample-size-bytes 6 --stream-sample-rate-hz 46875 --json capture.tbc.json --rf-video-sample-rate-hz 40000000 | ffmpeg.exe -f s24le -ar 46875 -ac 2 -i - -af aresample=48000 -sample_fmt s32 baseband_aligned.flac
ffmpeg.exe -i baseband.wav -filter_complex "[0:a]channelsplit=channel_layout=2.1:channels=FL[FL]" -map "[FL]" -f s24le -ac 1 - | VhsDecodeAutoAudioAlign.exe stream-align --sample-size-bytes 3 --stream-sample-rate-hz 46875 --json capture.tbc.json --rf-video-sample-rate-hz 40000000 | ffmpeg.exe -f s24le -ar 46875 -ac 1 -i - -af aresample=48000 -sample_fmt s32 baseband_aligned.flac
To embed an audio file use --audio-track
Example:
--audio-track HiFi_24-bit_48khz.flac
--audio-track Linear_24-bit_48khz.flac
To set the input file name, track name and language flags use --audio-track-advanced
Good to read: FFmpeg Audio Channel Mapping.
Language of the audio track i.g en
, jpn
, cn
, de
etc
[file_name, title, language, rate, format, channels, channel_layout, time_offset]
Example:
--audio-track-advanced '["tests/files/HiFi_24-bit_48khz.flac", "HiFi", "eng", "48000", "2", "00:00:00.000"]'
Example Basic
tbc-video-export --audio-track HiFi_24-bit_48khz.flac --audio-track Linear_24-bit_48khz.flac
Example Advanced
tbc-video-export Input.tbc --audio-track-advanced '["tests/files/HiFi_24-bit_48khz.flac", "HiFi", "eng"]' --audio-track-advanced '["tests/files/HiFi_24-bit_48khz.flac", "Linear", "eng"]' output-name
--process-efm
Decodes 4.0 or 5.1 AC3 to file
--process-efm-dts
Decodes 5.1 DTS to file.
Caution
Progressive video files will be 2x the size of interlaced files.
FFmpeg profile based de-interlacing can be done with bwdif
& w3dif
(see the tbc-video-export.json formatting for how to apply it to new profiles) for 50p/59.94p progressive files although QTGMC is recommended for the best results today.
We have implemented ProRes & PCM audio modes for better support in NLEs such as Davinchi Resolve.
Note
You can make/share you own FFmpeg profiles by editing the tbc-video-export.json
as needed, however it may need to be edited every major update.
Define your profile with for example: --profile ffv1_8bit_pcm
Standard Profiles:
Profile Name Codec Compression Type Bit-Depth Chroma Sub-Sampling Audio Format Container File Extension Bitrate ffv1 FFV1 Lossless Compressed 10-bit 4:2:2 FLAC Audio Matroska .mkv 70-100mbps ffv1_8bit FFV1 Lossless Compressed 8-bit 4:2:2 FLAC Audio Matroska .mkv 40-60mbps ffv1_pcm FFV1 Lossless Compressed 10-bit 4:2:2 PCM Audio Matroska .mkv 70-100mbps ffv1_8bit_pcm FFV1 Lossless Compressed 8-bit 4:2:2 PCM Audio Matroska .mkv 40-60mbpsEditing Profiles:
Profile Name Codec Compression Type Bit-Depth Chroma Sub-Sampling Audio Format Container File Extension Bitrate prores_hq_422 ProRes HQ Compressed 10-bit 4:2:2 PCM Audio QuickTime .mov 55-70mbps prores_4444xq ProRes 4444XQ Compressed 10-bit 4:4:4 PCM Audio QuickTime .mov 80-110mbps v210 V210 Uncompressed 10-bit 4:2:2 PCM Audio QuickTime .mov 200mbps v410 V410 Uncompressed 10-bit 4:4:4 PCM Audio QuickTime .mov 400mbpsWeb Profiles:
Profile Name Codec Compression Type Bit-Depth Chroma Sub-Sampling Audio Format Container File Extension Bitrate x264_web AVC/H.264 Lossy 8-bit 4:2:0 AAC Audio QuickTime .mov 8mbps x265_web HEVC/H.265 Lossy 8-bit 4:2:0 AAC Audio QuickTime .mov 8mbpsBroadcast Profiles:
Profile Name Codec Compression Type Bit-Depth Chroma Sub-Sampling Audio Format Container File Extension Bitrate D10 PAL MPEG2 Lossy 8-bit 4:2:2 PCM Audio Material eXchange Format .mxf 50mbps D10 NTSC MPEG2 Lossy 8-bit 4:2:2 PCM Audio Material eXchange Format .mxf 50mbps[!NOTE] The D10 or IMX50 standard is 720x512 NTSC & 720x608 PAL (SD + 32 lines for VBI area) standard for SD archive delivery and used by the BBC archive and many other broadcasters.
Warning
You probably missed the dropdown table tab directly above, this is very handy to read & reference!
Adjust values in ld-analyse to have your rough numbers
--chroma-gain
Chroma Level (Gain) factor applied to chroma components. 0 to 2
(Default 1
)
--chroma-phase
Phase rotation applied to chroma components (degrees) -180 to 180
(Default 0
)
--luma-nr
Luma noise reduction level in dB. (decimal range) 0 to 10
(Default 0
)
--transform-threshold
Transform PAL chroma sensitivity (decimal range) 0 to 1
(Default 0.4
)
Select Chroma decoder to use.
--chroma-decoder decoder
Defaults:
PAL:
PAL2D
(default for PAL/PAL-M S-Video)transform3d
(default for PAL/PAL-M CVBS)NTSC:
ntsc2d
(default for NTSC S-Video and CVBS)ntsc3d
(default for NTSC CVBS LD)Available Decoders:
Monochrome: mono
PAL: pal2d
, transform2d
, transform3d
NTSC: ntsc1d
, ntsc2d
, ntsc3d
, ntsc3dnoadapt
--export-metadata
Run ld-export-metadata before export. (default: no)
Note
The generated subtitles (closed captions) and ffmetadata will be used when encoding.
This is data such as Title
, Year
, Camara
, Tape Tape
and so on can be added to the final video files metadata.
Example:
--metadata [{"title": "Birthday 2008", "JVC-GX250": "LP"}, {"year": 2008}]
Too add a custom or standard field to the metadata you can do as such for each article manually:
--metadata "Video System" "PAL"
& -metadata title="my title"
Full Example:
tbc-video-export Munday-Walking-Around-Garden.tbc --metadata "Tape Format" "VHS-C" --metadata "Video System" "PAL" --metadata "Tape Mode" "SP" --metadata "Title" "Munday Walking Around Garden" --metadata "Date" "2009-07-24" --metadata "Capture Date" "2022-09-08"
Resulting Media Info Metadata Readout:
General
Unique ID : 189262425709654888565939284128344460936 (0x8E629E0783A839695528477CEDD64E88)
Complete name : Munday-Walking-Around-Garden.mkv
Format : Matroska
Format version : Version 4
File size : 386 MiB
Duration : 30 s 160 ms
Overall bit rate mode : Variable
Overall bit rate : 107 Mb/s
Frame rate : 25.000 FPS
Movie name : Munday Walking Around Garden
Writing application : Lavf60.3.100
Writing library : Lavf60.3.100
ErrorDetectionType : Per level 1
Attachments : Munday-Walking-Around-Garden.tbc.json
CAPTURE_DATE : 2022-09-08
DATE : 2009-07-24
TAPE_FORMAT : VHS-C
TAPE_MODE : SP
Time code of first frame : 00:00:00:00
Time code source : Matroska tags
VIDEO_SYSTEM : PAL
Standard Export (4:3 Standard)
tbc-video-export input.tbc
Standard 4:3 760x488 NTSC
& 928x576 PAL
Makes a Monochrome or Black & White Image using the Y channel or strips the colour carrier from a CVBS TBC.
16:9 Widescreen Letterbox Crop 928x432 PAL
& 760x432 NTSC
These commands work well for 16:9 widescreen letterbox crop tapes rendering only the centre making a ideal image for scaling/de-interlacing to 720p 50/59.94p for progressive Blu-Ray mastering.
Note
This produces an full-vertical image frame allowing you to visually see VBI data, and syncs however, if your image is not centred properly this may crop the data so ensure you adjust the active area in-ld-analyse before exporting.
Scaled 720x608 PAL
& 720x508 NTSC
(IMX/D10)
Unscaled 928x608 PAL
& 760x508 NTSC
includes top and bottom sync, useful for cropping and conforming in post rather than in initial export.
Full 4fsc Signal Frame
Note
This has to be manually done via editing the .JSON
files video area values, in a text editor, below are the fields you need to find and change to apply this effect.
625-line "activeVideoStart":20,"activeVideoEnd":1120}
525-line "activeVideoStart":0,"activeVideoEnd":910}
This option passes the TBC files to FFmpeg directly and exports only luma information.
Composite (Live CVBS Source)
S-Video (Y with C Ignored) (VHS PAL SP Source)
Viewing Video Files On ComputersDue to displays on TV and PC displays using different levels, you need to convert 16-235
to 0-255
this can be done via a simple shader filter in MPC for example.
Example Filter Settings:
The chroma-decoder can produce images that are not to correct black levels for PC use.
This can be corrected very primitively with initial export (no recommended if re-mastering in Resolve)
Example: --force-video-black-level 32
If a single value is defined such as 32
this will uniformly apply to R
/G
/B
colour if 3 values are defined it will adjust each channel.
This should only be used for PC as it will break TV Legal levels of 16-255
Note
Most BetaMax media releases will have Colour Bars at the start for reference use, which with a vector scope you can set the gain/phase correctly.
Test your images on different displaysDepending on the colour accuracy of your display these images can look bright white in the yellows this is called white level clipping
you need to adjust your display.
The example here is 2 sRGB photos taken of 2 displays
This is for altering the flagging data inside the mov
, mp4
& mxf
containers.
This is the reason why proper mastered media will only show above the active head switch area when displayed on a TV/System with standard video flag support.
AMCDX Video Patcher Live Link / Archive Link This tool is dead simple to use and allows you to correctly flag recordings colour/timecode information, this allows files to be hardware & industry standard software compatible.
Example of a blank and correctly flagged progressive exports for NTSC & PAL using ProRes HQ / V210.
Firstly, the correct way to word a video involving the projects is to state it's a FM RF Capture
decoded with xxx version
of decode not "domesday method" etc as this causes conflation about the methodolgy and workflow involved.
That RF Capture is decoded with VHS-Decode
, LD-Decode
and or HiFi-Decode
etc for example.
It's always best to separate current software generated video files, from the hardware capture projects reffering to it together as "FM RF Archival method" is the most accurate way to describe it.
RF captures are just the archive, not the decoded results as they can always be improved so never be considered a "final" so if possible It's recommended to always share source FLAC compressed RF on The Internet Archive if publishing any sort of decode related content.
To upload anything on YouTube in as close to orignal visual quality as possible you will need to upscale to the 2160p or "4k" bracket which is why we use 2880x2176
with 4:3
aspect ratio, encoded with HEVC 8-bit 4:2:0
at CBR 120mbps
for example.
Using the "4K" bracket of YouTube avoids horrible black/dark colour bias compression macro block artifacts, this is due to forced re-compression, so it's best to start with lossless FFV1 10-bit export and then do your upscaling and deinterlacing via StaxRip or Hybrid or Vapoursynth directly.
Here is 2 very basic all FFmpeg based scripts using BWDIF, but QTGMC is normally better and IVTC for 24p wrapped media, and if using Vapoursynth Spline64 is always a good goto for upscaling.
PAL
ffmpeg -hwaccel auto -i "%~1" yuv422p10le -vf scale=2880:2176 -c:v ffv1 -aspect 4:3 -vf bwdif=1:-1:0 -c:a copy -color_range tv -color_primaries bt470bg -colorspace bt470bg -color_trc bt709 "%~n1_YT_4k_Scaled.mkv"
NTSC
ffmpeg -hwaccel auto -i "%~1" yuv422p10le -vf scale=2880:2176 -c:v ffv1 -aspect 4:3 -vf bwdif=1:-1:0 -c:a copy -color_range tv -color_primaries smpte170m -colorspace smpte170m -color_trc bt709 "%~n1_YT_4k_Scaled.mkv"
For use on Odysee or anything that supports direct streaming:
You can pick either x264_web
or x265_web
depending on how much system/client support of streaming you want.
tbc-video-export --profile x264_web example_decoded_media
Note
Odysee uses direct streaming, no on the fly transcode/re-encoding so care has to be taken to avoid client side playback issues, i.g non phone users.
Jitterbug / Video Dump Channel for making tbc-video-export!
Harry Munday for tbc-video-export's development direction, feature testing and its icon.
Wrapscallion (Ty)
on the DD86 Discord for showing AMDCX Videop Patcher
this truly invaluable and easy-to-use cross-platform tool.
Next Page Post Processing
Sub-Page Jitterbugs TBC-Video-Export Releases Page
Back Page RF Capture Decoding Guide
RetroSearch is an open source project built by @garambo | Open a GitHub Issue
Search and Browse the WWW like it's 1997 | Search results from DuckDuckGo
HTML:
3.2
| Encoding:
UTF-8
| Version:
0.7.4