Deinterlacing in Staxrip does absolutely nothing?

[FAII]

I tried not checking "Deinterlace" for an obviously interlaced source. The results?

It only took half as much time to encode the file (x264/MP4) and the finished output was deinterlaced (somehow), anyway.

So what good does checking Deinterlace do, anyway? I accidentally checked it for encoding progressive source files once. They looked the same as normal to me (but they took twice as long as when not checking the Deinterlace box).

[LoRd_MuldeR]

The difference between deinterlacing or not deinterlacing is clearly visible with interlaced content! When you don't deinterlace, there will be nasty combing artifacts on all moving objects. As soon as you do deinterlace, those artifacts are gone! If you don't see any difference, then your source obviously was not interlaced (or your player does deinterlace "on-the-fly" and that's is why you can't see the difference).

Deinterlacing progressive (non-interlaced) content doesn't make any sens! The difference might not be as clearly visible as with interlaced content. But deinterlacing progressive content only hurts the vertical resolution and thus makes the video look unsharp. Also jaggies and other bad things might be introduced. So don't deinterlace, except your source really is interlaced!

[FAII]

I could've sworn I'd replied to this already:
That's the thing! I didn't check "Deinterlace" on an obviously interlaced source (checked it in VLC Player with Deinterlacing off). The output was deinterlaced, anyway (I also checked it in VLC Player with Deinterlacing off).

Meanwhile, when I accidentally checked for a Progressive source to be interlaced, the output file didn't lose half of its height or anything. In fact, it didn't look like anything was wrong, at all.

The only difference seems to be that it takes twice as long to encode something when "Deinterlace" is checked.

[LoRd_MuldeR]

I could've sworn I'd replied to this already:
That's the thing! I didn't check "Deinterlace" on an obviously interlaced source (checked it in VLC Player with Deinterlacing off). The output was deinterlaced, anyway (I also checked it in VLC Player with Deinterlacing off).

Maybe you did downsize the video?

Meanwhile, when I accidentally checked for a Progressive source to be interlaced, the output file didn't lose half of its height or anything.

It won't loose half of it's hight! Only every second line will be "interpolated", which makes the video look unsharp (which is obvious a bad idea for progressive material).

[foxyshadis]

Could be purevideo/avivo is deinterlacing it for you, somehow, or as part of the decoding chain (mpc, ffdshow and other decoders can do it). You'd have to check the individual encoded frames.

[FAII]

I downsized the video from DVD to 320x240. Does that have anything to do with anything?

[LoRd_MuldeR]

I downsized the video from DVD to 320x240. Does that have anything to do with anything?

Yes it does!

Resizing the video will destroy the interlaced structure of the video!

Downsizing (especially 2:1 downsize) might help to hide combing artifacts, but result is not as good as a "real" deinterlace.
Also resizing interlaced material can produce even worse combing artifacts that no deinterlacer will be able to remove later!

So always remember: First deinterlace, then resize!


// EDIT

See the first samples on the top of this site:
http://www.100fps.com/

[FAII]

But in Staxrip, I just check for it to both Deinterlace and Resize. Deinterlace is above Resize, though, so I assume it's done last?

[LoRd_MuldeR]

But in Staxrip, I just check for it to both Deinterlace and Resize. Deinterlace is above Resize, though, so I assume it's done last?

If Deinterlace is located above Resize, then it will be done first (exactly as it should be done)

[foxyshadis]

Barely matters though, when you resize to QCIF. You'll get some blended frames (like half the videos on amv.org (http://www.animemusicvideos.org/)) if you don't deinterlace at all, but at half size, you can use any deinterlace method including SeparateFields().SelectEven() - which costs nearly zero time - and hardly tell the difference. Might want to make sure nothing's ivtc'able first.