Upgraded Lens Tests !

Eagle's Eye Video · Jan 20, 2017

OK, now the real work begins. Some background first... I have 2 CGO3+ cameras from 2 Typhoon Hs
that I own. What I call Camera 1, is my primary camera. But my original camera that came with my H,
got bricked trying to get updated to the October 23 Massacre that took out so many innocents...

So Yuneec cross-shipped a brand new OEM camera and that is Camera 1.

Camera 2 is the camera that was part of a Typhoon I purchased from another member on the board.
Paperwork was included, confirming that Yuneec USA CS had done some maintanence on the H, including
a GPS module and camera replacement. Plus their checklist that everything passed quality control.

Specifics of parameters used for this test... Original Yuneec OEM lenses in both cameras. Both shots
are static, sitting on a table in front of a standardized lens sharpness target that is 2 feet square
and printed on a high quality oversize inkjet printer at 200 DPI. (This was the best resolution I could
find online.) Lit by natural light from an open sliding glass door to a balcony.

Settings on both cameras the same... auto WB locked... in this testing the specific color rendition of
each lens is irrelevent to the specifics of sharpness. This was shot in "Natural" image mode, with manual
settings on the camera of 150 ISO, shutter speed of 1/60. Original files are full 12MP DNGs. Sharpness setting
on the CGO3+, is at what seems to be the default setting of "6".

Original lenses were tested at a distance of 48" from the target, (being measured from the target to
the OEM UV Filter)

I have an older version of Photoshop, and I cannot open the original DNGs, but I convert them to uncompressed
TIFs, which should not result in appreciable image degredation. Once I open them I cropped the image to just
the target area. This intial test is not intended to test corner to corner sharpness. I also changed the image
to Greyscale... color rendition is not a factor. The only other manipulation is adjusting levels to give an
accurate representation of the original target. No sharpening or other image enhancement was used.

45B1F785-24C2-4FDC-9547-C62BE29EDE91.jpeg

Firmware

View media item 409
As you can see, there is a definite difference between these original manufacturer lenses. I actually was not
at all displeased with the quality of my Camera 1. However, based on the quality of the lens in Camera 2
(which had passed QC from Yuneec USA), I can see why the member gave up on his H...

What I have now is the following:

Camera 1 now has the PixAero 3.77 lens... Camera 2 now has the Peau 8.25... to attempt to keep this an
"apples to apples" comparison, in this second set of photos, the camera was moved back to approximate the
area of the sensor that is capturing the sharpness target. In the case of the PixAero 3.77 it was set at
a distance of 55". In the case of the Peau, it was set at a distance of 120" (10 feet). All other parameters
were set as described above, with the original lenses.

F814D683-A0B8-46DF-AFAB-06CCD78A4396.jpeg

Firmware

View media item 408
Difference, huh? PixAero has fantastic resolution, and I think the Peau is close to my very well focused OEM camera 1.

For anyone that cares to see them, all original TIFs are here:

Download file lens_tests.zip

And anyone that really want to see them, LMK and I'll put up the original DNGs... not trying to hide anything,
but I did want all lenses to have an even playing field.

Next up... the same sharpness target mounted on foamcore, and taped to an exterior wall of my apartment building.
I will hover the H in front of the target, for both photos and a brief video... 15 - 20 seconds should be more
than enough.

RiverRunner · Jan 20, 2017

I find this target very difficult to judge as far as sharpness. Maybe it's just my old eyes. I also think you might want to test sharpness over the whole frame and at a longer distance. I don't think you are going to be shooting much at 4 feet. I refocused my first camera using objects much farther away such as a brick house across the street. That camera failed later and was replaced. The replacement was focused very well out of the box.

Rubik · Jan 20, 2017

Interesting test. Looking at your first 2 target shots with stock CGO3+ lenses full screen on my 4k monitor, camera 1 has a little vignetting down near the center that is not apparent from camera 2. However, camera 2 has more noise in the smooth areas of gray and black. The 2 aftermarket lenses both show the vignetting and image noise. I don't find these results very conclusive. I'll be interested to see your outdoor test. Will you give us color?

Eagle's Eye Video · Jan 20, 2017

Oh, I intend to, but this was the closest I thought I could get as far as an indoor test and still have a reasonable focus... the rest will be under more real world comparisons with the cameras in use both on the H as well as with the Action Grip. The differences are more apparent in the TIFs (at least to me). And yes, everything else will be color.

gwhuntoon · Jan 20, 2017

After seeing this comparison between Stock, Peau, and PixAero, I've been leaning toward the PixAero.
https://www.youtube.com/watch?v=tVfQzVqhgWI

PatR · Jan 20, 2017

There's a produce packing plant across the way that keeps stacks of 4x4 grated bins outside. I use them for checking barrel distortion.. When the weather breaks I'll set an H in a table with the stock and PixAero lenses and shoot some jpg comparisons in Natural. I've bumped the sharpness up to 8 on the stock lens though so I'll shoot some 6's and 8's with the PixAero.

Sent from my iPhone using Tapatalk

Rubik · Jan 20, 2017

PatR said:
There's a produce packing plant across the way that keeps stacks of 4x4 grated bins outside. I use them for checking barrel distortion.. When the weather breaks I'll set an H in a table with the stock and PixAero lenses and shoot some jpg comparisons in Natural. I've bumped the sharpness up to 8 on the stock lens though so I'll shoot some 6's and 8's with the PixAero.

Sent from my iPhone using Tapatalk

Excellent. From some of the videos that I've seen, the Pixaero seems to have better color rendition and flare rejection than the stock or Peau lenses.
How does the sharpness setting work? I assume it's not like focusing, but more like sharpening in post-processing software, intensifying edge contrast. Over-sharpening can be distracting also.

yuneec53 · Jan 21, 2017

On the Peau lens site regarding the 8.25 lens it says
If you have the Yuneec black lens mount, it will have no threads to install a new lens, so you also need to purchase our Yuneec Extended Lens Mount.
1/ How do you know if you have a black lens mount Typhoon ? Is it from a certain date or serial number or do you simply have to look inside?
2/ If you purchase an 8.25 lens with the Extended lens mount you simply take out the two screws holding the mount with the old lens in it and remove. Replace it with the new mount and screw the new lens in.
3/ As this mount is only $10 it would be a better recommendation to buy one anyway to avoid the hazardous chipping off of the old glue with all its dangers. Plus if you need to put back the old lens it is complete in its holder a per the factory setting. No refocussing or other problems. I have quizzed Peau on this.

PatR · Jan 24, 2017

Due to issues with the site's software I cannot post the high resolution H photos here but a link to Dropbox files is posted for those that want to look. The photos are the initial shots taken using a stock camera lens and a PixAero 3.77 lens. They are not the same camera, nor the same firmware version. The PixAero has the latest firmware, the stock lens has the previous firmware version. No filters were used. Total time span between photo 1 and photo 5 was ~15 minutes. Changing cameras required binding the camera each time it was changed.

The stock lens camera has had the sharpness adjusted from 6 to 8 in the camera software. The PixAero camera is still set to 6. The gate posts are 25' from the camera lens, the fruit transport crates are ~150'. It was pretty cold so I didn't spend much time with it. I used auto white balance simply because it was there and auto to auto might have something useful to contribute in the comparison. Of interest is the PixAero has essentially the same image sharpness at the stock sharpness level of 6 as the stock one does bumped up to the 8 setting. The lens position was slightly different between the stock and PixAero shots but the distances were the same. The power poles are not true vertical or aligned so don't use them to judge barrel distortion. Camera was used for the shots but video was set to 4k to obtain the higher resolution. That's the part the site software doesn't like.

It was clear to me that settings we might be used to when using the stock lens will not directly transfer to the PixAero lens. To be expected and not a big deal, just another learning opportunity. I did like the way the PixAero lens rendered color in photo #5. That is extremely close, if not an exact match, to what things actually are.

Not a terribly scientific test and the light was flat, flat, flat. Histogram had no pixels at either end of the black/white scale, it was all center loaded.

#1: Stock lens, ISO 100, 1/400, Natural, Sunset/Sunrise, EV-0
#2: Stock lens, ISO 100, 1/400, Natural, Auto W/B, EV-0
#3: PixAero, ISO 100, 1/400, Natural, Sunset/Sunrise, EV-0
#4: PixAero, ISO 100, 1/400, Natural, Auto W/B, EV-0
#5: PixAero, ISO 200, 1/800, Natural, Auto W/B, EV-0

Dropbox - PixAero Comparisons

Rubik · Jan 25, 2017

Thanks, Pat for doing this test and sharing.

I see what you mean about sharpness being virtually the same between the two lenses (at different sharpness settings). But, boy, the white balance is all over the place. I've noticed that with the latest firmware (or somewhere along the way) the WB and exposure change very slowly in AUTO modes. Watch the histogram move very gradually to a steady position. This is good. But, it may mean that the software was still adjusting WB when you took some of these pictures. For example, there really should not have been such a large color difference between #4 and #5. Were these pictures taken in video or camera mode? Were you saving in .jpg or .dng?

I did some ad hoc testing myself, but I was looking at which settings produced the best images with my stock CGO3+. I got the best results (least noise, sharpest image, most natural color) by taking pictures in PHOTO mode (not video), Natural or Raw (Raw with post processing) and saving as .jpg files. The images saved in .dng were very noisy and the file size was huge.

PatR · Jan 25, 2017

All photos were jpg. Only the camera function was used to take the photos. The video setting was only used to establish the higher resolution that's provided to photos when video is set to 4k. No photos were taken with the cameras set to video mode, all were done in camera mode. Photo #5 I selected ISO 200, shutter speed 1/800 but looking at the file properties there is something odd as it shows ISO 100 and a shutter speed of 1/301. Photo #4 used ISO 100 and a shutter speed of 1/400.

All the photo dimensions are 4000x3000, 72dpi, bit depth 24, but interestingly photos 1 and 2 with the stock lens camera are 2.72MB while photos 3, 4, and 5 with the PixAero lens camera are 5.4MB. Looks like some Sherlock work is in front of me.

With any luck the weather will warm up enough this week end to where I can be a bit more comfortable and spend more time learning about the PixAero lens. It certainly has differences with the stock lens that indicate I'll need to adjust setting up the shot a little differently. Not surprised by that. Changing lenses on my DSLR's also generate the need to adjust things differently between one lens and another. The use of auto white balance was simply because it was there, not because it might have been better. An apples to apples sort of thing, which thinking back makes me think using it probably wasn't apples to apples after all. One thing was certain, where I typically use the Sunset/Sunrise W/B setting when shooting I found it didn't work well with a scene that was heavily loaded by white and shades of white subject matter under overcast conditon flat lighting.