Cornucopia is also a an old and experienced member, I miss his denoised version too...
The victory can not be complete without the triumph over Cornucopia and Poisondeathray... Since Selur Cornucopia and Poisondeathray are professional and experienced users on videohelp. I read so many posts of the Videohelp users... but others seem amateurs (Sorry, if I offended/hurt somebody, but this is the honest reality)
THE RACE IS NOT OVER...
Closed Thread
Results 211 to 240 of 359
-
Last edited by Truthler; 4th Oct 2021 at 05:23.
-
User Poisondeathray, and user Cornucopia!
So here is the original noisy camera footage , 9,8 GB lossless utvideo.
http://sendanywhe.re/GDRFLS08Last edited by Truthler; 4th Oct 2021 at 04:14.
-
I'm a bit busy this week, but I'll make a more detailed post later this week or on the weekend
Only the last 1/2 of that video is relevant for testing. You could have cut off the 1st 1/2 and saved everyone bandwidth
Overall that usage of Neat Video by Truthler did a good job with cleaning , revealing details such as the architecture, tiles, bricks , wall text "latogatokozpoint / visitor center", etc... ie. some good results, but some bad results too - it's oversharpened (halos, sharpened noise enhancement), and has stabilization frame edge and ghosting artifacts. It probably would have been better just using it "normally" without stabilization or such strong sharpening.
I'm certain you can get similar using other methods , or better results and without the ugly edge artifacts. I already mentioned the "how" earlier using the moving window stabilization method.
I will demonstrate how "silly" the sharpening is, because it's very unprofessional and looks terrible. Black objects don't glow white. White objects don't glow black. Halos don't occur in real life, unless you have ocular/eye problems. And I will similarly oversharpen a smdegrain version to compare - to "speak in words you can understand"
-
See, @truthler, this is where you are wildly out of your depth. I already said if we wanted to do comparisons (and it's not really supposed to be a "competition", but rather a scientific discourse with testing), we would need all those items from you.
Once again you have been asked for A, B, C, D. You provide only some kind of C and say, "I'm done, I won". Doesn't work like that.
Until you follow the rules, this is BS grandstanding and I'm done.
Scott
-
I already mentioned them. But here's the idea:
1. Source file, with its full provenance and metadata and settings of capture/creation.
2. Output files, direct w/o reprocessing or screencapping or reconversion.
3. Steps to ingest & process & output, EVERY piece of equipment and EVERY app used - including supporting decoding and encoding libraries and architectures, from start to finish, in order, with ALL of their settings each laid out in text format so that one can:
A. Reproduce the same outcome identically.
You just gave us a new source with no detail of steps or settings.
B. See what choices were made, because there might be better/worse/different ones to make, and it also guides one how to setup an alternative equivalent workflow (if using, for example, a different de-noiser).
Previously you gave us Output files, or rather screencaps of output files often with your own highlighting and then later gave us a single set of screencaps of 1 page of settings of solely the denoiser, which may or may not have matched some (but probably not all) of the sources or outputs.
Totally haphazard & piecemeal.
ScottLast edited by Cornucopia; 5th Oct 2021 at 13:09.
-
Do I have one? No.
Do I understand the details, regardless? Yes. I've worked with literally hundreds of cameras, both pro and consumer.
What's your point? The sticking point here is much less on the provenance of the source, and much more on your processing steps.
Scott
-
Hello Scott!
I use Camlink 4K in NV12 4:2:0 for the lossless record of the camera's HDMI out. I used OBS for capture.
In earlier times I used lossless NVENC HEVC with long GOP to achieve higher comprtession on my laptop, now I swithced to 4:2:0 Utvideo in avi container, because all frames are individualy compressed with the Ut video so seeking backward (with frame step) is faster.
I used Avidemux to cut the video without re-encoding.
Than I uploaded the file to the file sharing service, and posted the link on videohelp.
Should I desribe the Sony camera settings too?
Here is a sample file (9GB), can you see the quality of my recording chain? Especially between 1000 to 1300 frame.... It's really fantastic, isn't it?
Link to the sample video: http://sendanywhe.re/P160A3OU
Thank you for your reply!Last edited by Truthler; 6th Oct 2021 at 03:18.
-
The picture is correct. If I want to upload videos for youtube, sharpening is necessary, since the low bitrate will reduce the sharpness and remove many side-effects of sharpening. If I want to recomress the original lossless video into lossless HEVC, the compression will remove a lot of side effects. (Not all side effects, but most of them)
I know there are a lot of people who wear glasses for correct eyesight , but they did noit wear it on the streets for aesthetic reasons. They use glasses only for reading or working before the computer. If they had originally good eyesight in the beginning of their career, it was long gone, enough long time ago to be forgotten in their memory.
I also noticed that 90-95% of the people who deal with electronics, computers (videoediting programming etc..) during their worktime usually had to wear glasses. I can not prove that, but it has a high chance that Cornucopia, Selur or Poisondeathray use thick eye glasses during their work. LITERALLY their eyes are victims of their profession.
For them, it is unnatural if something is sharp like in real life, because they are used to their "dim reality" on the streets, parks and in the nature. For them the dim picture is the normal/natural situation of physical reality...Last edited by Truthler; 6th Oct 2021 at 05:44.
-
I was not rude I just wrote down a well known social phenomenon among tech guys... It was not my intention to hurt anybody with that comment.
There is nothing wrong with people who have to wear eye glasses, they are not better or worse than people who don't need eye glasses.
I did not attack anybody with my comment, just explained a pehomenon which may effected their taste regarding to video quality.Last edited by Truthler; 6th Oct 2021 at 07:31.
-
Of course user S-mp!
Neat photo denoiser for Adobe Photoshop.Last edited by Truthler; 6th Oct 2021 at 06:40.
-
Just to join in on the fun, here's one I did using VDUB2+MSU Denoiser, which I was happy to find out not only has GPU acceleration but it works with Intel's iGPU.
There seems to be a bug in the VDUB2 build I was using because when I tried to export the file with any sort of compression, I would get an out of memory error, even though I have plenty of ram. I ended up exporting a 60+gb uncompressed file and then encoding it with x265 12-bit via Handbrake.
-
Wellcome Sophicticated!
You can use Ut video lossless codec, and using https://send-anywhere.com file sharing service upto 10GB per registered e-mail.
-
I have no interest in uploading to that service, the truth of the matter is that while people may use UT for archiving footage it's not a delivery format and no one is distributing uncompressed or lossless compressed footage so I see no point in uploading a 10gb file.
What's going to happen is people will shoot, hopefully with the proper camera settings, something you show no desire to do, then if there is the need to denoise, adjust brightness, adjust contrast, sharpen or any other adjustment, they will do it and they will create a master for archival. The will then take that master and create a deliverable, be it via a streaming service or broadcast and that deliverable will be in a format such as AVC, HEVC or VP9.
No one will ever see your denoised version, they will see a lossy compressed version, so I don't see the point in uploading a 10gb file.
I also do not see the point of artificially creating noise by using improper camera settings.
Is this the way you normally shoot? If I hired you to shoot a sporting event of a local school or similar such event, are these the settings and processing you would use? If not, how about shooting something with the proper camera settings.
Also, if you want to test Neat Video, go here:
https://www.blackmagicdesign.com/products/blackmagicpocketcinemacamera/workflow
Download the Fitness Blogger video, it's incorrectly labeled as Content Creation 3.21GB, it's the one with the Asian girl working out.
See what Neat Video can do with that or some of the other samples.
And stop shooting like someone that was rejected from community college film school.
-
^^ TU for bringing back up the issue of ISO settings which I raised (and you responded to) back in #105 and #115
And even when he was challenged about it he simply stated (not an exact quote) "How else can I create noise ?"
Of course he will pay no attention to common sense. That was never the purpose here. He sees the program through rose-tinted glasses. Oh, hang on, I must assume, unlike everyone else on here, he does not wear spectacles. (PS I have worn glasses for the better part of 60 years. Does that cloud my judgement ? I trust not)
Similar Threads
-
Restoring a whole Anime movie using Neat Video
By xonathan in forum RestorationReplies: 4Last Post: 26th Apr 2021, 08:55 -
Neat video "colorizes" wrong color hands and face
By mammo1789 in forum RestorationReplies: 4Last Post: 25th Aug 2020, 15:21 -
Any Experiences with "NEAT VIDEO"?
By Avagadro1 in forum RestorationReplies: 3Last Post: 12th Aug 2020, 00:22 -
Create a function in Avisynth to apply this denoiser
By zerowalk in forum EditingReplies: 3Last Post: 19th Aug 2019, 07:45 -
is neat video my best option
By hdfills in forum RestorationReplies: 9Last Post: 25th Aug 2017, 12:57