Canon Vixia HF R70, R40 and Other Low-End Camcorders

[VidThreeNorth]

To start this topic I have posted my first video recorded on my Canon "R70" at:

"20180826 Isshin Daiko Japan Festival 2018 Mississauga"


The video was taken with a monopod, using 30 fps, the "Sports" mode and Power stabilization ON. Exposure was EV = +1.0, which was about a half stop under exposed. Exposure was corrected in post. No other changes were made in either the video or the audio.

Post processing used "Magix Movie Edit Pro 2016" on my ASUS M32BF_K30BF, AMD A8-6500 APU, Windows 8.1 system.

[2018-11-28]
One thing I did not mention is that this clip has the best sound that I have recorded on the Canon R70. It is significantly better than the previous clip. It used the built in microphone, but because it was outdoors, there is no indoor reverb. The drum sounds were very clear.

 

[VidThreeNorth]

Canon R70 v. Sony CX405 Image Quality Test:

Comparison frame captures:

This was a specific test of my Canon Vixia HF R70 and Sony HDR-CX405 camcorders.

I set the camcorders on my 3D mounting bar about 4" apart and tried to record matched, synchronized video clips. The results are not perfect, but are close enough for this test.

When I started using the Canon R70, I started by using the same settings that I was using on the Canon R40's. I did this deliberately so I would be able to use the new camcorder without putting up with quirks that might occur when using settings which were "foreign" to me.

Carrying this through my recent testing, so far my tests have used my normal settings, which for the Canon R70 means the "Sports" setting. This has probably been less than optimal -- like having "one hand tied behind its back."

The results are interesting. I had no "expectations", but I knew what I was looking for: As I have mentioned before, I felt that the Sony CX405 seemed less detailed than the Canon R70. As I write this, after my first serious comparison, the results are more complicated than I expected, but the Sony is, overall, better than I thought.

The Test:

As usual, I did not have time to do all I wanted to do, and due to recent weather, I was limited to trying to get it done on a Saturday late afternoon, which meant families, including children wandering around. I limited this test to two pairs of clips. One pair at the nearest wide angle and one pair at the farthest telephoto. Since the zooms do not cover the same range, I used the widest setting of the Canon and the longest telephoto setting of the Sony. From each pair of clips I found a stretch of about 2 seconds where I could identify matching frames, and captured all frames from each clip within that time into PNG's. That meant I was starting with over 120 frames for each test and over 240 frames for the whole project. I then further limited the selections to all frames within 1 second which, again provided identifiable frame pairs.

Then for each clip, I found the three largest image files. The largest files in these comparisons are assumed to be the best images in their respective groups. Starting from the largest file for each clip, I checked if there was a reason not to use the largest. The other frames are important only if I found a reason not to use the largest. I found no reason not to use the largest, so those were chosen for comparison. I then found the matching captured image from the other camera. So each set of comparison images has four:

Telephoto;
- Best from the Sony
- Canon frame that matches the best from the Sony
- Best from the Canon
- Sony frame that matches the best from the Canon

Wide Angle;
[as above]
...

Results:

The Wide Angle pairs show little difference except for the colour balance. The Canon looks warmer but with muted greens. The Sony has a wider variety of brighter greens, but lacks magenta. I was surprised by the similarity.

Wide Angle:
"W01C-19h33m39s256-C1.jpg" - best from Canon
"W01Cm-19h21m11s103-C1.jpg" - Sony frame match for the above

"W01S-19h20m55s839-C1.jpg" - best from Sony
"W01Sm-19h33m27s543-C1.jpg" - Canon frame match for the above

The Telephoto test pairs are more interesting. I think that at first glance, the Sony actually looks better, and for most people it is probably the better camera. Here's why:

Looking at the Canon images the obvious faults are a patch of grass above and to the left of the left goose, which continues off screen, and another similar patch above and to the right of the right goose, which also continues off screen, and the color fringe on the right (front most) edge the chest of the left goose. That colour fringe seems to be is some kind of chromatic aberration. It is surprising to see that it is not showing up anywhere else on these pictures, but for whatever reason, it is quite visible.

Seeing these image pairs, one might assume that big differences in the lenses are causing these differences, but I do not think this is what is happening. In fact, I think the lens performance is probably fairly similar. I think that what is happening is more the result of processing from the raw sensor data to the frames pixels. The Sony is probably cleaning up the colour fringing first, and then adding a bit more sharpening than the Canon.

The patch of blurry grass on the right side of the Canon frames is not in the picture on the Sony frames, so that cannot be compared. But the patch on the left side also shows up on the Sony frames as lower contrast. But in the Sony frames sharpening has brought them up more than on the Canon frames. Likewise, although the colour fringing is not on the chest of the goose on the left, the size and shape of the white area looks like what I expect to see if a "Chromatic aberration correction" tool is used.

So what I think the final images show is that the lenses (and sensors) are working similarly, but the processing is working differently. I do think that most people would prefer the Sony, but personally, I consider them about equal.

Telephoto:
"T01C-20h09m11s382-C1.jpg" - best from the Canon
"T01Cm-20h20m26s672-C1.jpg" - Sony frame match for the above

"T01S-20h20m07s396-C1.jpg" - best from the Sony
"TS01m-20h08m39s881-C1.jpg" - Canon frame match for the above

I do not think that I will try another "matched pairs" test, but now that I know the issues with the Canon, I think I will try to see if there is a setting which brings its performance at the telephoto end up.

 

[VidThreeNorth]

Cold Cameras

With snow already a problem, particularly in the East, I thought I would post a bit about camcorders and cameras in the cold.

I do not really have a huge amount of current experience in cold weather. I am not able to handle real cold for very long. And I never did believe in trying to use "ordinary" cameras in the cold.

1. LG 3D Android Smart Phone (LGP925G). This is significant because it had problems. The temperatures I tried it in were not that cold. Just a bit below freezing. The screen display was working, but the "touch" did not. I was recording a video and I could not stop it. Eventually I decided to pull the battery. Be aware that at least the older capacitive touch screens may have this problem.

2. Canon Vixia HF R40. This problem occurred in the Fall of 2017. I was recording a Remembrance Day ceremony at Toronto's City Hall and the temperature was just below freezing (light snow eventually fell). First, the screen display stopped working. The backlight was on but the LED pixels would not darken. I think the screen technology might not be Active Matrix. It certainly does not look like the Sony screen. Also, again, the touch screen stopped working. I recorded the ceremony essentially "blind". I did not even know if anything was being recorded. In fact, everything recorded normally. It might be possible to control the camcorder through WiFi, but then your phone might not be able to work in the cold either. The Canon Vixia HF R70 appears to use the same LCD and touch surface as the R40, so I am expecting the same problems, and am not planning to use it in the cold.

3. The Sony CX405. I have no experience with this camcorder, but its controls are all "real" buttons and joystick. There is no reason to expect any problems.

4. Yi Technology YiM1 with Kit Zoom. I used the camera below -15 degrees C. The screen displayed normally and all controls that I needed worked without problems. The only problem was the zoom. In the cold, the Zoom ring became almost unusably "tight". I decided to use it anyway. I gripped the ring and forced it into positions. The video turned out well. The only problem is that since that event, the zoom moves much more freely. That is not problem at all, since I felt that it was too tight to begin with. The zoom has not had usage problems yet. But I am expecting it to have problems later.

5. Panasonic Lumix GF-3. No particular problems. I have used it at and a bit below -10 degrees C. The only problem I have is that the camera body is cold and slippery.

 

[VidThreeNorth]

Canon R70 Minimizing Problems:
More Geese (and Ducks) in a Park:

An attempt to reduce the problems found in the previous test and specifically trying out the "touch screen point auto-focus".

"MVI_0007.MP4"
[Video not uploaded]
Created Oct 10, 2018, 17:00:26
Size 890,276,001 bytes
Video:
Length 00:04:59
Frame width 1920
Frame height 1080
Data rate 23553 kbps
Total bitrate 23809 kbps
Frame rate 29 frames/second [sic]
Audio
Bit rate 255 kbps
Channels 2 (stereo)
Audio sample rate 48 kHz

The frames I uploaded before in the "side-by-side" telephoto sets were exposed without adjustments and as a result, all were far overexposed. The point of the comparison was to see what the camcorders would do in exactly the same situation -- not necessarily when they were perfectly adjusted. If I could have practically made all the proper adjustments to each camcorder, I would have. But the situation I encountered did not allow that. Yet I felt there was enough to be learned from the results to make them available.

The question of what exposure adjustments you make start with knowing how much dynamic range with which you have to work. The geese are particularly good subjects for this because they have near white bellies and near black heads. If I darken the exposure to "contain" the highlights, then I might lose the heads. My experience has been that this sometimes happens, and overall I do prefer to keep the heads and let the highlights clip.

[2018-10-19 13:45 added this paragraph.]
The geese are really an unusual subject matter. I should emphasize that. If you look at the "Isshin Daiko" video linked in the first post, that video is a more typical "high contrast" video. In the final result, there was no clipping in the highlights and yet there is sufficient detail in the black clothes. I did have to re-adjust the overall level in post, but the dynamic range had sufficient coverage for it.

In the earlier test, both the Sony and Canon overexposed the chests, but the heads were all pretty good with no "0-0-0" blacks in any of the frames I posted.

I returned to the park intending to further test the R70 and see what I could do about the problems I had found. I made a few clips that I liked, and in particular, I had an opportunity to use a very similar situation to make comparison images from.

Clip "MVI_0007.MP4" was made at a camera location close to where I had set up before. The main difference is that this time I was using my mono-pod instead of my tripod, and there was no other camcorder being operated at the same time, freeing me to make adjustments.

The first change I made was to record using "P" instead of "Sports". This means I was using the R70 with more typical settings. The "Sports" setting prefers faster exposure speeds to stop action. In good lighting, the "P" setting should give a better balance by allowing smaller F-stops which, at the base ISO, require using a longer exposure time. In very bright light, this might not make any difference at all, but on cloudy days, or as the sun goes down either exposure times will get to the point where motion blurring becomes a problem or ISO might be raised.

The second change was to try to follow the lighting situation with EV compensations.

The third change was I had decided to use the "touch screen point focus" where it might help.


The Frames:

The first three images both confirm the issues I found in the Canon lens in the earlier tests but to an extent vindicate it at as potentially usable in the longer focal lengths.

The first two images:

"01-MVI_0007-00_28-20h49m49s790-C1.jpg"

"01-MVI_0007-03_48-20h12m29s674-rsz1892-C1.jpg"

The first two images are a familiar subject -- geese. And in the same general area where I made my earlier tests.

The first image was taken with the sun a bit shaded by clouds. The exposure is EV = -0.5. I correctly estimated that dropping the exposure further would lose the head. Checking this frame I found that there was a very small amount of "0-0-0 black", though possibly only 1 pixel. But certainly if I had dropped to EV = -0.75 or below, the head would start "going away". The goose's rear was already clipping, so there was no real room for choice. Overall, the grasses look much more detailed. But the softness to the far left and far right are still there, confirming the lens has problems.

The second image looks almost the same as the earlier tests. In this case I have compensated the F-stop further to EV = -0.75, but the cloud shading has gone away, probably completely, and it needs a further compensation, purhaps EV = -1.5. But I was busy watching the focus and could not be jumping back to the exposure adjustment. Here, the magenta-ish fringing shows up in the head, and also near the tail. The area near the tail is puzzling because it is both wider than mere "optical" fringing should be and appear to have a distinct edge. I do not understand that. Surprisingly, this frame file was larger than the earlier one, and in fact, too large to be uploaded in its orginal "FullHD" size. The C1 compressed file was 2,093,891 bytes. After looking at various possibilities, I decided this reduced size version (1892 pixels wide) was the best choice.


"FZ-01-MVI_0007-02_04-10h13m21s208-C1.jpg"

This was intended to be my first "full zoom" test, and in fact it is -- but not focused at the full distance. There were branches from another tree a bit nearer to me which screen the red leaves. I tried using the touch focus to see if I could get past the branches, but it did not work. I really should have tried the manual focus for this, but I decided to stay within my planned limits for this set of files. Manual focus will have to be tested on some other day. Still, the zoom is at full telephoto and the image is reasonably usable, possibly due to a smaller F-stop.


"G-Ducks-01-MVI_0003-13h43m52s964-C1.jpg"

This is from an earlier clip when the clouds were giving a bit more shade. It was the first test of touch focus. For this frame I had focused where the male mallard is facing left in the middle of the group of birds (about half way from the camcorder to the near edge of the pond). As I was recording, the female walked from left to right just a bit nearer. The grasses in the area look a bit more detailed than the grasses nearer or farther. I cannot tell which bird is actually in perfect focus (if either are). Depth of field has carried at least to the near edge of the pond.

[2018-10-19 13:30 added this paragraph.]
Results:

Was I Able to Improve on the Lens Performance:
I mentioned that I felt that smaller F-stops from using "P" instead of "Sports", and the exposure compensations, should improve the telephoto end lens performance. I think it did, but not enough to make the long telephoto settings acceptable. Both of the new frames for the goose look a bit better than the earlier video, but there is still enough off-axis softening to be noticeable. The colour fringing seems to be less as well, but it still sometimes visible.

Overall, I would say that the "touch screen point focus" is working fairly well. As for the lens, as I think I have mentioned before, I believe this lens was probably designed mainly for 720P ("HD Ready") use and I have found its performance in "Full HD" is not entirely successful. I found this before when I tested the Canon R40 and the Sony CX240 previously (conducted a few years ago for my own use and probably never reported). The reason I like the Canon R40 and R70 despite the lens deficiencies is because I have never used the zoom out that far into telephoto except for these specific tests. The 32x optical zoom is roughly 1000mm equivalent for a 35mm camera. The only way I could use it would be on a tripod, and I rarely use a tripod. So really, I do not use it beyond around 400mm to 500mm (35mm camera equivalent), and the colour fringing has not been showing up.

On the other hand the reason I have not used the touch focus, manual focus or exposure compensations on the R40 was because I mainly limited my use of that camcorder to 3D video, and one cannot make such adjustments on both camcorders at the same time. The R70 is going to be used for 2D and I will be able to use these controls which make this camcorder a fairly complete low cost video recorder, and I am looking forward to this possibility.

The future of these low cost camcorders is doubtful. I do not expect them to be further developed because there is a mistaken believe among customers, that 4K is a "necessity". I do not know if good 4K camcorders will become available anywhere near the prices of these lower end "Full HD" camcorders in the near future. So what is available now is probably going to have to suffice for "a while".

 

[VidThreeNorth]

Canon posted a sample video from the R82 (and R80 and R800) camcorder to YouTube back in 2017. There is no camera recorded sound. A music track has been substituted.

"Canon VIXIA HF R82 Sample Movie", by "CanonUSA" Apr 5, 2017,

 

[VidThreeNorth]

Canon has announced their HF W10 and HF W11 rugged camcorders. These appear to be based on the HF R80 series and will cost $399.95 US and $449.95 US respectively. From my experience, these need to have new display screen technology compared to the R80, in order to be usable down to -10 deg. C. The screens are 3". Also there is no information about whether the battery is removable, which is a key drawback of the JVC camcorders.

DPReview.com article:

Canon reveals VIXIA HF G50 4K UHD, W10, and W11 camcorders at CES 2019

This article also announces the Canon Vixia G50 which is Canon's first "low cost" 4K camcorder based on a 1/2.3" sensor and 20X zoom costing $1,099.95 US.