Length of the SCART cable matters?

[rgr]

Someone probably wondered whether the length of the SCART cable matters.

I used good Tech+link WiresCR cables to check.
0.5m - flat
1m - standard

MP4 files attached.
The last file is a comparison:
- top row: 0.5m and 1m with highly enhanced contrast
- bottom row: 0.5m and 1m unchanged

Everything was captured to the H265 codec with a 12mbit bit rate (yes, I know, lossless would be better), and then recompressed to H264 with CRF 17.
The difference can be seen in the size of the H264 file - for a 0.5m cable the file is 10% larger, which would mean a higher noise level.

Maybe someone will find it useful

[Sharc]

Length and screening/shielding basically matters, yes.
Not sure about your case though: Maybe noise, maybe different dotcrawl, maybe the non-deterministic encoder .... ?
Some observations:
The bitrate diverges increasingly over time between the tests
I,P,B frames were allocated differently
x264 has a setting for 'deterministic encoding' (or similar, I don't remember now). Try the deterministic setting.
Was the environment the same? Has an electric appliance been switched on/off between or during the tests, causing noise?

[rgr]

Capturing 10 minutes with a 0.5 m cable, turning off the equipment and changing the cable, Recording 10 minutes with a 1 m cable, then selecting 1 minute from the middle of the recordings and recompressing the source from H265 to H264.

The only downside is the use of ffms2. Only LWLibavVideoSource handles VFR correctly, ffms2 loses 1 frame per 20000 frames.

[pandy]

Use good cables - i have 20 meter long VGA cables and VGA signal bandwidth is at least twice higher than SCART (usually way more than twice).

[lollo]

Just as reference, when capturing I see and measure absolutely no difference between 1m and 2m lenght S-Video cable Monster Cable, series MonsterVideo3.