Hacker News Reader

Comment by dahfizz | Hacker News Reader

Here is a hint:

> Because H.266/VVC was developed with ultra-high-resolution video content in mind, the new standard is particularly beneficial when streaming 4K or 8K videos on a flat screen TV.

Compressing video is very different from gzipping a file. It's more about human perception than algorithms, really. The question is "what data can we delete without people noticing?", and it makes sense that answer is different for an 8k video than a 480p video.

View on HN

So compressing a 1080p video with H266 will not result in similar file size/quality improvements as a 4k video? How much are we looking at for 1080p, 10%?

not an expert, but from what i understand is more like that they "extend" the codec with technique more effective on higher resolution content, or new "profile" (parameter) more effective for higher resolution content (a bit like you can have different parameter when you zip file).

This new techniques can also be used for 1080p video (for example) but with lower gain. Also the "old" algorithm/system are generally still used but they may be improved/extend.

yup, from the example I remember(I read it through link on HN but cannot find it in quick search, I wish I could link it),

if you film(still film no movement) the macbook pro top to bottom in h264/MP4 at 1024 p resolution and again you take a picture from your camera.

the results will be shocking,

the video of 5-10 seconds will have lower storage size than the size of single Image. but when you inspect video carefully you will see the tiny details are missing, like the edges are not as sharp for metal body, the gloss of metal is bit different, the tiny speaker holes on top of keyboard are clear in Images and can be individually examined while in video they are fuzzy pixelated and so on.

so, The end results: A 5 second video with 10s of frames per second is smaller in size than single Image taken from same camera.

Let's also be clear, the still image will be the full resolution of the sensor. The video taken on the same camera is usually a cropped section of the sensor. You're also comparing a spatial compression (still image) vs a temporal compression (video), and at what compression levels are each image taken?

Additionally the images are not very compressed (edit: you mentioned that in your comment, sorry). While the RAW files can be a couple dozen megabytes and the losslessly compressed PNGs are still 5-15 mb, good cameras normally set the JPEG quality factor to a high amount and so even with a JPEG you're getting pretty close to a lossless image. Whereas in video you can often plainly notice the compression artifacts & softness. A more fair comparison would be the video file to a JPEG with equivalent visual quality.

see more details, https://sidbala.com/h-264-is-magic/

I know that's not fair comparison, but imaging clever compressions were not invented and you would download terabytes of data to view a small movie.

What question are you trying to answer? Nobody asked what is compression and why do we use it. I have been encoding videos since VideoCDs were a thing, so I have a pretty good understanding of how compression works. The fact that I differentiated between spatial and temporal compression should have been a clue. All I was pointing out was that compressing a postage stamp sized video and comparing its filesize to a large megapixel image isn't a fair comparison. (yes, I'm jaded by calling 1080p frame size a postage stamp. I work in 4K and 8K resolutions all day.)

You're thinking of "h264 is magic"

https://sidbala.com/h-264-is-magic/

GREAT article

yes,that is the article I was referring.

>How much are we looking at for 1080p, 10%?

We don't know yet. There are no public technical details (that I know of) for H266 yet, but if I recall H265 gave the same 50% reduction in bandwidth claims, and for years people stuck with H264, because it was higher quality due to dropping off less subtle parts of the video you really want to see. Only in the last couple of years has H265 really started to become embraced and used by piracy groups. Frankly, I don't know what changed. I wouldn't be surprised if there was some sort of H265 feature addition that improved the codec.

H.265 was always better from a technical perspective but that's not everything which factors into a video codec decision. H.264 was supported everywhere, including hardware support on most platforms. You could generate one file and have it work with great experience everywhere, whereas switching to anything else likely required adding additional tools to your workflow and trying to balance bandwidth savings against both client compatibility and performance — if you guess wrong on the latter case, users notice their fans coming on / battery draining in the best case and frames dropping in the worst cases.

Encoder maturity is also a big factor: when H.265 first came out, people were comparing the very mature tools like x264 to the first software encoders which might have had bugs which could affect quality and were definitely less polished. It especially takes time for people to polish those tools and develop good settings for various content types.

> H.264 was supported everywhere

this. our TV (a few years old, "smart") can play videos from a network drive, but doesn't support H.265. reencoding a season's worth of episodes to H.264 takes a while...

H.264 will become the "mp3" of video I think. Universally supported, and the pstents will run out micj sooner than the newer formats.

I think you’re right - for a lot of people it was the first to hit the “good enough” threshold: going from MPEG 1/2 to the various Windows Media / Real / QuickTime codecs you saw very noticeable improvements in playback quality with each new release, especially in things like high motion scenes or with sharp borders.

That didn’t stop, of course, but I generally don’t notice the improvements if I’m not looking for them. Someone with a 4K or better home theater will 100% benefit from newer codecs’ many improvements on all those extra pixels but if you’re the other 95% of people watching on a phone or tablet, lower-end TV with the underpowered SoC the manufacturer could get for $15, etc. you probably won’t notice much difference and convenience will win out for years longer.

Reencoding will cumulate H264 artifacts with H265 artifacts (and psychovisual optimization for one with psychovisual optimization for the other). Unless you reencode from source don't do that.

i appreciate your concern :) but honestly, is that a practical problem, or a theoretical one? the end result was fine to watch in our case. (and upload itself wasn't of great quality anyway)

It can be a practical problem depending on the source: if you're moving from a relatively higher-resolution / less-compressed video it won't be noticeable but if you're starting from video which has already been compressed fairly aggressively it can be fairly noticeable.

One area where this can be important to remember is when comparing codecs: a fair number of people will make the mistake where they'll take a relatively heavily compressed video, recompress it with something else, and get a size reduction which is a lot more dramatic than what you'd get if you compared both codecs starting from a source video which has most of the original information.

Another thing to consider, exactly the same thing happened when h.264 was first released.

Even though x264 quickly started seeing better results compared to DivX and XVid, you didn't see pirate encodes switch to x264 for years.

Yes and x264 is such an improvement that it was worth it. HW support became nearly universal. DivX and Xvid didn't have that.

So it was kind of a magical codec upgrade and now it seems more incremental to me.

The scene switched to x264 almost immediately. It did not take years.

Having GPU or hardware support to speed up encoding can make a big difference in adoption.

It was def hardware support that changes the piracy groups policies.

H.265 is patent laden. H.264 is much better in avoiding getting sued.

That's patently wrong[1]

Money quote :

"H.264 is a newer video codec. The standard first came out in 2003, but continues to evolve. An automatically generated patent expiration list is available at H.264 Patent List based on the MPEG-LA patent list. The last expiration is US 7826532 on 29 nov 2027 ( note that 7835443 is divisional, but the automated program missed that). US 7826532 was first filed in 05 sep 2003 and has an impressive 1546 day extension. It will be a while before H.264 is patent free."

(emphasis mine)

[1] https://www.osnews.com/story/24954/us-patent-expiration-for-...

Reply on news.ycombinator.com