There's nothing clever about the harsh sharpening that you've done, and resulted in unnatural halos. That's just a basic/generic "sharpen" filter cranked up to 11. It's not actually nuanced, and does not take bad side effects into consideration in its algorithm.Clever sharpening
What does "post sharpening of video camera records" mean? That's a nonsense phase.I know it from experience from post sharpening of video camera records
That's humorous. I started in the 90s, when it was still was not feasible at home. I was using SGI MediaBase for MPEG and LAN streaming. Trying to do anything digital was painful in those days, and I had to wait about 5 more years before digital was actually doable outside of a broadcast/studio/IT facility. I had to suffice with analog setups bolstered by digital techniques.long before you started to notice the simple existence of digital videos.
You've entered the site making asinine claims, and attacking respected when your claims are shown to be BS.
+ Reply to Thread
Results 31 to 38 of 38
The last time I tested NVEnc it was v3.27 and sadly H265 encoding was completely unusable.
The image quality in black/darker areas was so poor with blocky graphical glitches everywhere.
I'm happy to report things have much improved with NVEnc v4.09 regarding H265 encoding.
The overall image quality looks a little softer and less sharp compared to a H265 software encode. (I've not played around with VPP filters yet).
But the black/darker area image quality has been fixed and now looks much better.
It's far from perfect, but the results are now what I call 'acceptable'.
Encoded 1h36m58s 1080p/24 in 12 minutes @ 200fps.
i7 6700k @4Ghz / 16Gb DDR4 / GTX1080ti (Driver 398.36)
to those that have tried this and believe the results are decent or better.
how far have you been able to push the compression?
i tried for ~39% but it actually ended up being ~57% of my original files bitrate and it seemed pretty good. now that was a stand up comedy thing so wasn't a great test for fast action for example.
it was 720p 3.5Mb/s AVC.
next ill try to do something 1080p with some action. but i suspect ill have to bump it up to at least ~50 to 60%(attempted size), which i would be okay with.
wouldn't mind hearing other peoples thoughts on those that are making use of this.
thanks, and thanks to Comparison for the original post.
I read the beginning of this old thread, I ask if currently the X265 coding (via software) is still higher than NVENC (via hardware).
I have the temptation to use NVIDIA hardware speed using high-end video cards, but if the quality is less than x265, I can buy a less performing video card and save money.
Thanks for the answers
The general restrictions of GPU video encoding chips, compared to an all-purpose CPU with gigabytes of RAM, did not change. It is easier to speed up x265 with more RAM and a CPU with more cores and more efficient AVX instructions than enhancing a GPU hardware video encoding chip. And if you tried to copy more video frames into the VRAM to support its quality, it may not work in real time anymore.
Thanks for the reply.
Confirm that I should address me on the X265 encoding by focusing on the performance of the multicore CPU, I have in mind Ryzen 9 5900x and RAM 3600 MHz, DDR4 (32GB), and insert a video card that costs less (RTX 2070)?