markr041
-
Posts
892 -
Joined
-
Last visited
Content Type
Profiles
Forums
Articles
Posts posted by markr041
-
-
2 hours ago, bjohn said:
When I watch tests like this, it reminds me how much I dislike gratuitous slow motion. I think every filmmaker should be given a lifetime allocation of 10 minutes of slow motion that they can use judiciously throughout their career. It's like exclamation points: I remember reading a diatribe against them in which the author said that every writer should be given a lifetime supply of exactly one exclamation point.
I completely agree with you!!!
Limits on zooms and pans also!!
Here is my video that exceeds my quotas showing the slow motion capability of the fx30 (dancing!) and an extended zoom using clear image zoom plus power optical:
Really, it is just to test, but I actually like the dancing in slow motion.
-
-
-
-
Native WB on the GoPro is supposed to give you more color information, but you have to white balance in post. So, I tried out this setting. With Native WB, Protune Flat, 10bit color, and the 180 Mbps bitrate, this is the best the camera can do. Oh, it is 5.3K at 60 fps also. Linear mode, so no fisheye.
-
This is about bitrates, right?
Well, when might one need 12bit color and high resolution (8KDCI) and no compression = high bitrates?
Capturing Fall colors in sunlight!
You want to capture all those colors (12bit); you need to shoot the details of leaves blowing in the wind, so long GOP compression, which tries to track differences in subject's position, are really stressed; and there are lots of details. And, in Fall sun, highlighted backlit colors and deep shadows need to be caught without blowouts and without crushing. So you want to produce HDR videos, which at a minimum require 10bit color.
So, where do I draw the line - 2600 Mbps = Canon CRAW (R5) 8KDCI.
Watching in HDR is a treat! But not cinematic.
Computer disk space for storage is cheap; fast disks for real-time editing not so much. My MacBook Pro eats RAW for lunch using DaVinci Resolve Studio.
And, no, in a few hours of shooting the camera did not overheat, but it did run down the batteries fast (batteries are cheap). Cfast cards are not cheap. I used up 500 GB's of Cfast storage in shooting these two videos and exhausted about 1.25 batteries.
-
Hero 11 + Anamorphic Lens + ReelSteady in 10bit color and 180 Mbps:
The workflow: Shoot 5.3K 60P Wide 10bit color at 180 Mbps with 1.1 zoom, no stabilization. Then process in ReelSteady, which stretches and stabilizes. Then color grade and decompress more in Davinci Resolve. ReelSteady exports in Cineform 10bit 422 color in 5312x2320 at a very high bitrate so little loss.
-
14 hours ago, Emanuel said:
Thanks for still reporting it : )
Are we able to set and switch off BOTH screens in the front and rear too?
No overheating on 4K 120fps end either?
The front goes off as a setting - ie, never on. The back turning off interval can be set, like after one minute.
I have not used 4K 120p for long intervals. For normal shooting 4K 120p did not overheat - short interval shots in a long session with the camera always on.
-
-
I experienced another "overheating" test: I recorded an organ recital at 5.3K 10bit color for one straight hour in a hot church with no airflow, with the camera inert. The camera never shut down (of course, the screen turned off after a few minutes).
Here is an excerpt:
-
4K 60P Crop Mode:
-
15 hours ago, kye said:
A screen grab works as well, in a slightly different way.
Here's some fringing, but it does seem to be well controlled:
All lenses will exhibit some degree of CA - it's just a limitation on the various physics involved and especially on such wide angle lenses as these. It's not a criticism or attack. Also, people add CA in post when emulating film, so it's not always even a negative thing.
The above grab shows slight over-sharpening, as does the below:
It also depends on what you think the ideal amount of sharpening is, which is subjective. Personally, I prefer a high-resolution but low-sharpness image presentation, but everyone is different, depending on the needs of their project.
I saw the slight CA too.
Here is the full-resolution frame grab (the above was downscaled to 5K, this is the 5.3K) without any sharpening in post (the above had sharpening applied). The frame was shot with sharpening at "low" in camera.
-
14 minutes ago, kye said:
Looks like you missed both points I was making.
You are correct, I do not get what you want or why. What is the point of blowing up the images in a way that no one will view them to literally pixel peep? You can look at the actual images on your 4K monitor with your eyeball 1 " from the screen. I post examples of videos - do you really think they are over sharpened? Do you really see purple fringing? I don't mind you pointing out actual issues in the videos. Look as close up as you want. Or use a 4K projector to blow up the 4K videos to 110" and look close to the screen. Then look at some Netflix film and compare resolution and sharpness.
Or do you want a 5.3K frame grab to blow up, so no YouTube processing? This is a prime test of purple fringing. A 5.3K jpeg of a 5.3K frame shot Wide with Low sharpening in camera and at 200 Mbps, uncropped, not blown up.
In this frame grab I see no purple fringing, I see no over-sharpening halos, I see plenty of detail. I just do not get what you are talking about. But I am happy for you to point out in the posted images/videos the flaws you claim.
-
3 hours ago, androidlad said:
Comparison of the glass lens mod and original plastic lens:
The stock lens is evidently more hazy and less detailed, demonstrated here and by some of the face crops in a previous reply.
Look at the power lines, poles and especially the inscription text.
I certainly easily see a difference, but I don't understand this comparison - the "Original" lens is shot in Linear mode, which crops the sensor, and the mod lens shot is not cropping, it is shot in Wide mode. So, the original-lens image is blown up (zoomed) from a crop of the sensor to have the same FOV as the glass lens, which is not a crop. You can tell the "original" image was a cropped since there is no bending at the edges. That can account for the relative softness. Just explain why you cropped the image for one lens and not the other. Or are both crops? By how much for each? What is going on exactly?
-
7 hours ago, kye said:
The existing lens may well be soft compared to a high-spec alternative, but I'd suggest that the images from the GoPro are bordering on over-sharpened already, so if the lens is swapped for a sharper one then that is likely to be a net-negative on the overall image quality.
It might, however, reduce chromatic aberrations and fringing etc which would be an improvement.
I'd be curious to see a comparison between the stock and third-party lenses.
I'd also be curious to see a 1:1 crop on the 5.3k files you're recording - could you post a short clip to YT? Instead of downsampling the 5.3K file onto a 4K timeline could you simply put the 5.3K file onto the timeline at a 1:1 scale (which would crop part of the original image). If you're able then including a 2:1 where it's zoomed in to 200% would also be nice.
I do not downsample the 5.3K-shot videos to 4K in post. The full-resolution video is uploaded. That is YouTube downsampling. Given that YouTube does not display 5K video and recompresses, I do not think there is anything to be learned about resolution from watching YouTube videos.
When sharpness is set to low in camera I don't think the images are oversharpened. I do increase sharpness in post a bit. I do not see much purple fringing either. Resolution and these are just not the real issues of GoPro video. All the complaints are about fisheye bending, and the lens swaps offer longer lenses that are rectilinear. They are cheap glass lenses designed for old cameras that could not shoot 5.3K video that I doubt offer high quality.
-
-
1 hour ago, androidlad said:
The stock lens on GoPro is plastic with evident haloing and softness. A third party all-glass lens mod is available (Hero 10 for now):
Any evidence you have that this "all-glass" mod is actually better? This is just a claim. I don't see "softness" in GoPro videos. That is the least issue of GoPro image quality as far as I can see.
Post a video showing the alleged improvement.
-
8 minutes ago, kye said:
What was the ambient temperature during the test? Please tell me it was above 35C 🙂
No, it was not very hot, about 65 F.
-
And now my Hero 11 overheating test. 2.25 hours of shooting 5.3K light trails at night.
ISO at 100. Shutter at 1 second. Wide lens, Natural color, 3200K WB. Video is straight from the camera. No editing or processing.
The camera did not shutdown. I stopped recording at 2.25 hours, with 11% battery remaining from full.
4.5-minute video.
-
-
-
-
-
Sony FX30 (S35 FX3)
In: Cameras
Posted
I agree you can see the drop-off in 120P.
The 20mm f2 lens (night video) is nice glass, as is the 16-35mm (day video with slow motion), both FF. And with an APS-C sensor one is seeing mostly the center of the lenses where resolution is best. I do not think G-Master would make a big difference.
60P captures motion well. 24P does not. I do get that for telling a fictional story, 24P has some magic. But for videos aimed at conveying a sense of being there, where there is motion, 24P makes no sense to me. We all have our taste.