New H.265 twice as efficient as H.264!

[Penton-Man]

Originally Posted by Penton-Man

listed 23 enterprises owning essential HEVC patents...

http://www.mpegla.com/main/programs/.../hevc-att1.pdf

[Penton-Man]

Something from SMPTE 2014 to make vargo feel happy (and proud)…. http://www.tvtechnology.com/news/008...erified/272970

[hajiketobu]

I have a question regarding HFR. If you encode a clip with let's say 120fps, does it need more mbps because of more information per second? My logic says yes, but when I made a few tests with x264 with high profile, level 5.2 and cq16 the 120fps clip was only a little larger then the 24fps version. Is this normal? Has this do with the way that constant quality encoding works? I thought the clip would end up 5 times bigger.

In my head, theoretically, higher resolution and higher framerates need more mbps. So if an 1080/24 BD has a 50mbps cap, than a 2160/120p would need optimally 6 times as much mbps, theoretically. Lol, I am no expert, I am just speculating. And if a codec like HEVC is twice as efficient it would only need half the bitrate, right? Resulting in only half the filesize.
Can someone give me a professional evalution on this and debunk me if i got this totally wrong?

I made this table below to illustrate what I mean: Would it be right for optimal results? I know filesize and bitrate depend on the compression, but as a benchmark it would mean that with HEVC a 4K 60fps movie would fit on one BDXL/4K Blu-ray and a 4K 120fps movie or a very long 4K 60fps movie on two discs, like they did it with the DVD version and 3D version of Titanic. Is this wrong? Does 4K HFR not need such a high bitrate? Or will they just use higher compression?
In retroperspective, I think I messed up. I think the GB values I have given below must all apply to a 50mbps cap, because 24fps with a max bitrate of 625mbps sure would be bigger than 25GB. Or? Idk

90 min in H.264
1080/24p max.50mbps 25GB
1080/60p max.125mbps 62,5GB
1080/120p max.250mbps 125GB
2160/24p max.125mbps 62,5GB
2160/60p max.312,5mbps 156,25GB
2160/120p max.625mbps 312,5GB

90 min in HEVC
1080/24p max.25mbps 12,5GB
1080/60p max.62,5mbps 31,25GB
1080/120p max.125mbps 62,5GB
2160/24p max.62,5mbps 31,25GB
2160/60p max.156,25mbps 78,125GB
2160/120p max.312,5mbps 156,25GB

[vargo]

With higher framerate, the extra frames are mostly predicted frames, much smaller than i frames.

Each frame has less changes compared to a slower framerate, in theory the predicted frames can be coded more efficiently (albeit there is still more of them).

With a faster shutter speed you will have less motion blur. On the one hand this lends itself well to predicted frame compression because they probably bear more similarity to the reference frame. On the other hand all major H.264 encoders raise the quantizer in frames with high motion, based on the assumption that human eyes don't really notice or care about the quality of blurred objects (so much). So hard to say if this is a net gain or loss.

Overall you will require a higher bitrate, but it doesn't have a linear relationship with framerate. It's a pretty complicated subject so I wouldnt like to put a ballpark figure on it.

[Penton-Man]

Originally Posted by vargo

Overall you will require a higher bitrate, but it doesn't have a linear relationship with framerate. It's a pretty complicated subject so I wouldnt like to put a ballpark figure on it.

Indeed, but let’s kindly offer him an idea of what we do know so far….or at least what the evidence suggests.

We know that with HEVC (HM10), preliminary professional testing (in research labs the likes of BBC R&D, etc.) has shown that despite the fact that we are feeding the encoder twice as much data (120fps vs. 60fps), the bitrate overhead due to the increased temporal resolution is fairly limited and since "fear" seems to be the buzzword today in the Tech Forum, hajiketobu need not fear.

In fact, using HEVC temporal scalability, early testing has shown an average bitrate overhead of 8% more needed in order to transmit both 60Hz and 120Hz signals (within a single bitstream) with up to a maximum of 21% required for the most complex sequence, that of a merry-go-round…going round and round. Pre-processing is expected to get these numbers down even lower.

[vargo]

Those tests were conducted with short sequences whereas my reply is geared more towards a hypothetical motion picture.

Short sequences can be a little misleading, for sure they serve a purpose but what you see in them doesn't always scale out to a full length feature. I expect the min/max numbers to be higher in the real world.

Even judging the footage subjectively is quite difficult, because there's more than one variable.

[singhcr]

Originally Posted by Penton-Man

I didn’t get to watch the Scarsdale shootout yesterday, but a friend who did, mentioned to me that he stated during his presentation that Spider-Man (2) was the first feature film finished entirely in 4K and specifically mentioned that workflow included all special effects in 4K.

That is just plain wrong…and rudimentarily wrong to anyone familiar with post production from that time or simply now.
b.t.w., I don’t mean to exclusively highlight Joe, for I don’t - https://forum.blu-ray.com/showthread...ve#post9553148

Point is, aside from people involved in the entertainment business, generally it would be wise for folks not to blindly follow personalities based on professional credentials or past reputation believing therefore that whatever spouts out of their mouths is guaranteed to be accurate, just because they talk a good talk. More importantly, this advice extends to other more critical (than choosing TVs, kitchen appliances, etc.) aspects of life, e.g. financial advisors, realtors, medical providers, etc. For instance, I know of someone who went to an ophthalmologist who is nationally, even internationally, recognized as a leader in his field of eye surgery, regularly gets invited to and speaks at medical conferences, etc. who, yet despite all that, gave extremely *cavalier* opinion regarding an option for eye surgery to a patient which then unequivocally resulted in the patient needlessly losing a great deal of visual acuity in one of his eyes due to irreversible photoreceptor cell loss.

Although the patient lost much of his eyesight, he didn’t lose any of his intelligence and no longer goes to that particular eye doctor for follow-up care. He is already out thousands of dollars, the pain and discomfort of several surgeries attempting to correct the initial stupidly proposed operation and….the loss of photoreceptor cells which he will never get back as they don’t grow back.

I hear ya. When I had problems with my wrist, I was immediately recommended for surgery by multiple doctors at my local sports medicine clinic. I try to avoid such things at all costs so I tried physical therapy, supplements, etc. After a year I was really desperate and went to another doctor at the same clinic and asked for surgery.

"Why would we operate on you if we don't know what's wrong with your wrists?"

When I heard that, I realized just how stupid a lot of doctors are and I was thankful to find one in the minority who thinks.

[Penton-Man]

Originally Posted by vargo

Those tests were conducted with short sequences whereas my reply is geared more towards a hypothetical motion picture.

Short sequences can be a little misleading, for sure they serve a purpose but what you see in them doesn't always scale out to a full length feature. I expect the min/max numbers to be higher in the real world.

Even judging the footage subjectively is quite difficult, because there's more than one variable.

Yes, but the testing has to start somewhere. For example, the determination of how many bits per color per pixel in for theatrical D-Cinema all started out with an experiment in the old Digital Cinema Lab in Hollywood showing observers a total of 36 still horizontal and vertical square wave test patterns of differing luminance and modulation levels. The results of which eventually did carry over to the DCI requirements after proof of concept with StEM motion pictures for confirmation.

I think the range of from 3% (for soccer) to 21% (for merry go round) for the HFR video bitrate overhead is a decent ballpark estimate for hajiketobu to tuck away in his mind rather than his speculative theory of a linear relationship. Nothing wrong with your reply, I was just trying to elaborate more on it as to what the known state of the art is so far in regards to investigating bitrate overhead for HFR.

[Penton-Man]

Originally Posted by singhcr

When I heard that, I realized just how stupid...

Not sure if that was ‘stupidity’. Believe me, the surgeon who operated on the eye (OS) of the person noted above is FAR from stupid. As a rule of thumb it’s always best to take the time and go thru the inconvenience of getting at least a 2nd opinion. At least with Mayo in your neck of the woods , you should be able to ferret out somebody qualified capable of well balanced advice, no matter what your future healthcare needs turn out to be.

[Penton-Man]

Nifty free tool for comparing HEVC vs. H.264 quality….http://www.vanguardvideo.com/vct.php

[img eL]

More HEVC improvements

Version 2: (October 29, 2014) Second approved version of the HEVC/H.265 standard which adds 21 range extensions profiles, two scalable extensions profiles, and one multi-view profile.

[Penton-Man]

HEVC STB shipments poised to “explode” ….
http://www.businesswire.com/news/hom...I#.VHKeyovF-Sp

[Wendell R. Breland]

Any posters here with a UHD/4K computer setup? If true, would you mind posting a short list of the components, as in:

MB
CPU
GPU
Memory
Monitor
O/S

If already posted, can you provide a link?

[Penton-Man]

Retrospective milestone Look Back at HEVC for those who don’t have the time to read the last 15 pages or outside presentations, like so…

Originally Posted by Penton-Man

Scientific ‘shootout’ performed by the EPFL (https://forum.blu-ray.com/showthread...fl#post8587823) between three contenders….HEVC (HM 15.0), VP9 and AVC

encoding ‘4K’ material (well, to be exact, see slide #5 in the link at the end of this post).
http://soundbible.com/56-Boxing-Bell-Start-Round.html Annnnnnnnd the winner is....

HEVC (skip to slide #18)
For dem slides, see - http://www.slideshare.net/touradj_eb...014-hev-cvsvp9

HEVC (aka H.265 and, for short, MPEG H Part 2) stands for High Efficiency Video Encoding

Version 1 (Jan. 2013) Main and Main 10 (8-10bits, 4:2:0)
Version 2 (July 2014): Range Extensions (10bit or greater, 4:x:x) plus scalabilities
Still in progress: MPEG XYZ, HDR working groups

[Penton-Man]

Cutting edge news -
Handbrake 0.10 update released with HEVC encoding capability (using x.265 v 1.4)…. https://handbrake.fr/news.php?article=27

[Richard Paul]

Originally Posted by Penton-Man

Version 1 (Jan. 2013) Main and Main 10 (8-10bits, 4:2:0)
Version 2 (July 2014): Range Extensions (10bit or greater, 4:x:x) plus scalabilities
Still in progress: MPEG XYZ, HDR working groups

MPEG is looking into the idea of adding profiles for HDR and WCG but unless they can get a large improvement in compression it probably won't result in anything. The number of consumer video profiles is getting a bit high and I don't think they will want to add to that after they finish with the 3D profiles and the screen content profiles.

[faheem5]

Questions about all the various software starting to introduce x265 or H265 encoding.

I make backups of my blurays through few software ( lossless/pass-through in mkv format)

so like many of you i am sure.. i have hundreds of mkv titles backed up each 20-40gb each.. so i was excited to hear about the future of H265 (x265 its open source cousin)

but i am at a lost at which software or what are the steps to convert my VC1 or mpeg-4 titles to H265/x265 within the mkv container?

every software i touch talks about 1-2 pass.. not seeing any which offer the lossless/pass-through option which i am used to having with mpeg4 currently?

i was reading that x265 did release or has an cmd line switch for lossless.. so wondering if anyone can provide a guide or walkthrough on how to convert all my current titles to H265/x265 ?? (again looking for a pass-through option, but wanting to take advantage of the x265 compression.. if that makes sense)

[IndyMLVC]

Many apologies if this has been brought up here already but is it possible that cable systems might hop right over AVC and go to x265? I looked at some of the captures from the Grammy Awards and they're...not good.
Now, granted, this segment had a lot of flashing lights and fast movement but it ended up looking pretty awful in motion, nevermind in screenshots.

[dvdmike]

That's just American tv, it's always looked that crappy.
Hevc can turn too Lego also, it's all about compression.

Resolution is nothing without compression

[AK65]

Mayweather/Pacquiao in 4K please