3

There are 2 different things: The real width and height of frames (in pixels) - see the green part of the folowing picture. The displayed width and height (during the playback) - see the blue part of the same picture. The problem arose in old bad days, when the resolution of devices (number of pixels in the image) was very low, but the required ratio of ...


3

If the stills look sharp and detailed, then the sensor and optics are not to blame. I would blame the encoder and whatever sacrifices Canon made for the video mode. The 80d has the same Digic 6 processor as the older and cheaper models like the T6i. I have an older model T4i, and it is very low-res in 1080p mode, while in 720p mode it is chock full of ...


2

The 2.8 branch is too old. The loop filter was added on 2016-02-18. The 2.8 branch was cut from the master branch on 2015-09-05. Release branches do not get new features backported (only occasional bug fixes). Don't use such an old FFmpeg. Compile or download a recent build. You can see a list of available filters for your build with ffmpeg -filters.


2

Since you're asking for "uncompressed" I assume you don't care about file size. If so, I suggest rendering to mjpeg. This smoothing issue is because your command is defaulting to h.264. The difference between h.264 and mjpeg is that mjpeg is intra-frame compression only, but h.264 is intra- and inter-frame compression. As such, mjpeg is a massive file by ...


2

There are a number of comparison algorithms available beginning with PSNR through SSIM/SSIMPLUS to VMAF. VMAF is implemented as an ffmpeg filter.


2

I don't think that you really thought this through thoroughly. (what a sentence!) On a serious note tho, it would mean that you always have to carry around batteries on you, enough for at least while you're awake, as well as storage devices which you'll want to backup, or at least have in a raid-system. All in all, it's firstly going to be very expensive. ...


1

What you are looking for is called “no-reference quality assessment”. Google can give you specific examples.


1

You need to define what is acceptable to you. Record a well illuminated scene and increase the ISO while decreasing the aperture and compare what noise level on your specific camera is acceptable for you, and you will know what are the limitations of your equipment. In the case of the example video, the noise is "the same" because both moments, the dark ...


1

Your program probably re-encoded the resulting video file from scratch and used whatever setting it considered okay. If you want to have more control about the process and/or do not want it to re-encode the video at all (if both are of the same format), then I strongly suggest looking into ffmpeg for this. For reference, https://trac.ffmpeg.org/wiki/...


1

Yes, there is a method using AI to upscale images: if you convert your footage into an image sequence using ffmpeg, you could run every frame through such upscaling.


1

Ad 1: Try adding the -pix_fmt yuv420p10le option (e.g. before -vf). Or only -pix_fmt + for using the same pixel format as the input video has. If it doesn't help, you obviously have an inappropriate, albeit more common FFmpeg version — only for 8-bit colors. (Your original video uses 10-bit colors, and FFmpeg will try use the best 8-bit pixel format in ...


1

YouTube re-encodes whatever you give it. The process is lossy so there will be a quality reduction. All you can do is give it a very high quality: ffmpeg -framerate 10 -pattern_type glob -i 'resized/*.JPG' -vf scale=1280:-2 -c:v libx264 -crf 17 out.mp4 Use -crf 17 or -crf 18 for higher quality. Try without -pix_fmt yuv420p. I'm not sure what YouTube uses ...


1

You can use Premiere Pro or similar video editing software and plugin Flicker Free from Digital Anarchy with one of the preset. Result of plugin work Settings of the plugin in my case Also I know some other methods but they are poorer and harder.


1

That violates the rules of information theory. Look into the work of Claude Shannon if you wish to understand why what you are asking is mathematically impossible.


1

People need to define the priorities, if the video footage is important to assign a good place to the camera. A television studio with a live audience is a good example of the priority. Of course, this is not the case, but I am pretty sure these days is not the opposite to that. It has some importance to be streamed, "Youtoubed" or something. So, talk to ...


1

In Davinci Resolve, I would solve this problem by creating a multicam clip using Timecode as the method for synchronization and keeping clips organized by camera. This will create a multicam clip with only a single camera, but each clip spaced out according to its timecode. If you like, the multicam clip can then be decomposed in place and converted into a ...


1

Adaptive streaming is accomplished by switching between pre-encoded streams of different bitrates and/or resolutions. But the bitrate within any given pre-encoded stream does not vary too much. YT-DL, with your command, will select the 'best' video stream, usually the one with the highest bitrate and resolution, and return its URL. This URL is a direct ...


1

This looks to be a color space mismatch. You could be viewing a “clamped” (16-235) signal in VLC. This is a common issue with nVidia graphics cards. See VSG:Video:Color washed out.


1

I'd suspect anything purely automated is going to be pretty marginal. There's a lot of complexity to getting a good curve with the right gamma, but there's a variety of possible levels of what would be considered "good results". Personally, I'd rather try getting a good mapping manually with something like Resolve and then trying to see what scenes I could ...


1

As it seems both videos have a bitrate of 118 (117) megabits per second. This means - regardless off any other parameter - your video will take up 118 mb for each recorded second. But as you already noted, a video with double the FPS should have double the data, so if you originaly exported the 4k60 video with 118, you can export the 4k30 video with 57mb/s ...


1

If I was doing it I would probably give both speakers a wireless lavalier (lapel) mic, and record each to a separate channel. I would have two cameras (at least), one generally staying wide (which can be a static camera if you don't have more than one operator), and one getting close shots, and following on the action or dialogue, so when one of the speakers ...


1

Yes, the tech is advanced enough. That is not going to be the problem. You can have a very small, low power camera - look at the ones drone racers use. They aren't high resolution, but that could work on your favour as well, as they won't use as much storage. They also cope reasonably well with differing light conditions. You can power from a belt-mounted ...


1

YouTube will transcode your files regardless of what you upload. If you upload a 1080 file, it will create its own file using their new codec (which I am not sure is 100% deployed yet but is already widely in use), but its their propritary AV1 codec. That said, Lukas above is correct, just upload a 4k if the project was mastered in 4k, and let youtube make ...


1

If you can afford the additional upload and processing time to You Tube, you should use 4k. Not only will this look better for most people (Not actually look better, but rather it looks more modern to an average user to have 4k and some people don't even watch it if it says FHD). But also, YouTube gives 4k videos a higher bitrate which also FHD videos profit ...


1

Well, you are asking for a lot of things here. I'd guess the best option would be a Canon M50, it is not a DSLR camera, but the industry tends away from DSLRs anyway. It can shoot good 4k footage, is new and also shoots good images. You can get it with a double lens kit for around 900$. You can also adapt it to Canon EF. She has Connectivity to your phone ...


1

Intra frames (I-frames, keyframes) are independent of other frames - they contain the complete information for reconstructing the frame's image. There is no need for reading other frames and computing from them. So it's about a speed, not about a quality. On the other hand, more intra frames, grater size of the file. (Known tradeoff speed for space.) Intra ...


1

Neat Video has been working miracles on footage for us so far, you can give the trial a go! We've never tried it on this specific artifact (that I know of) but it does have features for restoration/archival type damage, and I highly recommend it.


1

In my experience (using handbrake) increasing the bit depth increases the file size but not by much. However, lowering the CRF is what causes significantly larger files. I encoded several times the same source with different settings to see what I'd get, so for a random source of 1.45Gb (H.264/AVC) it gave theses results: H.265 8bit CRF 22 (slow) > 362.6 Mb ...


1

A lower-than-optimal bitrate video with higher resolution could easily look worst than a smaller one with proper bit per pixel dedication. Of course a very important fact is the viewing device and condition. A smaller video should get upscaled to fill the screen, for example a 1280×720 will get upscaled in full-screen mode on a 1920×1080 screen. So, in ...


Only top voted, non community-wiki answers of a minimum length are eligible