Skip to content

Submitting & Downloading Sample Data

Harry Munday edited this page Oct 9, 2026 · 5 revisions

Data Repositories

Sample Dump Telegram Channel

Google Shared Drive Public Samples Folder

Decode Team - The Internet Archive

Sample Data Doc

If uploading a sample to the community shared drive, FLAC compressed captures are mandatory due to ease of upload and download for those of less fortunate network abilities, small 1-10 frame cut out samples should be used for bug reports.

If possible include VCR/Player/Camera used and or original source of generation if known, feel free to create relevant documents separately in a txt file or md format for this purpose one per each media article, or spreadsheet format for larger file collections.

Submitting Context

Please include your notes in a plain text .txt or markdown formatted .md file, google docs are also allowed but standard text data is preferred, due to ease of traslation & migration doing both or all 3 is fine.

Compressing FM RF Captures

Today all post capture compression and resampling has moved to the cross-platform FLAC-Chop.

If you encounter uncompressed data or legacy 10-bit packed .lds or wish to "chop" parts or start/ends of captures then this is the AIO tool for that task all modern captures should be made via MISRC GUI directly to FLAC and appropriate sample rates for the format being ingested.

Naming Rules

Naming should be the following or similar to keep files orderly and easy to copy/paste and index.

Both - and _ are legal separators, ware as ( & ) are illegal, so stick to the examples provided below.

Format: Username_Media-Name-MediaType-TV_system-Capture_Type-Bit_Depth-SampleType

Example 1: Harrypm_Colourbars_vhs_pal_CX-White-Stock_cxadc.betamax

Example 2: Harrypm_Dialog2001_svhs_ntsc_CX-Blue-Stock_Gain-22_cxadc.u8

Example 3: Harrypm_THX_ld_pal_CX-White-Modified_Gain-18_16bit-20mhz.u16

Example 4: Harrypm_Test-43_vhs_pal_DdD_gain-8.5.ldf

Example 4: Harrypm_Holiday_vhs_pal_DdD_gain-8.5_16mhz_8bit.vhs

Tape format, bit depth and sample rate should always be stated unless .ldf/.u16/.u8 is used which implies the bit depth by its extension type.

Files with just .flac/.oga and so on should always have descriptive names.

Help

CXADC Bit depth/Sample rate Designators:

Warning

These are now depricated.

Decoding Command Sample Rate Notes
--cxadc 28.6 MHz/8-bit (8fsc) Stock Crystal Native 8xFsc 8bit Unsigned Sampling Mode
--cxadc3 35.8 MHz/8-bit (10fsc) Stock Crystal Native 4xFsc 16bit
--10cxadc 14.3 MHz/16-bit (4fsc) Stock Crystal 1.25x 8xFsc 8bit Unsigned Sampling Mode
--10cxadc3 17.9 MHz/16-bit (5fsc) Stock Crystal 1.25x 4xFsc 16bit

Caution

35.8 MHz/8-bit is not recommended due to up-sampling, no benefits just wasted space.

Capture Formats

Capture Format Extension Compression Type
DdD 10-bit Packed .lds 10-bit Packed
DdD 16-bit Singed .s16 16-bit RAW
DdD 16-bit FLAC .ldf 16-bit FLAC
CX Card 8-bit .u8 8-bit RAW
CX Card 16-bit .u16 16-bit RAW

List of flac compressed naming designators:

Warning

These are now depricated for in-file naming and embedded vorbis metadata so .flac should be used.

Format Compressed Extension Compression Type Channels
Composite .cvbs FLAC 1
VHS .vhs FLAC 1
S-VHS .svhs FLAC 1
Video8 .video8 FLAC 1
Umatic .umatic FLAC 1
Umatic SP .bvu FLAC 1
Hi8 .hi8 FLAC 1
BetaMax .betamax FLAC 1
Video2000 .v2000 FLAC 1
SMPTE 1” Type C .type-c FLAC 1
SMPTE 1” Type B .type-b FLAC 1
EIAJ .eiaj FLAC 1
BetaCam .betacam FLAC 2
BetaCam SP .betacamsp FLAC 2

Always include the sample rate of a compressed capture if using the format extension naming method, i.g 20msps 10-bit.

With Betacam, betacam-y and betacam-c has to be discriminated, unless you have both channels interleaved or somthing, betacam is still in alpha workflow wise, now that syncronised CXADC capture is viable.

With βetamax be sure to include if βI βII or βIII speed.

Etc more formats will be added as captures are obtained for testing.

Current FLAC compression standards are:

40msps 8-bit, 16-bit & 10-bit packed - Decodes automatically (DomesDayDuplicator Stock) (40mhz CX Card Crystal Mod)

20msps 8-bit - Requires -f 20 (Down sampled File)

16msps 8-bit - Requires -f 16 (Down sampled File)

CXADC Sample Rates:

17.9msps 16-bit - Requires --10cxadc3

14.3msps 16-bit - Requires --10cxadc

20msps 16-bit - Requires -f 20 (40mhz CX Card Crystal Mod)

16msps 8-bit - Requires -f 16

Mass Archive Submissions

We all love large datasets for testing a good example of this is Itewreed’s RF Archive witch has 5.7 Terabytes (TB) of German off-air TV recordings.

Sadly however to changes in Google Workspace as of July 2023 we no longer can accept mass submissions to the google shared drive, we do accept physical submissions in LTO5/SSD/HDD/Optical with index files on the shared drive however, sadly due to the 5TB cap the public shared drive will be limited to 1-2TB of active currated data at any given time for the foreseeable future.

Checksums

Checksums are recommended typically CRC & Sha3-512 for verifying data integrity are appreciated in separate files & index names alongside captures to verify a clean file upload/download and local transferring.

Example of a CRC checksum hash embedded: Testbars-PAL-CXADC-[248955C9].svhs

Example of a Sha3-512 checksum: b984fa07139ddaf2a1c4b8c6b2a05ca5c4c0e1119616ce329bdb72916c86f107

Easy to use GUI checksum tools RapidCRC (Windows) GTKHash (Windows & Linux)

If making optical archives or using the .ISO format then look at Dvdisaster for embedding/creating ECC data and verifying discs.

Page End

Clone this wiki locally