In fact almost no camera in pal shoot 25p, they are 50i. Camera advantage in PAL is the higher resolution.

To answer for 25p, it could be a real adavantage if the AG-DVX100 is really recording on its ccd a full frame AND if it makes an even an odd field from this frame with the even takne from the same image than the odd. This would be good. But, and it is the sense of my question, is there an NLE capable of this. PLEASE VEGAS, TELL ME IF YOU CAN HANDLE 25p... so the disappointment you heard from pal owners is probably related to the absence of NLE, att least to their knowledge. If Vegas handles the 25p, then obviously PAL model is far better than NTSC model.

Stefan

The disappointment was related to the amount of artifacts generated by shooting 25p, that was before the footage went anywhere near a NLE!

As far as I know cameras such as the PD150 will shoot 25p.

The spatial resolution is exactly the same at 25p or 25i. You just need to merge the two fields to get it which is what our eyes / brains do.

The temporal resolution is halved at 25p

Stefan,
I just went back and read through what you were saying again and I think you've fallen into the same trap I had for a while.

When a PAL camera shoots 25i it take one field 50 times a second, NOT one frame 25 times a second and then splits that frame into two fields. You can see this very easily in VV on very fast motion, you get interlace artifacts on the edges if you go through the footage frame by frame. It looks fine at normal speed although I've had it cause grief when its mpeg encoded.

Even most still cameras shoot a frame interlaced, mostly because CCDs are scanned one pixel at a time. The other reason why television uses interlacing is to reduce the bandwidth needed to transmit at a given resolution, there's no loss going down that path.

Also from what SoFo are saying if there is an issue with the DVX100 its going to be in the camera not VV.

The only advantage I could see with 25p and the DVX100 is for the film look, but again the NSTC users report that it looks closer to film but cetainly it isn't film, far more issues to contend with than progressive scan.

Anyway that still doesn't answer your question, I'd suggest you ask Panasonic if they correctly mark the footage as progressive, if they're not doing it right thats a pretty major impediment to sales if no NLE can cope with it due to incorrect headers.

Panasonic claims they record on ccd a full frame and then generate two equivalent even and odd field for each full frame. This would be great because it totally different than recording an even field and then an odd field, because in this case (50i) i have a movement between the two field (in thecase i moved the camera during shooting), a movement i don't have if the odd and even field comes from one single full frame. This is why i absolutely want the 25p. If panasonic claims that than please VEGAS, tell me if you are supporting it. But i should also ask panasonic also.

For the artifact, i can tell i have none, the camera is wonderful. And the PD-150 is nowhere able to record a full frame (25p). Only the XL-1 did a frame mode which an attempt to mimick part of the 25p process. Sort of 1.5 times the field resolution, or 0.75 times the full frame resolution.

Stefan

Why not throw a short source clip online -- keep it to 50mb or so -- and I'll throw it on my V4 and see if it reads properly?

- jim

Stefan,
you are quite right about what happens with 50i when the camera or subject moves. The two fields when merged into one frame will have interlace artifacts. The DVX100 shooting 25p should give the same results as say a telecine, each field is derived from the same frame and can be merged together to give a full res image.

The artifacts I were refering to are motion artifacts, a significant issue when shooting film. You don't see them on a frame by frame basis, from memory they are a big issue with tracking shots where you have a complex motion relationship between background and foreground. 50i can still produce these but the higher temporal resolution makes it less likely.

So you are trading off one set of issues for another, either way when viewed as video the resolution of both systems in DV is still 720x576, viewing footage with motion, frame by frame will look very different between 50i and 25p. but thats not how the finished product is viewed. Certainly if you plan to print back to film I would think 25p a much better starting point or if you want to pull stills from your footage.

But up against that you are stuck shooting 4:3 with the DVX100, I would see that as a big disadvantage when in this counrty we are inexorably going to 16:9 for all television broadcast. I had given the DVX100 serious consideration, I liked the fact that it does a better job of audio than most cameras but the lack of 16:9 has put me right off it.

I work in a business that hires cameras, I have never had anyone ask for something to shoot 25p, 50i at 16:9 all the time, 720p and 1080i quite a lot, its just out of most peoples budget at the moment.

If you really want higher resolution I'd be having a think about the JVC GR-HD10.

I've tried the AG-DVX100E (PAL version) with 25p shooting. Vegas 4.0c will able to recognize automatically the DV stream (progressive/interlace) when I capture using Vegas capture utility. I can correctly see the captured file information on Media Pool regarding Video size, fps, Alpha channel including Field Order. The capture utility is able too recognize my NV-MX500 with Progressive field order when I use the Frame mode of NV-MX500. We can customize the project property by modifying the DV PAL template on Field Order to None (progressive scan) and save to another template name.

Many PAL Mini-DV-camcorders can use some sort of frame mode to record 25p footage. While there is some difference in the way the chips are read and consequently some difference in resolution, there is and was absolutely no problem for Vegas capturing this footage as 25 frame progressive. Even frame mode video from my 5 year old NV-DX100 is captured that way looking as fine as it could. Still hesitant about buying the AG-DVX100 for at least twice the money of the XM-2. There will be some interesting new models by the competitors in just a few months.

Can someone explain to me how there is a difference in resolution if the same camera shoots 25p or 50i?

Maybe I've missed something here but Dv is at best 720x576. The combination of lens and CCD resolution will be less than that but I cannot for the life of me see how wether the CCDs are scanned interlaced or a whole frame at a time the resolution is going to change. If anything given that an interlaced scan means scanning half the number of pixels in a given time the resolution is more likely to be better with interlace.

" ... frame mode video from my 5 year old NV-DX100 is captured that way looking as fine as it could. Still hesitant about buying the AG-DVX100 ..."

Funny you should write this Alex - I was talking about the DVX100 to a colleague recently and he said - why bother - it won't look any better than you can already shoot. I have a 7 year old dx1/EZ1 (pre-firewire 3CCD predecessor of the NV - DX100 ) which also records in Frame mode.

Maybe I'll keep the biccies in the tin ready and waiting for the "right priced" HD camera ...)

I see it like this with the Panasonic range:
Standard def filmic Panasonic NTSC (AG-DVX100) cameras can do things three quite different ways for their 720x480 grid.

- 29.97 progressive, where at each sampling unit the whole grid is sampled.
- 59.94i where two sampling units are taken at equal distances apart (for the length of time sey by the aperture)
- 24p where the same as 29.97p occurs but the camera dictates that you control the lighting with iris/aperture and balance being manual.

Panasonic PAL camera (AG-DVX100E) have two modes where the scan rate:dimension changes:
- 50i - two 25 full frames per second split into two fields from different intervals of time.
- 25p - a single frame captured at each time interval.

I believe that the 2 provided preset "files" for the 25p also go-manual for iris/aperture/"latitude"/gamma. I'd expect to have auto-modes also on the Panasonic in progressive mode but can't see much about whether this is possible on the PAL version. I believe the NTSC version is stuck in manual mode in progressive - making it progressive - but not plug-and-play for point and shoot work.

I'd like to have a 50p mode to pay lip service to a progressive slow-mo - even if the tape has to shift faster or have finer track pitch. Not at this price point. It'll probably be a feature of stills cameras before we see it on video cameras for the prosumer market.


Vegas, Premiere, VT-Edit all support progressive 720x576 25p out of the box - even the earlier incarnations. You might find the earlier versions need to you promote the properties into progressive mode (and possibly adjust the aspect ratio if you have an 16:9 converter lens/adapter fitted).

The difficulties only arise if you expect these NLEs to automate your creation of NTSC 24p target product. This is a manual resize and pitch conversion - no script or templates come with it (maybe they don't exist - don't know). If you print-to-tape to the camera - if your footage is still ALL progressive - you'd probably want to convey that in the PTT. I think Vegas can do this as it can do it for NTSC, even though it manually adds the pulldown to 24p when heading out to the non-"E" AG-DVX100.

I believe that Vegas filters/FX/transitions are all resolution independent and are fine for both interleaved and progressive media. You can test the DV functions of Vegas even without buying it - so test away.

Making a 25p PAL DVD is a little more quirky for a TV RGB/YCrCb target. We'll not get into talking about that for any NLE if thats OK?

Any DV 25p camera is supported. Most PAL cameras are interleaved scan. Even fewer have HD modes and there is no off-the-menu "PAL HD" setting for HD capable devices.

If I was North American (or anywhere NTSC59.94 oriented), I'd actually be keen to get PAL resolution at 25p and convert it to 24p in post. Then again - I'd want 29.97p aswell, so I'd like to have a dual-mode firmware and dual-mode onboard DV codec (which are probably common anyway). I've got eyes far far bigger than my pockets!

"I'd actually be keen to get PAL resolution at 25p and convert it to 24p in post"

This is a new area for me, so please tell me - what, apart from a mathematical conversion task, is the difference between 25p and 24p - is it discernible, and what do you look for?

I find this topic one that can open a whole can of worms:

The length of the film is different if you convert 25p to 24p (you also have to apply the right pitch conversion). Additionally like you say, mathematically, you really should have to frame your 25p viewfinder to suit the final crop to 720x480.

25p is a better match for footage which is mostly going to be shown on PAL TV but might just also be played on film and will have a higher resolution on the film transfer (720x576). If it goes to film - it can also have a pulldown inserted to go (with a crop/noisy-rescale) to NTSC frame sized/rated video (720x480). So PAL 25p is quite universal for movies.

24p (or 23.97..) is ideal if the target is film. If it goes to video, it might even look like converted film by the the time the pulldown has been inserted. It can also be converted to 25p, but will end up a bit shorter and won't have the same vertical resolution as if it were 25p PAL. So NTSC 24p might look quite poor played out on PAL 50i.

~~~~~~~~~~~~~~
Always choose to spend your budget for the camera based on your target format. Transcode as few a times as possible. These are the conclusions that I've established so far.

OT: I've heard people say that 60p is an ideal frame rate as you can always drop a few frames to get to 24,25 or "30" progressive and have enhanced slow-mo options. However I refer back to the fact that you should primarily match your target with your options. Dropping frames and thereby that moment in time can't be taken lightly and might need further signal processing to really be the right thing to do. Like supersampling, smart resampling and de-interlacing/blur - these things all have their price.