Info: Comb Filter Testing — Page 3

Does anyone here have an ATI Theater 550 Pro?

drngr said:

So, any checkerboarding on althor's full cap of SW JSC? I couldn't find any.

No, none!

Now we just need to figure out if the "AGC issue" is real or not.

Here's the 2D 5-line comb filter in the Extron YCS-100.

Click here for a side-by-side comparison of the Extron YCS-100's 2D 5-line comb filter and the Faroudja VP-100's 2D 3-line comb filter.

drngr said:

Trying to do the same trick with an LD would be like averaging YCS-100 capture 1 with VP-100 capture 1; they won't cancel out.

As long as there's at least one analog->digital->analog conversion in the chain, then the phase-inversion trick is possible.

poita has tested this with many LD captures and has verified it to work.

The key thing to realize is that the video-DAC in the device determines the phase; not the source.

Extron YCS-100: composite=>([A->D]->[digital processing]->[D->A])=>s-video

Also, an easy way to get the device to invert the phase is to turn it off and then on again.

So, if you want to test this...

Do a capture, turn everything off, and then turn everything on again and do another capture.

If that doesn't do it, then turn everything off and try again.

Repeat until you get a capture with inverted phase. It sometimes requires several captures until you get two with a 180-degree phase difference.

One last thing...

This trick wouldn't work with just a single analog->digital conversion in the chain; e.g., connecting the composite output of an LD player directly to a capture card.

You need an analog->digital->analog intermediate step before the capture card gets the signal to digitize.

AntcuFaalb said:

Also, an easy way to get the device to invert the phase is to turn it off and then on again.

So, if you want to test this...

Do a capture, turn everything off, and then turn everything on again and do another capture.

If that doesn't do it, then turn everything off and try again.

Repeat until you get a capture with inverted phase. It sometimes requires several captures until you get two with a 180-degree phase difference.

Interesting... so, I pass the LD video out to a DVD recorder in, then DVD recorder out to the video capture card in, I have the first capture; then I switch off everything, do a second capture, average the two and obtain a chroma noise free result, right?

How could I check the second capture has inverted phase? I mean, with a test disc it's quite obvious, but with real moving pictures?

I thought to do this: capture the test disc, then capture the LD; switch off everything, capture the test disc, check if the phase is inverted; if so, I continue and capture the LD, if not, I repeat until the phase is inverted. This is the best technique I have in mind... the only downside is I have no PAL test disc, so I can't apply this to PAL LDs... (**/)

_,,,^..^,,,_ said:

Interesting... so, I pass the LD video out to a DVD recorder in, then DVD recorder out to the video capture card in, I have the first capture; then I switch off everything, do a second capture, average the two and obtain a chroma noise free result, right?

Correct, but only if the second capture has a 180° phase difference from the first. It sometimes takes several captures before you get one.

Also, it's binary. A capture either has a 180° phase difference from the previous one or it doesn't; there's no in-between. You'll never have a 57° phase difference, for instance.

It's always either 0° (no cancellation) or 180° (perfect cancellation).

_,,,^..^,,,_ said:

How could I check the second capture has inverted phase? I mean, with a test disc it's quite obvious, but with real moving pictures?

The easiest way is to look to see if the chroma noise (rainbowing) has been canceled-out. If it's canceled-out in one place, then it's canceled-out everywhere else.

This is particularly easy with Star Wars since C-3PO and the spaceship models rainbow like crazy.

_,,,^..^,,,_ said:

I thought to do this: capture the test disc, then capture the LD; switch off everything, capture the test disc, check if the phase is inverted; if so, I continue and capture the LD, if not, I repeat until the phase is inverted. This is the best technique I have in mind... the only downside is I have no PAL test disc, so I can't apply this to PAL LDs... (**/)

This technique would probably work just fine. I haven't tried it yet, so I can't make any guarantees, of course.

Also, one thing to keep in mind guys is that I've only tried this with 2D comb filters: the Faroudja VP-100 and the Extron YCS-100.

I'm 99% certain it'll work with other equipment, such as 3D ones, but I can't be sure.

After some testing today, it seems likely that I'm flat-out wrong on this.

I'm going to do more testing and discuss my findings with poita. I might be doing something wrong.

drngr said:

AntcuFaalb said:

Does anyone here have an ATI Theater 550 Pro?

There is a still of the second S&W PAL test here: http://forum.videohelp.com/threads/286099-Blackmagic-Intensity-Pro-Can-t-capture-from-any-VCR-source?p=2200637#post2200637

Interesting! I wonder if it exhibits any checkerboarding.

I'd be willing to buy someone a used 550 and have it shipped to them if they'd be willing to do some testing on it for me.

(My motherboard has no more PCI or PCI-x1 slots left and I tossed my other testing motherboards in the garbage during a move.)

Also, forgot to add: No strings. You can keep the card in exchange for the testing.

AntcuFaalb said:

Also, forgot to add: No strings. You can keep the card in exchange for the testing.

Really? Count me in! (^^,) You know, I love testing, and I miss another kind of capture card...

_,,,^..^,,,_ said:

AntcuFaalb said:

Also, forgot to add: No strings. You can keep the card in exchange for the testing.

Really? Count me in! (^^,) You know, I love testing, and I miss another kind of capture card...

Yes, really, but I'm sorry to say that I don't want to pay the expenses for shipping to Italy!

If you're willing to take care of the S&H costs, then I'm willing to pay for the card.

35US$ shipped to me... what do you think? By the way, is it possible to avoid the MPEG2 encoder, and capture lossless/uncompressed, right?

AntcuFaalb said:

One last thing...

This trick wouldn't work with just a single analog->digital conversion in the chain; e.g., connecting the composite output of an LD player directly to a capture card.

Unless your LD player has a digital frame store, a la the D925?