View Full Version : XL2 25p question


Asa Phelps
March 7th, 2005, 05:25 AM
Hello everybody,
maybe a stupid question, but I never worked before with progressive video.

On page 42 of my XL2 PAL manual is said "25p mode: Records in 25fps progressive mode and outputs signals converted into 50 fps interlace".
So even in 25fps the output is intelaced?

It's important to me, because importing with the wrong settings on After Effects may reduce quality.
So I have to interpret footage as "interlaced lower field" importing and rendering in AE?

Thanks

Mike Hardcastle
March 7th, 2005, 06:05 AM
I Know what you mean, thinking the same thing, Anyone.???

Pete Bauer
March 7th, 2005, 06:51 AM
Hi Asa,

I'm in NTSC-land, but the principle is the same: both NTSC and PAL TV signals are interlaced. For PAL that would be 50 fields a second, usually called 50i but sometimes people will label it 25i...but always with the "i".

When shooting 25p, this simply means that the two fields that make up a frame are half of the exact same image instead of two different images shot 1/50th of a second apart. Those two 25p fields can be put back together to reproduce the single, original image that came off the CCD block of the camera. So most people advise to just use "none" or "frame" for interlace settings...that's what you started with, and the image was only split into two fields on tape and in AVI files to comply with the PAL DV specification (which mirrors the TV signal requirements), which is 50i.

Am I making sense before my morning coffee? ;-)

Geoffrey Engelbrecht
March 7th, 2005, 10:05 AM
Just to try to make clearer what Pete was saying:

All PAL video is made up of 25 images every second.

Interlaced video records 50 images every second each with half the vertical resolution. Every two of these images are blended together to form a single image where each horizontal line of video alternates between each image to produce 25 images per second.

Progressive video simply takes 25 images every second. No blending so each frame is one complete still picture very much like a frame of film.

How these are recorded to tape or in AVI files I am not certain. As I understood from Pete's mail it may be that the images are recorded as half images 50 times a second and then reinterpreted by your TV or editing software (i.e. with interlaced recording a different image 50 times a second and with progressive recording the same image twice 25 times a second - resulting in 50 frames a second). However the end result is what I stated above.

Regards,

Geoff

Chris Hurd
March 7th, 2005, 11:07 AM
<< How these are recorded to tape or in AVI files I am not certain >>

I suppose I need to add this to our XL2 Watchdog FAQ (http://www.dvinfo.net/canonxl2/xl2faq.php). Progressive video is broken down to interlace when it's recorded to tape. It has to be; otherwise you wouldn't be able to play it back from the camcorder nor view it on a standard television or video monitor.

Although it becomes interlaced, the progressive temporal motion is preserved because there is no longer any time difference between the two interlaced fields within any given frame. Hope this helps,

Asa Phelps
March 7th, 2005, 01:29 PM
Thanks to all.

So in AE I'll have to acquire and render 25p video as "interlaced lower field" every time and I'll have anyway progressive video look preserved.
Perfect.

Rob Lohman
March 12th, 2005, 04:58 AM
Let's get a couple of things sorted out:

1) the difference between interlaced and progressive is nothing in the digital the world, the storage is exactly the same (except for some indicators which can tell which of the two it is)

2) the difference is in the content, and here the only difference is a TEMPORAL one!! In other words, the two fields have a time difference (which creates a higher temporal resolution and the well known "lines" when looking at the footage on a progressive scan system)

The DIGITAL (very important!!) information is simply stored in two
fields. Think of this that 50% of the signal is stored in one file and
the other 50% in another. These can simply be recombined to get
the original signal.

So the progressive signal is "stored" in (two) "fields". The signal
itself is NOT TOUCHED (except for compression, of course)!

So in the end it makes no difference how things are stored, but
how they where acquired where the real (temporal) difference
between interlaced and progressive rears its head (as Chris
explained).



Asa: no, the content and signal is STILL IN PROGRESSIVE (if you
shot it that way)!!!! Keep AE and all the other programs in
PROGRESSIVE only!

Interlaced is for when you shoot interlaced. The settings tell the
software in what way the footage was SHOT, not STORED!

Hope this explains things a bit better.

MC Smith
April 7th, 2005, 05:34 AM
I have learnt a lot from this thread but it does give me a question

Assume i am recording a moving subject

If i get it right with 50i what comes out of the end of of my firewire cable is 25 odd and 25 even fields of video.
All 50 are slightly different because the subject is moving.

In 25P i get 25 odd and 25 even fields but only 25 of those fields (in pairs) are different .

So why do matrox say clearly they cant capture progressive video on the RTX100 if its still 50 frames coming out of the camera ?

Pete Bauer
April 7th, 2005, 08:12 AM
Hi MC,

I'm not a Matrox user so not sure what context their comment is in. I'll try to take this briefly end-to-end and hopefully that will help you to interpret what the Matrox documentation says. (And if I make a technical boo-boo, someone please correct me).

The camera CCD block images either 50/60i PAL/NTSC (a new field is created every 1/50 or 1/60th of a second) or true 25/30p (a whole image is created every 1/25th or 1/30th of a second, and then split into fields for output). Fields are always output continuously alternating between upper and lower (that can become important for 24p work, but I won't go there today).

The camera VCR section records the image as 50/60i in all cases, recording lower field first to the tape because that's the DV standard. It doesn't know or care whether the CCD block recorded the fields at the same time or separately. It simply records the data to the tape as it comes in and is processed.

The computer captures the tape data through your firewire hardware/cable unchanged, except that once on your hard disk, a "wrapper" is placed "around" the data to create AVI (or Quicktime) computer files. But the actual video information is absolutely unaltered. I believe that a NLE's capture module settings will not affect that captured file; it'll still be a 50/60i PAL/NTSC video signal wrapped in an AVI, regardless whether the fields started out as progressive or interlaced. What it may affect is how your NLE handles the video; if you tell it that the video is progressive by choosing "no fields" then your NLE will combine the two fields into a whole progressive frame in the timeline. Using Lower Field First (the DV standard) will tell the NLE that the footage is indeed meant to be interlaced.

There is a distinction between using "no fields," and other similar-sounding features such as "frame blending," or "field interpolation" to de-interlace 50/60i footage. These may apply algorithms such as discarding one field and doubling the other to minimize the combing effect of converting interlaced footage to progressive (obviously, with some loss of resolution), whereas the "no fields" capture is simply intended to recombine the upper and lower fields of a frame back into a whole progressive frame within the NLE.

Your NLE can then export to a plethora of formats, which can vary in frame rate, resolution, and whether they are progressive or interlaced. If it is a true PAL/NTSC compatible output, then it will be 50/60i. What makes it "interlaced" or "progressive" at that point is whether your settings direct that the two fields ARE (progressive) or ARE NOT (interlaced) two halves of a single image.

Yes, this is confusing -- and a lot of the details are not found in user's manuals. Some points I've made are the result of deductive reasoning based on what I do know, and not hard fact that I've found anywhere. So, again, I really encourage anyone who has deeper knowledge to join in the discussion...it is flustering to a lot of us!

Bob Thompson
April 7th, 2005, 03:29 PM
Thanks for all the information. I am still a little confused - Is it possible to edit 25P and 50i DV footage in the same sequence using FCP4.5HD and what should the field order be set to.

THANKS

Bob

Kevin Kocak
April 7th, 2005, 04:28 PM
Maybe this will help
http://www.adamwilt.com/24p/index.html#24pRecording

MC Smith
April 7th, 2005, 09:39 PM
Actually I think the matrox tech guy may have been confused by the question he asked but here is the response

http://forum2.matrox.com/cgi-bin/rtx100/ultimatebb.cgi?ubb=get_topic;f=24;t=001371#000001

Mike - Still confused but gettin' closer