David Landelle

Shure, production is 16/9 and 1080i, as well as broadcast on major networks, but the cameraman uses 4:3 markers on the viewfinder while shooting to enshure it will look fine when viewed on a 4:3 display (90% of the world viewers for soccer).

I am not a cinematographer, so this is only my point of view. If SD is 4:2:2 or 4:4:4 the horizontal chroma resolution will be the same or even better in SD than in HDV (4:2:0 or 4:1:1) so upsampling will still give a good impression. Also, HD is not easy to perceive all the time, and it relies on the ability to get a sharp focus during shooting (shooting with a true HD viewfinder of zoom, an HD screen sideways, or measuring as in the old days). In my mind "visibly HD" images are maybe 10% of the time (unless you make a demo reel for an HD camcorder :) In the end, HD is like having twice the resolution (horizontal and vertical) : this is exacteley what you get when you zoom by 2 with a good optical zoom. And the eye does not actually "see" all the big screen. So if you want to tell a story in SD I presume you may tell it using close-up rather than wide angles (as before). In the end, I think very little viewers see HD in good conditions, so wide success may not rely on technical details. You may know that soccer (the highest quality broadcast in europe, all shot with highest quality HD trucks) is still shot taking into account that 90% of the world is watching it in 4:3 SD. Stanley Kubrick made the same choice I think :) My conclusion is that HD was more a tactical move from the broadcast industry to escape the PC and consumer stuff, but for end users, this is not a big deal. Anyway now everybody can shoot crisp AVCHD images with a camera and an SD card ... well I am only talking about technical facts ... this rarely looks like Coppola <_<

Thank you Simon. You are right, the first test I did (not published) was using an HD lens, and the image was naturally sharp. The rent company told me that the lens for the second test was not HD. I - stupidly - said well, I'm only shooting SD, no problem :rolleyes: they probably boosted the sharpness to try to recover that real problem. Even though, I did not get the image I saw the first time with a better lens. :( Also, the MPEG2 encoder probably suffers from excessive sharpness, and adds it's own artifacts.

my appologies for the dead link, clips have moved on the new web http://www.lvideoservices.net/ and the first "uncompressed" rush named "MVR16b9at25p.m2v" is a rough edit of the original material (ingest of DVCPRO50 tape by the MVR). In multicam, up to 10 feeds may be captured in real time. This is a "real-life" test with pedestal, shoulder & funny camera movements. Also shutter is sometimes not set to 1/25, as it should be to provide smooth motion blur. I got only 704 pixels of signal, instead of 720, as it should be in SDI. Panasonic told me @ IBC last week that this was not normal. I'm a programmer, not a video specialist :) and I'm interested in your opinion about "subjective" image quality (compression artifacts, chroma for keying...).

I'm trying to "go progressive" for what we call in french "multi publication", I don't know the equivalent in english, but it implies viewing the same content at different resolutions, and I don't like resizing interlaced content :( When displaying the content on a video monitor, the monitor outputs a field each 20ms in PAL (16,6 in NTSC) but does not know if the content is progressive of interlaced (like most video equipments), so fields taken at the same time (in progressive shooting) are displayed at the wrong time. My question : Aren't the CRT monitors responsible for some of the "judder" often associated to progressive ? (Assuming correct shutter and shooting)

I recorded a program in 25p last week end, and captured it onto my server using the SDI OUTPUT option card of the SDX900. It seems that I get only 704 pixel (and black bars sideways). My server usually records 720 pixels of active signal. Is is the camera head that outputs 704 (most unlikely) or the SDI OUTPUT card ? Also, It doesn't seem to output 10bits in "live output" mode (not using DVCPRO which is inherently 8 bits). Thanks for your opinion !

Another trick is "color cancel" that removes the chroma-key color from the final program. Luma key offers indeed better quality, I have been using it on the ground and on snow on outsite broadcast (virtual advertising). Also, the background beeing uniform, you might select a range for the lumakey which does not need a saturated or dark signal to work (tuning of threshold, clip...). In the end you could even try to build a key combining the luma and the chroma key. But generally, you get acceptable results with the simple method if you got a good signal and a correct lightning :)

Gordon said all. But frame rate should be guided by the use of media. Are you broadcasting in NTSC ? then 30p is better than 24p (which will lead to 3:2 pulldown). Are you planning to do a film print, 24p is mandatory. If you want to burn DVDs, I'm not shure, but both should be OK. Hope this helps.

I have recorded a video clip in 25p with my MVR in 16/9 mode. Here is the result after blow-up in Windows Media : Test SDX800 25p (a 3GHz PC with 1280x1024 resolution is needed to view it correctly). Do you think I should get a better result ? Conditions where difficult, backfocus does not seem to be correct, and lighting...well, this is a test :) Thank you for your opinion !

New for NAB2005 ! The MVR (Multicam Video Recorder) now has SDI inputs. Max configuration is 10 genlocked SDI inputs and 20 XLR independant analog audios inputs - DVCPRO25 quality. 16/9 and 576p support. Public price for 10 channels is 70KE. Licensing, Dealers & Oem welcome - same conditions.

Yes, There was a SONY presentation of HDV in Paris last week. The "progressive" mode is not a true progressive scan, you are right. Keeping half of the vertical resolution is theoretically possible, but should require filtering (like any subsampling). You may try this in VirtualDub, but this is not straightforward. I agree with you that Z1 should be used only for interlaced shooting. This is only my opinion :ph34r:

It depends on the content. If fast moving, 60i (or 50i). Otherwise, choose progressive video mode. Avoid 24p which will lead to frame rate conversion (disaster). David.

Well any card with a BT848 or BT878 should do the trick, provided that it has a mini din Y/C connector (better than composite if you have SVHS tapes) miro pctv, pinnacle pctv, ads tech, hauppauge you can find some for less than 40$ in europe (in supermarkets). Check-out VIRTUALDUB site to get a cool software. You may also get a generic driver from sourceforge if the one provided does not handle full resolution (720x488 if you wish NTSC) to put on top of the original driver. To do real time encoding you need a fast CPU (like P4-2,4) and take a codec that fits your requirements (links on virtualdub site). You may also check out "winpvr" boards a bit more expensive, but with hardware accelerated hw (200$) Hope this helps ! David

hdforum argue with kell factor that 1080i is better (for Europe) than 720p. I disagree :P [with PAL numbers] The test is quite easy with VirtualDub : -take a progressive video (Vp) -build an interlaced video (Vi) -build a resampled progressive video (Vr) Now, compare : -Vr scaled properly to recover correct aspect ratio -Vi deinterlaced properly (using DScaler) Vr is better. I have been looking closely at slow motion machines (like EVS) and building full frames from interlaced material is really difficult, even with smart image processing trying to overcome the prehistoric subsampling of interlacing (even with a broadcast camera which DSP's are working hard to make it acceptable). This is only my european developper opinion. Sincerely.

I don't know what DV (please tell me in a few words) do, but fields are not as friendly as progressive frames, because they are subsampled with no filtering, and DCT assumes signal continuity. We would get better results by sooting with a "half vertical resolution" progressive camera...and up-scale the image after decompression B) maybe this is what dv does ? Of course these techniques were not available in the old days of interlace, but now, we could choose this way instead of interlacing a material which is very difficult to deinterlace properly :rolleyes: (scaling is easier than deinterlacing). Sincerely. David.