Jump to content

Alvin Pingol

Basic Member
  • Posts

    677
  • Joined

  • Last visited

Posts posted by Alvin Pingol

  1. Personally, I wouldn't use the cine lenses, as you'll become very limited in the wide angle range; a wide 35mm cine lens will still be telephoto with a 1/3" CCD, and I imagine it'd be a bit difficult to compose your shots.

     

    The manual lens is probably the best solution, as the P+S Technik Mini35 adds its own textured/diffused look that you may or may not like, and the Pro35 would, I believe, require some additional adapters (and a lot more money).

  2. >>I dont think they are 35mm still camera lenses, they are the same lenses we

    >>used on the arri 35mm.

     

    Still, the lenses are designed for 35mm, where the negative is a lot larger than the XL2's CCDs. Without using the P+S Technik adapter, you will still gain a significant amount of telephoto.

     

     

     

    >>By the way, UR-A-HO?

     

    IMHO, or, In My Humble Opinion.

  3. Moreover, dropouts will be more visible on the HDV format; the MPEG2-based recording has a GOP of 15 frames when shooting 1080i, meaning any dropouts will last for at least 15 frames (1/2 second). DV is I-frame only, meaning a dropout can exist in just one frame, in which most cases error correction will catch it and you'll never know about it.

     

    I've heard the GOP length changes when shooting 720p (something like a 6-frame GOP) but am not completely sure.

  4. We took 1/6th of a Z1 HDV still and magnified it full screen. It was equivelant to a full screen DV image.

     

    The only real equivalence I can think of here would be in terms of pixel count. What camera did you use for the SD image? It makes a world of difference; I'm pretty sure one-sixth of a frame from a Z1 won't be as sharp as a full frame from an SD camera such as the SDX900, or even your common professional ENG camera.

  5. Diagrams would be useful. Unfortunately, I can't seem to find any.

     

    This pertains to auxiliary lenses ("attachments"), not the lens itself. I am familiar with the 2-element wide angle lens, the easiest, most inexpensive design. Correct me if I'm wrong, but the front element is what takes in the 'wider' image, and the rear element is what flips the image so that it can then be photographed by the main lens.

     

    Although economical, this design is flawed in that barrel distortion cannot be avoided.

     

    It seems that better, more expensive wide angle attachments contain a third element that greatly reduces (if not eliminates) barrel distortion. I would like to know how this element bends light so as to cancel out the distortion caused by the front element. I've also noticed that at very wide angles with some lenses, there is no barrel distortion, but rather what appears to be mild pincushion distortion! This confuses me, as I normally associate pincushion distortion with the long end of a zoom.

     

    Thoughts?

  6. >>To get a motion blurred and brighter image similar to slow shutter speeds in

    >>the video world, I would have to shoot a lower frame rate?

     

    Yes, as this allows each frame to capture a longer period of time, meaning more blur and increased exposure.

     

     

    >>And if that's the case, how do I slow the footage back down to normal speed?

     

    You mean speed it up? Normally, footage shot with slower framerates is not sped up, but rather printed with duplicate frames to get back to 24fps. If you were to speed up the footage, everything would be playing in fast-forward.

     

    In fact, a lot of fight sequences in martial arts films are shot at ~22fps and sped up to 24fps. This is done to increase the speed of the choreography, as well as to keep things safer on set. Full-speed, full-power martial arts is deadly!

  7. >>If it's something like 15 meg or something you could probably just sign up to a

    >>free online FTP like geocities

     

    Keep in mind, though, that GeoCities and similar free webhosting sites give you very limited bandwidth transfer. Bandwidth transfer is very important, especially when it comes to large files like video - for example, three people may watch the clip and then the account will be suspended for 24hrs (or one person can let the clip load, then refresh, three times, same thing).

     

    I second looking into the webspace your ISP provides. Mine gives me 25MB per account (which, I believe, I can have 5 of), and so far what seems like unlimited bandwidth transfer.

  8. Side-scrolling is the devil. ;)

     

    (Just a quick note to Daniel and anyone else posting images, please resize them! While it is not a written rule, I would recommend keeping widths under 800 pixels - this way, side-scrolling will be unnecessary for most users, and those who still run at 800x600 won't need to scroll too far. Thanks)

  9. BTW, setting the detail to "0" in Sony HD cameras is not the same as turning the detail off. Sony uses a scale of -99 to +99, where zero represents the middle of the scale and the "nominal" amount of sharpening.

     

    Thanks for that important correction. I tend to forget about negative values for things like sharpness because, frankly, I don't think it makes sense - "Negative" sharpness should technically be adding blur... but oh well. :P

  10. >>Until somebody can figure out a way to fit a USB socket into the back

    >>of the cinematographer's head to inform the processing software what is

    >>intended, video cameras are always going to produce un-natural looking

    >>images!

     

    Always?

     

    Setting detail/sharpness to 0, or Off, effectively solves this (assuming, of course, the camera is capable of delivering sharp images without the need of electronic edge enhancement! :D )

  11. Hi Alex,

     

    >>what type of camera

     

    You'll want something that offers as much manual control as possible. Manual focus, aperture, and shutter speed are a must. More professional cameras will offer additional image settings, such as custom white balance, edge enhancement ("detail"), saturation, knee compression, gamma/master black, color matrix, etc. Expect to spend at least $2500 for just the camera - other money going towards necessary accessories like tripod, batteries, filters, etc. Be happy you're not required to buy the camera lens separately.

     

    Consider the following "pro-sumer" 3CCD camcorders, in no particular order:

    Panasonic DVX100, DVX100A

    JVC DV300

    Sony VX2000, VX2100

    Sony PD150, PD170

    Canon GL2

    Canon XL1, XL1S, XL2

     

    Plus any I may have forgotten. Another option to consider is HDV, though it is still an early format.

     

     

    >>sound equipment?

     

    Find an inexpensive shotgun mic, (Azden SGM1X, for example). Almost anything is better than a camera's onboard mic. Using an external allows you to practice proper mic placement, and teaches you how to light and compose shots while avoiding shadows or seeing the mic in the frame. External audio recorders really aren't necessary, especially if the camera you choose has XLR inputs to begin with.

     

    Eventually you'll want to move on to better audio equipment, but if you are just beginning it'd be overkill to buy a sound package worth more than your camera package.

     

     

    >>What camera would be good to shoot with and will also be good for the big

    >>screen?

     

    That's a tough question because it really depends on how you define "good."

  12. Thanks for the replies.

     

    >>What law did you use to recomposite them?

     

    With layers in Photoshop, the overexposed frame was placed atop the underexposed frame using Screen transparency, overexposed layer set to 50% opacity. This method raised shadow areas and brought detail from the underexposed layer into the comp, while returnining the "overall" exposure close to that of the control shots. I tried using Multiply instead of Screen - I think the results may have been a bit better, as shadows appeared not to become muddy as quickly.

     

    >>I think the shot of the candle makes the improvement clearest and it looks

    >>very good.

     

    It's actually an incandescent light bulb. Using the Multiply setting on this particular image provided a noticeably different composite, with shadows on the white fan coming out a little darker.

     

    Kai.w thanks for posting those links. A lot of great looking images on there.

     

     

     

    >>You didn't say if your images were captured with a digital still camera or a

    >>video camera.

     

    These are from my JVC DV300. While it usually isn't mentioned alongside the usual pro-sumer 3CCD's, like the PD170/VX2100/DVX100/GL2/XL1/2, it's a good camera and I like it.

     

    >>The gamma of a video camera is not equal or linear throughout its whole

    >>dynamic range; the highlight portion is typically much steeper than its shadow

    >>response.

     

    Something I hadn't considered before doing the exposures!

     

    >>What all this means is that adjusting exposure can also change the contrast

    >>significantly in differently lit areas.

     

    Exactly - this is why I think the 4-stop-difference images look somewhat foggy or washed out. If it makes any sense, the underexposed images were not as dark as the overexposed images were bright. In other words, opening up two stops placed too much of the image in the highlight range, which is quite limited on this camera, effectively washing out the midtones.

     

    I'll be checking out that HDRSoft site. I just love the way the images come out from what they show in the gallery. Very nice.

  13. I had a bit of free time today. (Re-)Inspired by this thread, I ran a test that I had been planning on for quite some time. Read on if interested...

     

    First images are the control images, "properly" (to my eye) exposed. Underneath them are the composites of the same scene underexposed and overexposed by equal stops.

     

    The first set of composites consists of images 1 stop underexposed and 1 stop overexposed.

     

    Control

    1stop_control_a.jpg

     

    Comp

    1stop_comp_a.jpg

     

    As expected, more detail in both the highlights and shadows, though saturation was lowered somewhat. A better example, note the detail recovered in the hotspots:

     

    Control

    1stop_control_b.jpg

     

    Comp

    1stop_comp_b.jpg

     

     

    Wanting to see how much dynamic range I could "replicate," I did the same test (new scenes) but used 2 stops for the over/underexposed frames.

     

    Control

    2stop_control_a.jpg

     

    Comp

    2stop_comp_a.jpg

     

    Very low contrast. I attempted the same thing with a day EXT shot, but the results weren't so pretty...

     

    Control (odd color due to improper white balance, my bad)

    2stop_control_b.jpg

     

    Comp (neighbor came home and left his garage door open, sorry)

    2stop_comp_b.jpg

     

    I did a quick Levels adjustment to bring the blacks and whites to their normal levels but, as was stated in an earlier thread, re: Phil's 16mm shoot I believe it was, doing so gets you almost right back to where you were in video:

     

    2stop_comp_bl.jpg

     

    Obviously it's not exactly like the control, with highlight ares having lower contrast and shadows rendering differently, as well.

     

    Looking at the levels-modified comp versus the control (for the EXT shots), I can't really say which one I like more. While the control looks a bit more natural, the lower contrast of highlights in the composite really stand out to me; the smaller contrast ratio between the front and sides of the lightly-colored houses makes a huge difference.

     

    Thoughts?

×
×
  • Create New...