Welcome to InspirePilots.com

Join the leading DJI Inspire community for free!

What do you think about this test?

Discussion in 'Zenmuse X5' started by Djalma Ribeiro, Jan 20, 2016.

  1. Djalma Ribeiro

    Joined:
    Jun 26, 2015
    Messages:
    50
    Likes Received:
    12
    Location:
    Cuiaba, Brazil
    What do you think about this test?

    It's not mine, just found on youtube.I do not understand what they say or write.

    Compare: Phantom 3 Standard, Phantom 3 Professional, Zenmuse X3, Zenmuse X5, GoPRO Hero 3, GoPRO Hero 4, Panasonic GH4

     
  2. DrMrdalj

    Joined:
    Jan 18, 2016
    Messages:
    81
    Likes Received:
    51
    Very nice video, but there are some details of test procedure to be reconsidered and, possibly, result to be questioned...

    All cameras have best dynamic range and noise performance when they are set to native sensitivity of sensor - native ISO. I know for sure that GH4 has native ISO of 400, I asume X5 to be 400 as well. None of these cameras have native ISO of 100, so Dynamic range results are questionable...
    Resolution test depends heavily on lens perfromance - we do not know which lenses were put on X5 and GH4, so results are unconclusive... Not to mention differences in codec, bitrate and so on, which are not documented.
    Rolling shuter test is directly dependant on angular speed and focal distance, which are not documented either...

    So, what we have learned from this video is what we already know by "naked eye" - Hero4 and X3 and Phantom 3 Pro are in one league, X5 and GH4 are one big step above, while basic models are one step below in performance...
     
  3. DaveInspire

    Joined:
    Mar 20, 2014
    Messages:
    167
    Likes Received:
    15
    My understanding is native iOS is 200 for the x5
     
  4. Luke Neumann

    Joined:
    Jan 20, 2015
    Messages:
    59
    Likes Received:
    40
    I would really like an answer from DJI on this as they know for certain. I have always assumed it was 400 as well.
     
    DrMrdalj likes this.
  5. DrMrdalj

    Joined:
    Jan 18, 2016
    Messages:
    81
    Likes Received:
    51
    You could be right, but my testing and comparison made me believe that it is 400. I did not find any official data, but it would be technicaly very challenging to make such small sensor with native ISO200 to perform without obvious noise under ISO1600 and above, what X5 succeeded at least as good as GH4, if not better (GH4 has native 400)... Do you have any reliable source for such info?
     
  6. DaveInspire

    Joined:
    Mar 20, 2014
    Messages:
    167
    Likes Received:
    15
    There are a lot of claims and no official report. Interesting, many claim gh4 native is 800. In any case, I think the safe takeaway is don't use 100. And 200/400/800 go with what looks best to you or what light conditions allow in that range.
     
    DrMrdalj likes this.
  7. DroneTechAerial

    Joined:
    Sep 26, 2015
    Messages:
    70
    Likes Received:
    19
    Location:
    Malibu, CA
    It is my NEW understanding that you want to use multiples of the native ISO to reduce the . I used to think to dial down the ISO as low as possible to reduce the amount of gain that the camera was adding to the image. However come to find out that if you operate below the base/native ISO the camera goes to the native ISO and then adds in negative gain. Is my understanding of this correct? If so I need to get some stronger ND filters to kick my ISO up to the base and then adjust exposure with the aperture. And as of now there is no native ISO published by DJI?

    DrMrdalj or DaveInspire Do you never shoot below native ISO?

    As a side note about the video. It is hard to believe that the X5 is really better than the GH4. It has been my experience doing a side by side to my eye and my editors eye comparison with our own footage that the X5 with the same lenses is about 80-90% of the GH4. However I am not a pixel peeper and I have only been doing this for 3 years. I mean DJI, a company that routinely builds things and releases them to the public without being fully tested beat out Panasonic in the camera game? I could be wrong though.... I think the KGB and DJI are working together to develop propaganda pieces...
     
    DrMrdalj likes this.
  8. DrMrdalj

    Joined:
    Jan 18, 2016
    Messages:
    81
    Likes Received:
    51
    To my best knowledge - there was no official statements on native ISO of X5 so far... I believe that has something to do with not yet released X5R and D-Log updates, as native ISO is more important for RAW&Log workflow then for regular more processed and compressed color. While noise artifacts in CinemaDNG/RAW workflow come from the sensor, fIlming 4K in MP4 compression at 60Mbit introduces much more of "codec noise and blocking"... Therefore, I personally believe that for X5 (not raw) filming in native ISO is not of paramount importance - any ISO from 100 to 800 will get fair results...

    For best dynamic range (sunsets, strong contrasts etc) I tend to use ISO 400 with ND filter when appropriate (in strong light) or ISO 800 when I strugle for greater sensivity (dusk, dawn)... In strong light, I use ISO 200 as well, but I can not remember if I filmed with ISO100 - as I usualy use ND filters to correct exposure...

    So, before flight I measure light of the scene and put appropriate ND filter to get proper exposuer in average shot with f5.6 ISO400 1/100... When I want to get more light I increase aperture by 3 stops to 2.0, then I get 1 stop by lowering shutter speed to 1/50 and 1 stop by incresing ISO to 800 - 5 stops in total. In the opposite direction, when I want less light in the picture, I close iris by 2 stops, lower ISO by 1 stop and increase shutter speed by 2 stops to 1/400 - 5 stops in total... I think that ISO100 will do OK but I did not test... This is by no means scientific or final method, it is just what I do and I consider to be good practice...

    Regarding comparison of X5 to GH4 (which is my favorite camera) - pleas note that overal image quality depends not just from lens and sensor but from image processing engine (Panasonic has it own proprietary) as well as from codec and bitrate (at 4K GH4 has 66% MORE bitrate - 100Mbit)... So, when you say that X5 delivers 80%-90% of GH4 quality, I believe that you are quite right.
     
  9. DroneTechAerial

    Joined:
    Sep 26, 2015
    Messages:
    70
    Likes Received:
    19
    Location:
    Malibu, CA
    Please consider that I am asking these questions based on the assumption that I am filming rather than taking photos, which the rules don't really apply. These are just questions not saying you are doing it wrong just looking at your thoughts.

    Any reason why you are not following the 180Deg shutter rule with your shutter speed by shooting at 1/100 and 1/400?

    Also it seems that you are opening your aperture way up lending yourself to vignetting. I tend to never go that low to preserve a deeper depth of field and keep the image sharp by shooting between F4 - 5.6, my usable range is between F2.8 - 8 (Olympus 12mm). I am also using ND's to control light as well.

    I never let my ISO go above 800 even at night/dusk I just feel there is too much noise. I have been running between the ISO 100 -800 however I think I am going to stop using ISO 100 and keep the range between 200-800. Can running the ISO down to 100 induce any more sensor noise in your opinion?
     
  10. DrMrdalj

    Joined:
    Jan 18, 2016
    Messages:
    81
    Likes Received:
    51
    I strongly believe that ISO100 does not induce noise, it just lowers maximal dynamic range of your shot. So, ISO100 is usable, while ISO1600 is almost unusable, I agree. Please note that I shoot in 4K but mostly deliver in 1080p, so some noise gets smoothed in donwscaling...

    180Deg shutter is just idea of what pleases average eye - depending on motion blur look you like and depending on motions in your shot, you can modify this rule freely. When filming landscapes I do not go for mortion blur but for clarity, where 1/200 shutter is common for video - look for DJI official guide how to set up GH4 on S900+Zenmuse and you will find 1/200... When I shoot close ups of moving cars or other very dynamic motions or close subjects then I tend to use 180deg rule (1/50) as well, but these are rare occasions for me... In some slowly moving evening landscapes I filmed even with 1/25 and got results which I consider nice...

    When filming X5 in 4K I can not notice significant vignetting - because sensor is croped to central part (crop factor of whole sensor is 2.0 but sensor has more then 4000 pixels in width so for 4K 1:1 pixel readout crop factor gets to 2.3 or 2.4). Same applies to GH4. You can easily notice that additional crop when switching from 1080p to 4K... If one does not film fully even backgrounds such as clear sky, I find that regular DJI/Panasonic 15mm/f1.7 lens set at aperture 2.0 does not produce important vignetting with this additional crop (I did not test Oly 12mm, but I got similar results wit Oly 45mm). Depth of filed gets narrower, but consider hyperfocal distance at which you can stop thinking about depth of field - when X5 (and GH4) with 15mm lens at f2 aperture gets focused on 15m everything from 7m to infinity is sharp and clear - so I believe that f2.0 is good for landscapes where I do not ask for clarity of close objects which just fly-by anyway... For close up filming, when subject is just few meters away I tend to use f5.6-f8.0 as well (but then I use lower shutter 1/50 for motion blur, so I get aprox. same light)...

    I find that you and me do not have significantly different practices: ISO 100-800 (priority 400), aperutre 2.0-8.0 (priority 5.6), shutter 1/50-1/400...

    But there are problems with X5 which bother me much more then noise and dynamic range - custom white balance does not work, D-Log and Cinelike does not produce sharp images, some flickering in auto-exposure even in Aperture priority (when camera changes ISO and shutter in very fine steps)...
     
    #10 DrMrdalj, Jan 31, 2016
    Last edited: Jan 31, 2016
  11. DroneTechAerial

    Joined:
    Sep 26, 2015
    Messages:
    70
    Likes Received:
    19
    Location:
    Malibu, CA
    I had no Idea that the custom white balance was not working....Is it not accurate? I see that there is warming and cooling taking place...
     
  12. Luke Neumann

    Joined:
    Jan 20, 2015
    Messages:
    59
    Likes Received:
    40
    It works, it is just not accurate...at all. 5600K is more like 4800K with a magenta shift.
     
    DrMrdalj likes this.
  13. The Editor

    The Editor Moderator
    Staff Member

    Joined:
    Aug 7, 2013
    Messages:
    6,705
    Likes Received:
    3,861
    No - its broken (on the X3 but I haven't confirmed on the X5 - although it's a firmware problem that R&D have confirmed with me).
    The custom setting does not retain its value or even set until you touch the slider. If you change any mode the value is lost even though the slider stays where its put. Touch it again and the value changes to your pre set.
    I was astounded nobody else picked up on this. I NEVER use auto WB or any pre set but prefer to set my own and leave it there.
    Guess everyone is using auto which is dire (as opposed to mired :p)
     
    Floaf and DrMrdalj like this.
  14. DroneTechAerial

    Joined:
    Sep 26, 2015
    Messages:
    70
    Likes Received:
    19
    Location:
    Malibu, CA
    I don't use AWB, it will give too much of an inconsistent look throughout the day. I usually manually set it, however I am not on the editing side of things so I am not seeing how its being graded and corrected. Double wammy for me today, D-Log and WB are not working and I have been unaware.
     
  15. Luke Neumann

    Joined:
    Jan 20, 2015
    Messages:
    59
    Likes Received:
    40
    I'm talking about the X5. It's buggy, to say the least. AWB can ruin shots, although I'm sure many people are in fact using it.

    Yet another reason why I am pining for the X5R. I shoot raw video on the ground so sometimes I forget how much of a luxury the eyedrop WB tool in post production is :)
     
    DrMrdalj and The Editor like this.
  16. DrMrdalj

    Joined:
    Jan 18, 2016
    Messages:
    81
    Likes Received:
    51
    I find analysis of my shots in the post to be very important for learning my tools, their limitations and strenghts. I never miss to watch each and every take on my MacbookPro before I deliver to the editor, and usualy if I had not participated in postproduction I compare my source shotsh to final edited film... In such way I learned that custom WB is way off (it is possible that I am constantly doing something wrong, which is hard for me to believe as I have good results with so many other cameras)... AWB produces more accurate results but can vary during a shot, what is unacceptable... Therefore I rely on presets, to get some approximately good results and then color correct in post... That is not time-saving or best-quality way, but for now I am not aware that there is better one... Is there anyone with better idea?

    Regarding "shutter angle" there is nice video at Videomaker
    Camera Controls & Settings: Shutter Speed & Angle
     
  17. DroneTechAerial

    Joined:
    Sep 26, 2015
    Messages:
    70
    Likes Received:
    19
    Location:
    Malibu, CA
    On the GH4 and other DSLR cameras you could always take a picture of a gray card and set the WB. But on the Inspire there is no such option, so I have always set it by knowing some of the key ones like full sun,shade and sunset kelvin temperatures and then kind of moving it around to my taste, which is really not the best way. So I am also looking into other suggestions or tips to setting the WB with any sort of systematic methodology?
     
  18. Floaf

    Joined:
    Oct 18, 2015
    Messages:
    12
    Likes Received:
    1
    I actually noticed that on my first startup, got an inspire 1 pro a month ago. Wanted to test the image quality (I can't use auto WB, it's horrible to adjust in post). But people has been talking about that it doesn't keep the camera settings between battery changes and things like that, so I assumed it was one of those things. Great to hear it's going to be fixed, it's quite annoying :)