Mixing S-Video Input w/ Component Output For VHS Capturing?

[companda]

I'm playing with a new capture setup, which consists of: Mitsu HD2000U > S-Video Out > ES 15 > Component Out > Capture Card @ 720/480p 59.94/.TS file. I'm hoping someone smarter than me can point out the flaws in this setup? Is the mixing of component and s-video pointless? Also, with this setup, I noticed a significant amount of interlacing which, I'm under the impression is something that you want for archiving/deinterlacing later. I also have a setup that encodes directly to h264 deinterlaced. Those look "better," but is essentially a deliverable "finished" product. I've included a snapshot of an animation video of the component out setup, any and all advice is appreciated.
[Attachment 70466 - Click to enlarge]

[dellsam34]

De-interlacing on the fly while capturing is not recommended and in most cases will result in a permanent damage of the captured video, A snapshot will not tell the whole story, it hardly tells anything about the video, a raw sample clip is the only way.

[lollo]

can point out the flaws in this setup

Analog SD signals should be captured interlaced (29.97 frames per second), in YUV 4:2:2 colorspace, and with a lossless codec. The closer signal format to what is on the tape is Y/C.

[companda]

Edit

[jagabo]

Pretty much everything that can go wrong has. Bad black/white levels, bad automatic gain, repeat frames, blended chroma, dot crawl artifacts, field order switches, maybe a bad PAL-to-NTSC conversion, h.264 encoding has baked a lot problems into the video putting them beyond repair...

[companda]

Pretty much everything that can go wrong has. Bad black/white levels, bad automatic gain, repeat frames, blended chroma, dot crawl artifacts, field order switches, maybe a bad PAL-to-NTSC conversion, h.264 encoding has baked a lot problems into the video putting them beyond repair...

To be clear, you're referring to both clips?

[lollo]

TS interlaced clip?

It is a ~ 59.94fps video, h264 compressed, in a transport stream container, at 720x480 SD resolution, but with a 709 color matrix.

It shows blended frames and fields:

[Attachment 70478 - Click to enlarge]

I do not understand what it is

edit: and jagabo went deeper in finding other defects

[oln]

Component out from the ES15 is fine. It is going to be converted into component internally in the dvd-recorder anyhow so no reason to go back to s-video on the output if you don't need to. I would use interlaced rather than progressive scan output though. You can do much better deinterlacing in post. Many of the US panasonic dvd-recorders can only output deinterlaced/progressive scan video on the component output but according to the manual this one should be capable of interlaced.

[jagabo]

Pretty much everything that can go wrong has. Bad black/white levels, bad automatic gain, repeat frames, blended chroma, dot crawl artifacts, field order switches, maybe a bad PAL-to-NTSC conversion, h.264 encoding has baked a lot problems into the video putting them beyond repair...

To be clear, you're referring to both clips?

I had only seen the first clip (ts1.ts) at that time. The second clip (h2641.mp4) is a little better in some ways, worse in others. I recommend you use a cleaner tape as your source. Something you know for sure isn't a PAL to NTSC conversion. Then you want to capture 29.97 fps interlaced -- with a lossless encoder, not h.264. If you're forced to use h.264 try to convince your device/software to encode it interlaced, not progressive.