I saw what you did there ;)

@egozi44

Hello! what do you mean? ^_^’

Sankyuu!!

@SomaHeir
Enjoy it ^_^.

Thanks mate

@ShuKu
You are welcome friend :)

So, idk if you watched the video linked in https://nyaa.si/view/1140063#com-10 and agree with it, but I feel like I have to clear up a huge misconception here.

Yes, using 4:4:4 chroma subsampling allows for more data to be preserved arcoss the chroma planes. No, this does not mean releasing 1080p 4:4:4 is fine because BDs are released with 4:2:0 subsampling and at that point you are purely upscaling chroma and not preserving any information. You just generate more information from thin air basically, only to inflate the video size, which is referred to as bloating.

In order to display video data on an RGB screen, the color needs to be converted anyway, before which chroma gets upsampled to the luma resolution if needed – by the video player. You don’t need to encode the upscaled chroma into the video because users do the same thing at playback but without wasting bandwidth or disk space.

Technically, you performed 2/3 of a 1080p-to-4k upscale with only the luma plane left out.

@FichteFoll: Encoding bluray data from high sampling rate 8bit 4:2:0 (5~10 GB per episode) to lower sampling rate 10bit 4:4:4 (700~850 MB here) is not a bloating encode.
There are a lot of data thrown away and raising color depth and chroma subsampling allows preserving more data through encoding.

If 4:4:4 chroma subsampling is a bloat here, then UQW’s encode has up to 50% bloat data i.e. 350~450 MB per episode!
The Airota&VCB-Studio’s HEVC encode in 4:2:0 with 600~730 MB size per episode has comparable quality as this release.
So in this case,

Airota&VCB-Studio’s HEVC: 600~730 MB per episode

loses to

UQW’s AVC non-bloat 350~425 MB per episode

@Nyann that’s not how compression works, buddy

Encoding bluray data from high sampling rate 8bit 4:2:0 (5~10 GB per episode) to lower sampling rate 10bit 4:4:4 (700~850 MB here) is not a bloating encode.

Sampling rate plays absolutely no factor here because the metric for video is frames per second and I sure hope the framerate hasn’t been changed. You were probably looking for “bitrate”.

Comparing an encode to the Blu-ray source and saying “it can’t be bloated when it’s smaller than the source” is like saying “MP3 with 1000kbps can’t be bloated because it’s smaller than WAV”. If you couldn’t make it smaller, you failed as an encoder. Or you love dynamic grain and bloat, which still makes you fail.

There are a lot of data thrown away and raising color depth and chroma subsampling allows preserving more data through encoding.

You don’t seem to understand. Yes, raising color depth and using 4:4:4 chroma subsampling increases the maximum information conveyable. But when that extra information has been generated with a very simple algorithm, there isn’t any benefit in it, because it might as well be generated by the player as needed.
Actually, raising bit depth is different because ot actually helps with compression for Anime content specifically while making it look better due to how color gradients work, but that does not apply to chroma subsampling.

If 4:4:4 chroma subsampling is a bloat here, then UQW’s encode has up to 50% bloat data i.e. 350~450 MB per episode!
The Airota&VCB-Studio’s HEVC encode in 4:2:0 with 600~730 MB size per episode has comparable quality as this release.

I don’t care if someone else’s encode is larger or smaller than this, because I never have and will do actual comparisons. The technical background already marks chroma upscales as bloaty and your comparison is just pointing fingers to unrelated issues (that may also result in bloat, such as an unnecessarily high CRF or bitrate).

@nyann https://www.animemusicvideos.org/guides/avtech31/theory-videocolorspace.html