Two Types of Interlace?

[FLM]

Hi, everyone...

for several months now, i've been capturing SD TV Shows using an Analog TV capture card: Leadtek Winfast PxPVR2200.
I record using the supplied software with the card: Winfast PVR2.
more than half of my recordings are PAL Interlaced Sources.
over the past few months, I Burned and played the recording on a standalone DVD Player connected to a CRT TV and I noticed that some of the interlaced shows look like they play as progressive video at 25fps.
other tv shows that has the same horizontal interlacing lines on the computer plays on the crt tv very smoothly at 50fps.

I wanted to ask: what is causing this and is there anything that can be done to make the 25fps files play at 50fps without re-encoding using a double framerate deinterlacer (bobbing?)? or is it "normal" behavior for interlace (some play smooth, some don't)?
another thing I forgot to mention, those shows play the same via the cable Set-top-box I use to record from...

I uploaded 2 sample video clips (http://www.fileflyer.com/view/raKfOAW).
for some unknown reason, winfast pvr encodes the shows as Bottom Field First. I fix this using ReStream to make them Top Field First and play them on the dvd player.

[Revgen]

Have you tried using Dscaler? Your card has a conexant chipset, so dscaler may support it.

[Guest]

The one that plays like progressive does so because it *is* progressive (ignoring chroma artifacts on one field).

You need to realize that just because the physical display or capture source is interlaced, the *content* itself may be progressive, i.e., there are pairs of fields that come from the same temporal moment.

[FLM]

The one that plays like progressive does so because it *is* progressive (ignoring chroma artifacts on one field).

You need to realize that just because the physical display or capture source is interlaced, the *content* itself may be progressive, i.e., there are pairs of fields that come from the same temporal moment.

Wait a second... Now I am totaly shocked!

How come the content is progressive when on the pc every frame has Interlacing lines? and every frame has two distinct pictures from different position in time?
You said there are two fields from the same temporal moment but to my eyes it looks like completely different fields from different point in time which made my assumption that the content is interlaced.
for an unprofessional eye (like mine), it looks like it is interlaced because every guide\tutorial on interlaced video shows these exact characteristics (every frame has two distinct pictures in time).

what are chroma artifacts? I know a little about chroma subsampling and YUV colorspace but nothing major that can help me with this.
are these chroma artifacts are the cause for those "splitted" pictures in every single complete frame (which, again, looks like pure interlace)?

I said that I play the files on a dvd player- does the dvd player do any altering to the picture while it sends it to the TV set?
how can I get rid of these horizontal lines while playing the video on the PC using any directshow player without losing too much quality? de-interlacing is unnecessary, right?

Thank you very much for the reply and I hope I'm not bothering you too much with my beginner's questions. I just don't know where else to go to find info about this...

Have you tried using Dscaler? Your card has a conexant chipset, so dscaler may support it.

Hmm, I don't mind watching this content as it is on the PC cause I mainly watch it on my CRT TV as I mentioned earlier.
If Dscaler will get rid of these lines without losing too much quality, I'll use it. but it's not my main concern.
I only wanted to know why two similar-looking (even fooling) contents play differently on a CRT TV...

[Guest]

How come the content is progressive when on the pc every frame has Interlacing lines? Because there is a one field phase shift. Pass it through Telecide() to see the progressive frames.

and every frame has two distinct pictures from different position in time? That's not true.

what are chroma artifacts? I know a little about chroma subsampling and YUV colorspace but nothing major that can help me with this. If you separate the fields and step through, you will see them.

[FLM]

Because there is a one field phase shift. Pass it through Telecide() to see the progressive frames.

That's not true.

If you separate the fields and step through, you will see them.

Ok, I see. I'll try Telecide()...

for future reference, Can you tell me how to differentiate between real interlace and one field phase shift? I heard this term but I didn't know it really resembles true interlace...

and for the burning question and the reason I started this thread: What makes this happen? something in the analog tranfer process?

Thank you very much for your kind help... :)

[Guest]

for future reference, Can you tell me how to differentiate between real interlace and one field phase shift? I heard this term but I didn't know it really resembles true interlace... See the first two questions here:

http://neuron2.net/faq.html

Specifically, for normal progressive you have this, where the [] denotes frame alignment:

[a a] [b b] [c c] ...

For phase-shifted progressive, you have:

[z a] [a b] [b c] [c d] ...

All the right field pairs exist but they are aligned wrong in frames.

and for the burning question and the reason I started this thread: What makes this happen? something in the analog tranfer process? Usually bad edits in the source where a field is lost, although the capture card can cause it also. It doesn't really matter because you will see it often in the wild and have to deal with it. The phase can change throughout the stream, so you typically use an adaptive field matcher like Telecide() or TFM().

Thank you very much for your kind help It is my pleasure to assist one who thirsts for knowledge. But as a seeker after truth, I have to mention that you appear to have encountered this issue before and subsequently forgotten about it:

http://forum.doom9.org/showthread.php?p=1191458#post1191458

The alternation you mentioned is, of course, the phase changing throughout the stream. As described there and above, your solution is to use an adaptive field matcher.

[FLM]

Thank you so much for this explanation!
Finally I'm starting to grasp this issue...

so, if I understand this correctly and correct me if I'm wrong: one field is lost and the other that is still present, makes all the other fields to "shift forward in time" and makes all this mess to appear?
another thing I still don't get is how this content is progressive if it is break down to fields? somewhere in the edit process the editor wanted this content to be progressive and deinterlaced it incorrectly?

about the DVD player and a CRT\LCD TV:
1. I now watch this content on a CRT TV as I mentioned and I want to know who makes the progressive frames to appear: the DVD Player or the TV? if it is the DVD, how does it "know" to drop the problematic field?
2. If I connet the dvd player to a LCD TV, will it deinterlace this phase-shifted progressive content and degrade the quality?

about that old thread I opened: I didn't forgot about it and I'm glad you mentioned it.
the content that I was talking about had a different characteristics: it was 80-90% progressive and only some scenes had interlacing lines. but now that you explained this phase-shifting, I understand it more clearly... ;)

Thank you so much for helping me with this subject.
after so many months I now understand this "mystery"... :P

[Guest]

another thing I still don't get is how this content is progressive if it is break down to fields? If there are field pairs which sample the same temporal moment, it is progressive content, regardless of how the fields are presented to a display device. In other words, you can have an interlaced display device and play progressive content on it.

somewhere in the edit process the editor wanted this content to be progressive and deinterlaced it incorrectly? Possibly but more likely the source content just was progressive, such as film.

1. I now watch this content on a CRT TV as I mentioned and I want to know who makes the progressive frames to appear: the DVD Player or the TV? if it is the DVD, how does it "know" to drop the problematic field? The content is progressive. If you are asking why you don't see frames with blends due to bad field alignment, I suppose something in the display device is doing field matching or phase shifting.
2. If I connet the dvd player to a LCD TV, will it deinterlace this phase-shifted progressive content and degrade the quality? It depends on whether the TV does field matching or phase shifting.

Your uploaded stream also has the wrong field order compared to what the MPEG2 syntax signals. This may be the cause of difficulties as well. Finally, it occasionally randomly generates a frame out of sequence. That capture card and/or its software is likely the source of all this (unless you have improperly transcoded the stream) and you should ditch it and get a more reliable solution.

[FLM]

the TV does field matching \ phase shifting on it's own? it seems too sophisticated for a 10 year-old 20-inch CRT TV to do... (or not?)
I really want to reveal which of the devices causes this field phase shift because soon I will ditch this CRT TV and buy a big screen LCD HDTV. I wouldn't want that HDTV to de-interlace this content and destroy the quality of the not-so-fine archive that I already captured...

I don't think my capture card is to blame because I watched the same episode via this capture card when it was on re-run and it didn't have these fields \ blends I see now with this capture.

about the wrong field order: I know for a fact it is the software that came with the card (Winfast PVR2) and I said that in my first post.
I ditched that software a few months ago and from then on I use the Capture Tool of Mainconcept MPEG Encoder. using this Tool I specifically capture everything Top Field First.
on the CRT TV using the DVD Player it plays EXACTLY as it plays via the digital cable box I capture from, meaning I don't see blends of frames as I see them on the computer screen.
here is a sample capture from mainconecpt (http://www.fileflyer.com/view/7V8k3Bq) - maybe now things will be more clear (?).

this is the 3rd analog capture card I bought in the last 2 years. If I will buy another one it will be a HD Capture device like the Hauppauge HD-PVR maybe...

all I wanted to know is what is causing this and you have helped me a lot and enlighten me very much on this subject and for that I thank you so very much. :)
Never have I encountered such a knowledgable and experienced people as I encounter here... ;)

[SeeMoreDigital]

the TV does field matching \ phase shifting on it's own? it seems too sophisticated for a 10 year-old 20-inch CRT TV to do... (or not?)What's the make and model no. of your CRT TV and DVD player. And which type of "input" connection (composite, s-video, SCART, etc) are you using?

[FLM]

I'm not so sure you will find any documentation on the CRT TV on the internet because it's a really old, cheap and unfamiliar model. the back of TV says Sharp 20R-SC.
the DVD Player is Apex DVX-811.
I connect the dvd player using a Composite cable because the TV only has RCA inputs (no svideo, no scart, no component, no nothing :P). plain and simple composite and only one audio RCA input (white cable) lol.

[SeeMoreDigital]

If your CRT TV does not support progressive video and your DVD player has not been configured to pass progressive video, then all you will ever see are interlaced images.

And what-ever you do, "don't" attempt to set your Apex DVX-811 DVD player to output progressive video, as you may end up seeing a "black" screen!

Perhaps it's time you considered buying yourself a flat-panel TV. Most, if not all, support progressive video via HDMI and/or component phono inputs ;)

[Guest]

@FLM

The phase shift can be in the source content. It is not necessarily introduced by your chain.

[FLM]

@FLM

The phase shift can be in the source content. It is not necessarily introduced by your chain.
I know it can be in the source, but because you said that a capture card might cause this also, I wanna be 100% sure it's not my card\software that is at fault. how I can be sure is beynod me... :\

@SeeMoreDigital

about the CRT TV and the DVD Player: I never touched any of the settings of the dvd player. it does have an ability of Progressive Scan but only via Component. my CRT has only Composite so it's irrelevant. as neuoron2 said, the content is defentially screwed up somehow...

I know that LCD \ Plasma TVs do deinterlace any interlaced content so I am not eager in purchasing one. Interlaced content plays exactly as I want it with my CRT (no added artifacts because of deinterlacing)...

[SeeMoreDigital]

I know that LCD \ Plasma TVs do deinterlace any interlaced content so I am not eager in purchasing one. Interlaced content plays exactly as I want it with my CRT (no added artifacts because of deinterlacing)...Maybe that's the case with some of the smaller cheaper flat screen TV's, but it's certainly not the case with any of mine... It depends on the quality of the display ;)

[GodofaGap]

the TV does field matching \ phase shifting on it's own? it seems too sophisticated for a 10 year-old 20-inch CRT TV to do... (or not?)
A good old TV doesn't need to correct for phase shifting because it is actually displaying one field after another and not two fields in one frame like a computer display.