DV Info Net

DV Info Net (https://www.dvinfo.net/forum/)
-   Sony VX2100 / PD170 / PDX10 Companion (https://www.dvinfo.net/forum/sony-vx2100-pd170-pdx10-companion/)
-   -   Flash motion effect, resolution loss? (https://www.dvinfo.net/forum/sony-vx2100-pd170-pdx10-companion/27961-flash-motion-effect-resolution-loss.html)

Douglas Akers June 23rd, 2004 09:21 AM

Flash motion effect, resolution loss?
 
I've owned a VX2000 for over a year now, and have shot countless hours of footage.
However, I've never really looked at the "digital effect" and "picture effect" settings until the other day.
I turned on flash motion, dialed it down to the first "notch"and WOW!
It looked like about 48i.
Now, I only did a quick test through my Sony 8 in. PVM but, it looked damn good.
Does anyone know what in camera processing is involved with this setting?
Is there resolution loss?
It's gotta be too good to be true.

Ignacio Rodriguez June 23rd, 2004 10:36 AM

Sorry, what do you mean by 48i?

Douglas Akers June 23rd, 2004 01:52 PM

What I mean by 48i, is that the "look" of the effect is 48 interlaced frames per second.
Or maybe I should have said 24i.
My bad.
Again, I know that it's not really affecting frame rate, only some sort of image resample/shutter effect.
Still, the look is interesting.
I'll just have to do a more full scale test.
I did try multiple searches on this subject but, nothing came up.

Dave Croft June 24th, 2004 02:31 AM

I too have often wondered about this effect, and found with the setting at the first notch, the more stroby nature (but not too much so) seemed to lower the frame rate producing a mildly 'film like' image. In the manual it says that it is producing a succession of still images, and this becomes more apparent as you turn up the setting.

Does anyone else know anything about this effect? As Douglas said - does it reduce resolution? and what is its pseudo 'frame rate' that it produces.

Dave.

Jeff Sherman June 26th, 2004 11:18 PM

I tried talking to Sony about this ( my DSR PD-150 ) but didn't get past first tier, however - I shot a feature last year and in one of the pick-up shots I had it set on this setting ( Flash ) at the very lowest; during post while de-interlacing footage, this particular piece ended up looking double studder, so we un-did and left it as was. The result was that shot matched the de-interlaced footage ! I don't know if that helps, but most of those effects can be achieved in post.

Douglas Akers June 27th, 2004 09:13 AM

I did a more extensive test and I still could not see any jaggies or resolution loss.
Everything looked great, just a nice motion blur.
If it turns out to be about the same as a smart DI, that will save major render times in post.
I wish I would have experimented with this before I started shooting the feature I'm working on.

Barry Green June 27th, 2004 04:31 PM

The best way to know if it's causing a loss of resolution, is to shoot a resolution chart.

Download one from bealecorner.com at:
http://www.bealecorner.com/trv900/respat/eia1956.jpg

Print it on your printer, frame your shot so it fills the screen, and shoot it both ways. Examine the footage on a computer, and you should be able to tell instantly whether it's causing a resolution loss, and if so, how much.

Douglas Akers June 27th, 2004 06:47 PM

Thank you Barry, will do!

Douglas Akers July 8th, 2004 06:39 PM

Alrighty then!
I finally had time to print and shoot the resolution chart for the test.
The results are better than I thought.
I shot a standard 60 shutter p.s. clip and a 60 with the flash motion on the first "notch".
Comparing the two still frame exports in photoshop, verticle resolution loss was slighly noticeable but, minimal.
I also did an "instant sex" DI to the non flash clip and there was significantly more loss than the "flash" clip.
I'd say that the Flash Motion effect is somewhere in between the two, (no flash & DI).
This would seem to be about what you'd expect from a "smart" de-interlacer.
Personaly, I'll take the Flash Motion, screw the cost of Plug-ins and render times!
Well, at least until I get a DVX!
I played around with some high shutter speeds with flash on and it looked great! Very similar to a friends DVX at high shutter.
I'll try to post the test pics and provide links if I can.
Cheers, Doug.

Douglas Akers July 8th, 2004 07:05 PM

No flash.
http://www.imagestation.com/picture/sraid127/p6efd10b0ab8639fc090b088e335fc31a/f7ed9523.jpg
Flash.
http://www.imagestation.com/picture/sraid127/pcd1246bb56159f18af5f6908b1614be6/f7ed9524.jpg
No flash with De Interlace.
http://www.imagestation.com/picture/sraid127/p68aa76d697093b2944b24468844f4208/f7ed9529.jpg
Hope these links work.

Ignacio Rodriguez July 12th, 2004 06:12 PM

Have you tried panning the camera and checking the resulting frames in an NLE? Reports from a PDX10 user suggests some frames that pass through the effect are still interlaced while others are deinterlaced. Very confusing. Here is the thread from the TRV950/PDX10 section:

http://www.dvinfo.net/conf/showthrea...threadid=28751

Boyd Ostroff July 12th, 2004 08:57 PM

(am posting a response here as a result of another thread in the PDX-10 forum). That's interesting Douglas, but I'm confused as to exactly how the test was performed.

Were you using 16:9 mode on your VX-2000, it looks like it? Did you frame the overall chart so that it fills the frame vertically (there should be whitespace to the left and right if you were shooting the 4:3 chart in 16:9 mode)? If you shot in 16:9 mode then I don't think the test has a lot of meaning because that mode pretty much trashes the image on the VX-2000. Here's what I got at 1/60 sec in 16:9: http://greenmist.com/dv/res/x-e2.JPG However in 4:3 mode a 1/60 sec I got considerably better results: http://greenmist.com/dv/res/x-d2.JPG

Now these may very well create an effect that works fine for your purpose, so I'm not really questioning that. However I don't think you're getting the most out of your camera this way. But it is a fascinating topic, and thanks for taking the time to explore and share your results.

BTW, you can make your links "clickable" using the following format:

[ url ] www.anywebsite.com [ /url ]

or if you want to turn some text into a link:

Check out this [ url=www.anywebsite.com ] terrific website [ /url ]

In these examples you must delete all the whitespace between the url tags and the brackets in order to make it work.

Douglas Akers July 13th, 2004 06:00 PM

Boyd,
No, the camera was NOT in 16x9 mode.
I left my century optics 16x9 lens on for the test.
I did notice the slight squeeze which I think is because the lens and FCP have slighty differnt ideas about the anamorphic ratio, which some other users of this lens have noticed.
Or the difference between square and unsquare pixals when in Photoshop?
I should have resized it after I noticed, but I don't think it matters for this little test.
The test was performed quite simply.
I taped the chart to a chair in my living room under practical lighting and pointed the camera at it with the exposure set to open, no gain, and shutter at 60 per second, white balance at tungsten.
Oh yeah, I did leave my custom preset on which has color and sharpness down 2 notches.
I should have turned custom off, but I don't think it matters because we're still comparing non flash to flash and DI.
Two passes, one with "Flash" on and one without.
Once imported into FCP, I did the instant sex de-interlace to a copy of the non flash clip.
Everybody knows what that is, right?
If not, look here .
How's that HTML? :-)
I made the frame grab through Quicktime export format TIFF, best depth.
Opened up in Photoshop, enlarged about 280%, crop it, saved as JPEG highest quality and then uploaded.
I think the Image station photo host I'm using has a resolution "cap" because when I tried to upload a TIFF file, it converted it to a JPEG so I went back and just did the JPEG comperssion myself.
Am I missing anything?
No wavefrom monitor or capture software, just camera straight in from tape via firewire.
If you think there is something I could do to make this test more conclusive, let me know but as I said, I'm sold on this.


All times are GMT -6. The time now is 06:26 AM.

DV Info Net -- Real Names, Real People, Real Info!
1998-2024 The Digital Video Information Network