Xl2 Camera To Tv (quality)
 

(hopefully Pete Bauer reply's)

Hehe, pete its people like you that make this forum awesome!

Anyhow, my question is (can you tell I just got the camera?) When you record with your camera, and you use the cables to connect it to your TV to see your current work, does the quality degrade a little, or is it better to transfer it to CPU first than tweak it and see better results? The reason why I ask is because I have seen some AMAZING footage but I just can't seem to replicate it, or at least its not looking like that from my TV... Through the small LCD screen it looks AMAZING, but after I see it on TV it just doesn't look "cinematic". I know their are lots of tweaks that I still have to learn but I guess I wish it was a little more view and shoot. Nonetheless this camera is AMAZING!

Thanks

You're too kind, Armando (but that's just the way we like to be here at DVinfo!). I'll start the ball rolling, but others will probably be able to be more helpful since I do 90% of my stuff strictly on the computer, and the remainder goes to DVD+R. Although I did hook my GL2 to the TV a couple of times, I don't think I've ever done so with my XL2.

At 200K pixels, the camera's viewfinder LCD is nowhere near full resolution. It also has adjustable brightness (in the menus) and a focus slider. As is typical of most camera viewfinders and TV sets, it overscans -- that is, doesn't quite show the full recorded image all the way to the edge. So flaws in the recorded image may well not be apparent on the LCD. Depending on their purposes, some folks use the optional higher resolution black & white CRT viewfinder and others use separate monitors when they shoot. But most of us just get familiar with the limitations of a viewfinder through trial and error.

As an aside, for a quick demo of the overscan in the XL2 resolution test I recently posted on my web site:
http://www.geosynchrony.com/scratchpad.htm
I used blue painter's tape to create a 16:9 target. When properly framed so the outside edge of the blue rectangle was right at the edge of the recorded frame, the blue tape was just barely entirely hidden "offscreen" by the viewfinder. I captured directly to HDD, so did the framing of the shots using the computer monitor view.

When you do an analog-out to your TV, there is, of course, a small amount of signal degradation from the D-A conversion and analog cables. But the bigger variable is your TV. If you have a high quality video monitor that is properly calibrated, the picture should still look very nice even if analog. If you have an old, cheap TV with the brightness, contrast, and color saturations all out of whack as they come from the factory, the footage won't look so good. Most TVs never get the tender loving care to be properly calibrated...even my 55" HiDef, I'm embarrassed to say.

As you've probably read in many other threads, the cinematic look is a "beauty in the eye of the beholder" thing, but some of the usual attributes people accept are the frame rate, shutter speed, shallow depth of field, gamma, and color saturation. The latter two can either be done in-camera or in post, or both. As it still is for me, that'll be a combination of learning and artistic license on your part. For example, too high a shutter speed will kill motion blur; the result is that as the clip is played, the movement within frame won't look as fluid.

FWIW, except for test shots, virtually none of my footage ever ends up getting exported to DVD, WMV, or whatever, without being tweaked in my NLE. So all other things being equal, my eye will, by definition, tell me that my footage looks better on DVD than it did straight off the camera because I changed it to look the way I wanted.

How does the same footage look when captured to your NonLinear Editor software? Assuming your computer monitor is reasonably well calibrated, if the footage looks good at full-rez in your NLE or even Windows Media Player, then you can feel comfortable that the degradation you're seeing is somehwere in the analog chain from D-A conversion to display on your TV. No biggie because your final viewing format probably won't be analog-out from the camera.

Let us know what you find!

I too look at most footage on the computer monitor (which has been calibrated) but recently have started looking at it on the TV and I've got to say I think it's pretty impressive for SD video. The TV is hooked up to a Canopus ADC300 (think that's the model number) via the component outputs of the Canopus box so that the chroma signal has never touched the analog world (i.e. it's not converted to analogue chroma to go out over an S video cable or, worse yet, into a composite signal, and then back to color difference signals in the TV set). The Canopus box gets its DV from a Mac Mini (run headless via VNC from a laptop) on a LAN so I can pull raw or edited footage from whatever disk it happens to reside on. Both look great with the edited looking a little better or worse according as to whether my edit decisions were good or bad ones!

All video from the XL2 is captured to HD (using BTVPro). Thus the raw video files can be displayed directly by having the Mini Mac play them back through BTVPro. Edited files are exported and saved as QT movies which BTVPro also plays out as DV through the Fire Wire into the Canopus box.

Armando,

I've a tale to tell that I think fits in this discussion. Recently I purchased Canon's new Realis sx-50 projector. It's a very high resolution beast (1050x1400). Immediately I hooked up my xl2, hoping to be blown away by the "cinematic" quality of the output. Well I was underwhelmed to say the least, as the highlights looked video-ey, the footage looked aliased, and other problems. Then I took the footage into FCP straight...no adjustment, and played it back on the projector through my laptop. Wow...the image was more finely detailed, the highlights looked normal, aliasing was virtually nonexistent. This was a big surprise for me, as I had always expected the opposite. One surprise was this: The oft reported red/blue macroblocking...large aliased halos around saturated red and blue objects(in my case mostly red) was not apparent in the footage viewed from the camera, but very apparent in the playback off the QT file from my computer.

Barry

Barry,

That's most interesting. How about a few more details? Was the camera hooked up though S-video, for example. And what was the connection between the projector and the laptop. If it was DVI much of the credit would go to the video card in the laptop which would treat the projector as a computer display and do the interpolation between 720 x 480 and whatever pixel size the projector demands depending on the mode you have set. This is very important in reducing aliasing as you have noted in another current thread.

You also mention that you went straight into FCP without adjustment and then played back a QT movie. That does imply a conversion from DV to one of the QT formats. Which Codec was used? It probably deserves the blame for the the macroblock haloing as I've never heard of this with respect to DV coding (which doesn't mean that it doesn't happen with DV - just that I've never stumbled across it).

If you want to experiment further you might want to try BTVPro from http://www.bensoftware.com (you can try it free). It allows you to capture DV as DV and play it back as DV to a second monitor (or over firewire which is what I do but to an SD set). It's engine is QT.

AJ

Now that I'm thinking of it, the stuff I tested was actually captured and exported as a .dv file in iMovie. The camera was hooked up with s-video. Laptop was connected through the dvi port on the powerbook into the vga (or whatever the computer monitor RGB plug is called) on the projector. I'm going to try a direct dvi to dvi connector to see if this changes anything (difficult to find in my little burg...and expensive).

The macroblock haloing is something Adam Wilt detailed on his 24p webpage, specifically citing the dvx100 and xl2 in the discussion. I've noticed it mostly in footage with highly saturated reds (some of the early clips on the xl2--i think kaku ito's--showed it in the red tailights of cars.) Wilt also desribes saturated blue's as causing a similar problem ( I think this has to do with the lower sampling rate for red and blue in the DV compression). I'll try to throw up some stuff from the shoot to show what's going on. It's not something I'm worried about...its rarely a problem...I was just surprised to see it played back differently off the tape.

Regarding the laptop playback looking smoother. I agree that the scaling should probably be a lot better, although I was surprised (pleasantly) at the resolution and highlight improvement. I have played a number of older FCP projects through the projector, and have found that the quality surpasses the direct video output and dvd playback by a significant margin.

OK, now I'm tracking. Yes, this phenomenon is caused by lower bandwidth (sampling rate) in the Y-R and Y-B channels. This is part of the DV standard (and also part of the NTSC standard) so that there shouldn't be a whole lot of difference between the extent to which one manufacturer's implementation shows it relative to another's. By design, you shouldn't be able to see it at normal viewing distances.

<<<-- Originally posted by Barry Goyette : AJ

Now that I'm thinking of it, the stuff I tested was actually captured and exported as a .dv file in iMovie. -->>>

Hi Barry,

If you imported to FCP from the iMovie dv files, I am surprised that it looked good at all. I tried that workflow recently and was horrified at the results on my 65" tv. I recaptured using FCP directly and what came out even after being authored to DVD from DVDSP3 was awesome. Apparently, FCP tries to transcode the iMovie footage and uglys it up in the process.

-gb-

Greg..

hope I didn't give that impression...the files I first tested on the projecter were captured in iMovie, exported as DV files then played back in the Qt player...

But I did drag some iMovie clips into an FCP bin the other day...seemed fine to me...but I haven't looked close...hmmm....

Barry