Welcome to PhantomPilots.com

Sign up for a weekly email of the latest drone news & information

Filters again and again

Discussion in 'Phantom 2 Vision Discussion' started by Klaus, Feb 21, 2014.

  1. Klaus

    Joined:
    Dec 30, 2013
    Messages:
    281
    Likes Received:
    0
    First of all, English is not my first language so what I'm going to write now can sounds very wrong or maybe even arrogant, please don't take it that way, I'm only writing this to help.

    Ok here we go.

    The last month or two this Forum have been filled with Gimbals and lens Filter talk.

    I'm not going to talk about the Gimbals (whish I had one)

    I'm going to talk about the filters. Again please don't get this wrong, but I think a lot of the people who want the filter and filter mount actually don't know what they are talking about. But when so many people are writing about it, they also want one :cool:

    So why do I think that people don't know what they are talking about?

    First of all, I have been a professional Photographer for around 20 years (and still are), so I know a little about cameras and filters.

    I really cant see the idea with a filter on a flying thing, because most of the filters people are talking about are polarizers and ND (Neutral density) filters, which all are taking or blocking some light to the camera. A polarizer can easy take two whole steps (4 times less light) of light which mean you get slower shutter times to compensate (for photography)!
    When the camera are hanging underneath a flying and vibrating object you want as fast shutter time you can get, or you photos will be blurry, and with this camera that already are bad quality you will not be happy with the result. Remember the Vision have a fixed aperture (f2.8) which means it can only change the shutter-time to get a correct exposure.

    The ND filter is different because you can buy them the with different density but again why on earth do you want slooooooower shutter time???? Please if anybody can explain that for me.

    Maybe you will say ok, that is for Photography but I'm taking Video :D
    Think about this: A polarizer have different effect compared to the angle of the polarized light. Normally when you are using a polarizer on a normal camera on the ground, you point the camera towards what you are going to photograph and the you turn the polarizer the get the wanted effect. You can't do that when the Camera are hanging 100m above you. Which means when you are shooting your video, the effect of the polarizer will change all the time when you are rotating the Phantom which you will do a lot. So everything like the Color saturation, the clarity, reflections and so on will change.
    Ok I haven't tried it myself on video, but that is how a (circular) polarizer are working.

    For the ND filter, its kind of the same but you use them for different things.
    Many photographers are using them when there are a lot of light, yeah so much light so the can't open up the aperture to get a more shallow depth of field, like a portrait where you want a blurry background. Normally the Depth of field are controlled by 3 things:
    The aperture, the focal length and the distance to the subject and background. But again remember you cant control the aperture on the vision. Maybe the ND can give you good results in one direction but as soon as you turn the Vision it will be totally different.
    So again, please enlight (is that a English word?) me here.

    Another thing, I can see the filters people are buying are very cheap filters (I have seen Peter have been writing the same thing)
    which means you are putting a cheap peace of glass in front of another cheap piece of glass, and you still expect great results. Yeah you actually expect better results, that why you are buying the filters :oops:

    Hope it all makes sense.
     
  2. pault

    Joined:
    Jan 4, 2014
    Messages:
    567
    Likes Received:
    1
    Location:
    Liverpool UK
    The filter holder has been out for several weeks now I cannot recall anyone posting any photographs which they say have been dramatically improved by the use of a filter. Maybe I am wrong but I agree that I have never seen the logic that I have read which says that the camera uses a small sensor and poor lens therefore you may as well use cheap filters for it. If anything the reverse should be true ie give it all the help it can get.

    I think the only real use of the adaptor is to hold a proper lens cap - oh, and generate more cashflow for DJI ;)
     
  3. Pull_Up

    Joined:
    Nov 5, 2013
    Messages:
    3,684
    Likes Received:
    48
    Location:
    South Oxfordshire, UK
    Yep, I'm going to have to agree with you about stills after finally getting some good light to play with my (cheap) filters.

    The cheap circ. polariser certainly cuts out reflection/glare from water (plenty of that around these parts at the moment!), but almost all of my shots had a faint (sometimes quite bright) semi-circular flare effect over the bottom half of the shot, this includes where I was trying to take shots at 90 degrees to the sun. I had marked up the filter so I knew where maximum polarisation was and set that. The unprocessed jpeg images look nice colour-wise, the sky's blue depth certainly shows, and contrast with clouds is nice. There is a reduction in haziness. I'm not convinced anything's dramatically better than could be achieved with post-production

    A cheap grad ND filter didn't show any obvious changes to my eye, and in fact there was obvious vingetting at the corners with this particular filter and the circular flaring effects were much worse. But it was extremely cheap!

    Looking at it again I think the flaring is actually caused by light reflecting back off the lens surround and bouncing about between lens and the new piece of glass that's now in front of it.

    Video, however, looks like it might be a slightly different story (I only took a couple of very short clips at 1080p/30 full width). The circ polariser shows much less lens flaring effect and you still get the glare/sun from reflective surfaces removed entirely. The clouds also "pop" against the blue sky and give a slightly 3d effect with the movement of the camera. Hills in the distance show that haze is still reduced. I quite like the effect of this on video and will experiment a bit more. I did only keep the camera pointed in one direction and just let it do an atti drift, so there are going to be issues when panning and changing the angle to the sun I suspect. Cheap grad filter I didn't really see much difference.

    So based on my very limited experience I am thinking cheap filters for stills - nope, no point. Manually knock the exposure down in bright conditions and do everything in post (shoot in raw for max tweakability) unless you are desperate to remove reflections and reduce haze. For video? Might do some more experimenting with that but I'm not sure any benefit will outweigh the loss of flexibility to flick from video to stills and not get flaring and so on. What I don't know is if a DIY lens hood would help, or even painting the lens surround matt black to cut down on reflections.

    Of course the caveat is I used cheap, cheap filters - if you've got some quality ones hanging around it might be worth having a look at how they perform, but unless you have a specific shot you want to get and can manage flaring caused by the lens surround...

    I enjoyed the experiment, though, and those with more skill, knowledge and better equipment than I may get different results. Thanks to my 7 year old son for his piloting skills whilst I hit the shutter buttons :D
     
  4. pault

    Joined:
    Jan 4, 2014
    Messages:
    567
    Likes Received:
    1
    Location:
    Liverpool UK
    It is interesting to note that your cpl seemed to do the biz as iirc that was the one that was panned in a critical review on Dpreview - just goes to prove one should not believe everything you read on the web !

    In fact, the results are so diametrically opposed that maybe they have changed the glass they use.
     
  5. Pull_Up

    Joined:
    Nov 5, 2013
    Messages:
    3,684
    Likes Received:
    48
    Location:
    South Oxfordshire, UK
    Here's the brief bit of video I shot. Nothing done to this at all, straight from the camera at 1080p/30...

    [youtube]http://www.youtube.com/watch?v=lcCKkUy7F80[/youtube]

    You can see there's no reflections off the water at all...
     
  6. pault

    Joined:
    Jan 4, 2014
    Messages:
    567
    Likes Received:
    1
    Location:
    Liverpool UK
    Yes, presumably by the look of it the sun was just out of shot on the right. Hard to be sure without seeing an on/off comparison but I quite like the effect on the water, gives a sort of placid look. Maybe it is time to think of the filter kit (but I am probably going to go down the P2/Rx100 route when I have seen how Peter (Evans) gets on with the Droneexpert conversion and I would not get the money back when I come to sell the P2V).
     
  7. iResq

    Joined:
    Dec 23, 2013
    Messages:
    545
    Likes Received:
    2
    Location:
    Annapolis, Maryland
    There certainly has been a lot of misinformation on proper filter use.

    Photography is photography, doesn't make any difference if it's land based or from an aerial platform.

    Filters all have a time and place. I use cpl's frequently on ground and in the air. True you can't adjust them in the air. They can be a pain on the ground as well if using a hood. If you can approximate your shot on land, you can adjust before lift off. But even without the benefit of fine tuning a cpl, it can certainly reduce reflections, make skies bluer and trees greener. However, the camera lens is wide enough that you will probably see some color variation from one side to the other. I certainly would not use one for pano work.

    As for increased shutter times needed to compensate for the light loss...check your exif data. I believe that most shots taken in significant light have shutter speeds fast enough to eliminate camera blur. CPL should not be used in low light situations. The image problems that I encounter are not related to camera shake.

    Video clearly poses a challenge because as your angle in reference to the sun changes, so will the coloring of your shot.

    I don't use ND filters in the air (and typically only graduated ND filters on land). There seems to be a school of thought that slower shutter speeds (for video) reduce jello. But I have also read posts where the opposite is claimed. Doing a comparison is still on my to do list.

    The linear polarizer has been suggested. It is widely understood that linears can't be used with auto focus systems (which is not a concern here). But it is my understanding that they should not be used with auto exposure cameras, which would come into play here. Something to consider.

    As for cheap filter argument. I agree to a point. I certainly would not use a $9 Polaroid filter. But Hoya makes some less expensive filters that would probably do a commendable job.
     
  8. iResq

    Joined:
    Dec 23, 2013
    Messages:
    545
    Likes Received:
    2
    Location:
    Annapolis, Maryland
    OK, I'm convinced. Thanks!
     
  9. Klaus

    Joined:
    Dec 30, 2013
    Messages:
    281
    Likes Received:
    0
    Very convincing, but can anybody explain why and how it is working?
     
  10. jimre

    Joined:
    Nov 24, 2013
    Messages:
    371
    Likes Received:
    1
    Location:
    North Bend, WA
    First, ND filters mostly affect CMOS jello caused by prop vibrations. I haven't seen much - if any - of that with the P2V. The props seem to be well-balanced and smooth, at least compared with older Phantoms. So unless you're definitely having this problem, it's not going to do much for you.

    I believe the ND filter mostly works in this case by forcing a slower shutter speed, which introduces more motion blur. The high-frequency vibration from the props is still there, but it's now more of a smooth blur, rather than a sharp discontinuity. Those sharp discontinuities - where part of each frame gets recorded while shifted a couple pixels up/down - are what causes the "rolling shutter" jello effect.

    On GoPro cameras, people have also reported that increasing the video frame rate to 48fps or 60fps can sometimes also help reduce rolling shutter jello. I think this is more a matter of timing & luck. In those cases where the prop-vibration frequency is some multiple of the video frame rate, increasing it might make it so the "vibration shift" doesn't occur in the same place each frame.

    Like I said - I have seen little if any rolling shutter "jello" in my P2V footage so far. I do see a lot of fisheye/motion warping - eg, the "bouncing banana" effect. But that's something completely different.
     
  11. Visioneer

    Joined:
    Jan 2, 2014
    Messages:
    287
    Likes Received:
    0
    Location:
    Louisville, Ky, USA
    Just a guess ... the frames per second are going to be the same in either case but with an ND filter "blocking" light the lens must either have a bigger aperture or stay open longer to admit the correct amount of light. I don't know how these cameras compensate ... a wider aperture will reduce depth of field (acceptable focus depth) - not sure how that would eliminate jello ... a longer exposure will "blur" each frame more than a shorter exposure. So perhaps this blurring during a longer exposure tends to mask the minute changes in view that occur during a frame's exposure due to vibration (which creates the jello effect). As each frame is exposed over a longer time and is less sharp, stacking these frames in playback is less likely to reveal small changes in the camera's view.

    Another way of thinking about it ... if the camera is shooting 30 fps and each frame is exposed for 1/90th of a second, the change in view that takes place in the unexposed 2/90ths will be revealed on playback. But if the exposure is 1/30th of a second, no change will be revealed on playback because any change in view was "blurred" into the frame.
     
  12. jimre

    Joined:
    Nov 24, 2013
    Messages:
    371
    Likes Received:
    1
    Location:
    North Bend, WA
    As far as I can tell, this camera always shoots at f/2.8. I don't think the lens even has a variable aperture. The camera varies the shutter speed (and ISO if you have it set to AUTO) in order to set the proper exposure. Depth of field isn't really an issue with small-sensor, fixed-focus cameras like these. Everything from about 6 feet away out to infinity should be in focus - even shooting wide open at f/2.8.
     
  13. iResq

    Joined:
    Dec 23, 2013
    Messages:
    545
    Likes Received:
    2
    Location:
    Annapolis, Maryland
    These cameras are fixed aperture, so shutter speed is the only variable.

    Sent from my DROID RAZR using Tapatalk
     
  14. jimre

    Joined:
    Nov 24, 2013
    Messages:
    371
    Likes Received:
    1
    Location:
    North Bend, WA
    I purchased a filter holder and 46mm polarizer - but weather has sucked here pretty much since it arrived, so haven't had much chance to test it. A polarizer should help darken skies and reduce glare from water. But like all polarizers, you'll have to ensure you're shooting at a 90° angle to the sun for it to work. Our ultra-wide fisheye lens may also be an issue - the polarizing effect works across a 90° field of view, and our lenses capture about 140°. So you may see the sky darkened unevenly across the entire image width.

    I use ND filters routinely on my DSLR for landscape photography, but I'm skeptical of their value for the Phantom. Like I said above, I'm not seeing the kind of prop-vibration jello that might benefit from slower shutter speed & more motion-blur. An ND grad filter might be useful - to balance bright skies with dark foreground - if you have a flat horizon and have good enough pilot skills to aim it precisely (I don't).

    Finally, I believe UV filters are a complete waste. They don't affect picture quality in any way, and just add weight - and an additional reflective glass surface to cause more sun flares & ghosts. The ultra-wide lens means there's little chance of finding a lens hood that actually reduces flares without significantly cutting off some of the picture.

    If you just want to protect your lens for travel & storage - since the filter holder renders the original lens cap inoperative - then get something like this:

    Op-Tech Neoprene "Hood Hat" - micro size. Designed to fit snugly over a lens with attached lens hood - this size works with my P2V camera with filter-holder attached. http://www.bhphotovideo.com/c/product/5 ... Micro.html
     

    Attached Files:

  15. Hiway

    Joined:
    Nov 28, 2013
    Messages:
    491
    Likes Received:
    2
    Al technical photographic considerations aside, I am very pleased with the experimenting I did with my ND filter on the gopro.

    I posted a vid a while back where it was used... but the positive results were not noticeable in the video. In fact, I have stopped using the slip on ND filter I bought- but not because of video quality or lack thereof, but instead due to my improved skills flying.

    You see- the slip on filter wasn't helpful on the quality of footage but it did protect my lens when I hit the granite statue- and cracked the filter rather than my gopro lens.

    So score one for the filter?
     
  16. AnselA

    Joined:
    Jan 2, 2014
    Messages:
    896
    Likes Received:
    1
    Location:
    Declination +8° 11'
    Yes, fully agree with you.
     
  17. BobUnplugged

    Joined:
    Nov 17, 2013
    Messages:
    235
    Likes Received:
    24
    That looks good, but it was shot in the UK. Do you think the filters will work on American waters?
     
  18. Pull_Up

    Joined:
    Nov 5, 2013
    Messages:
    3,684
    Likes Received:
    48
    Location:
    South Oxfordshire, UK
    No. If you wanted our water you should never have sought independence. Can't have your cake and eat it.

    ;)