help wanted: good examples of poorly-decoding red pictures...

[happycube]

A common LD problem that isn't solved properly in ld-decode yet is that reds on LD are very hard to get right. So if you have any samples, I'd like to collect them to adjust filtering better someday ;)

(edit: otherwise strong disks with >40dB SNR are best for this, since there's only so much that can be done for rotters and soft signal disks...)

[Gamnn]

JSLA022292-wibble.ldf.zip

[IanSmallshire]

The Image entertainment Catalog LDS file on the server (Banana1/Misc/) has a large sections of frames with varous background colours, usually a minimum of 100+ sequential frames per colour.

ld-decode -s 340 for the lead-in frame The entire disc only takes 30-40 mins to decode.

Frame background colours between: 991- 1508 - Red 1510- 2033 - Green 2035 - 2638 - Black 2640 - 3147 - Magenta 3149 - 3414 - Dark Yellow 3416 - 3689 - Light grey 3691 - 3869 - Cyan/Blue 3870 - 3935 - Almost white 3937 - 4135 - Orange ish... 4136 - 4237 - Dark Red 4239 - 4864 - Light Blue 4866 - 4915 - Light Green 4917 - 5000 - Pink 5002 - 5175 - Dark Grey 5177 - 5550 - Yellow ish 5552 - 5609 - Pink 5611 - 5812 - Blue

[AkBKukU]

I have an RLV disc with SMPTE color bars in the lead in that may be helpful for this. I could not ld-cut the file down small enough to fit within github 25MB limit without ld-decode not being able to find frames. It's kinda janky, the VBI data is static during this. I've uploaded it to my own server to share here: https://nc.akbkuku.com/s/demTaARK6HXP5Lz and I'll keep it available for some time.