Who has the best technical standard worldwide?

No broadcaster in North America broadcasts in 1080p with the current ATSC standard, or the ATSC-MH (Mobile/handheld TV's) standard. Dish Network & DirecTV for satellite in the USA & many cable systems upconvert the HD of broadcast networks to 1080p, even if US networks like ABC, Fox, Ion TV, My Network TV, & some PBS stations broadcast in 720p, & CBS, NBC, CW, Telemundo, Univision, Telefutura, & most PBS stations broadcast in 1080i. For PBS, they're typically 1080i, but some stations don't have the latest codecs, & use 720p instead (or a PBS station might be running 2 HD channels, like my local PBS station, WYIN Gary, IN, which runs 2 720p HD channels along with 2 widescreen SD channels). The North American ATSC standard still uses MPEG-2.

Yes, the BBC. But only 1440x1080p25 (switching between that and 1080/50i for each GOP) - even 1080/50i is better than that, and so is 1080p50/60.

NHK? Super-Hi Vision (7.6K) at 120 fps. But I could think of an even better standard than that.

not sure I get it ... They all use the same standards for broadcasting, best picture is only a matter of bandwidth

They use the same basic technical interconnect standards in production, but not always to the same quality, and the transmision standards vary country by country.

Some countries are using mainly H264 (aka MPEG4) for HD broadcasting to the home, others are still using MPEG2. H264 delivers is a much more modern system, and capable of delivering decent quality at lower bitrates, with far less objectionable compression artefacts. (In Europe there are no major MPEG2 HD broadcasts to the home anywhere - HD = H264 here)

Some countries have a distribution infrastructure based around compressed distribution with multiple MPEG2/MPEG4 encode/decodes between the viewer and the source material, whilst others may only have a single encoder at the end of a centralised broadcast chain.

All of these make a major difference in picture quality terms.

Sky in the UK, for instance, will have just a single H264 encoder between their HD-SDI playout area and the viewer at home. Other broadcasters may use MPEG2/MPEG4 to distribute to a local station, then that station will decode this to HD-SDI and re-encode to MPEG2 for OTA broadcast, and there could be further decode/recodes if the signal is then broadcast on satellite as well. This concatenation of codecs hammers picture quality.

Of course there are differences in standards upstream of the distribution and transmission systems.

Some broadcasters may still accept transmission masters on HD Cam tape (1440x1080 3:1:1) whilst others may mandate HD Cam SR (1920x1080 4:2:2 or 1280x720 4:2:2). Some may mandate a minimum acquisition bitrate of 50Mbs Long GOP or 100Mbs Intra, whilst others may accept 35Mbs Long GOP (aka Sony XD Cam EX) or even 25Mbs Long GOP (aka HDV) for acquisition. Some may accept Super 16mm as being suitable for HD broadcast, others may not. Some may standardise on DVC Pro HD for shooting (1440x1080 4:2:2 at 50Hz, 1280x1080 4:2:2 at 60Hz), some may not think this is acceptable.

Some broadcasters may use 1440x1080 for their transmission resolution (UK OTA HD and satellite, France OTA HD, Japan OTA HD I believe) whilst others may mandate 1920x1080 (Sky HD, and US OTA broadcasters)

Whilst the 1920x1080 4:2:2 and 1280x720 4:2:2 interconnects are standards, and there are many broadcast recording, distribution and transmission standards, there are lots of combinations of these, and lots of different bitrate choices that can alter quality to the viewer at home.

And then of course there are the technical standards involved in ensuring decent production values (camera set up and matching, decent camera work, decent sound acquisition etc.) You can shoot with the same kit and have the same distribution model - but deliver totally different qualty results.

Personally I think Sky have very decent technical standards for delivery to the viewer, and a very clean chain. However outside of sport and news they don't actually make much content of their own, so it's difficult to judge their production technical standars. (I sometimes think they over detail-enhance some of their live sports to make it "crisper")

When it comes to great HD production I've seen - I think some of the European HD state broadcasters do a pretty good job. SVT HD (1280x720/50p H264) on Canal Digital usually looks cracking, with no obvious compression artefacts visible, and some of their HD music shows have incredible standards, with very well mixed 5.1 sound at 640kbps and very clean pictures. But then their SD broadcasts on the same platform also usually look excellent, and they've been doing 5.1 sound for SD 16:9 broadcasts for years (it was part of their SD OTA digital standard)

I'd also add that programme source and style plays a major role in perceived picture quality.

The Japanese broadcasters, for instance, use huge amount of lighting in their studios, which means they're running with the cameras stopped down, and a large depth of field meaning everything in the frame is in focus pretty much. They also run with a very video look on much of their output with highly saturated colours, and this can give everything a crisp, pop out of the screen, look. The same techniques are used on the infamous HD demo loops of brightly sunlit, saturated exteriors, with 60Hz interlaced motion. They have that "looking through a window look"

In contrast, many European and US broadcasters, will go for a totally different style (particularly for drama), with low lighting levels, shallow depth of field (throwing backgrounds out of focus), to give a more cinematic feel to their HD. They may well heavily grade their material to make it look "filmic" de-saturating it, colour-biasing it, shooting with prime lenses (or trying to simulate the look of primes) etc. They will often shoot 24/25p progressive, with the much lower temporal resolution (i.e. film look motion).

Put the two programmes next to each other - and people will undoubtedly say the Japanese stuff looks sharper, crisper, more HD-like etc. (It's why HD demo reels don't often have heavily graded, filmic drama clips in them!)

What I was trying to say is that broadcasters are committed (more or less, obviously even in broadcast there are some lousy stuff) to maintain signal quality along the acquisition, production and contribution links. They are aware of degradation in image quality after cascaded codes connections, that's why they're using 4:2:2, large bandwidth feeds, and contribution encoders.
No one is trying to compare apples with oranges - meaning there is no takin'bout H.264 vs Mpeg-2 or something.
It's the final compression (i.e. in distribution) which put the greatest stress on image quality, and that's why I said it that
best picture is only a matter of bandwidth.

There, in distribution, encoders must heavily compress the signal, with minimum delay and, in statmux, it's the multiplexor who allocate the bandwidth. Ok, there might be someone coming and say that some distribution encoder suppose to have some mambo jumbo capabilities but, from what I've seen, even the best (arguably) contribution encoder can't go below 10 Mbps (in H.264/4:2:0/8 bit/D) and still show some reasonable images. Not mentioning that you hit water (sort of speakin') it's bye bye

Moreover, broadcasters knows that their distribution signal shall be used by network distributors (cable, DBS, IPTV), sometimes re-encoded, trans-coded or whatever, and that's why (usually) use decent bandwidth and CBR, and that's why image quality, between a broadcaster and an distributor, (than again, usually) is better with first. Heck, not long time ago, Dish had some 6-7 HD channels in one TP using DVB-S (around some 40 Mbps for all). On the other hand, I think that I've seen one broadcaster with some 18 Mbps HD channel and boy, especially in live events, that was killer

obviously, all above was my 2 cent opinion

sometimes I think andy might be putting up with too much ...

No Wowow is not 1080p. It is 1080i/60.

It's not the standard, it's how it's implemented.

I agree and think sneals2000 did a very good summary.

Interesting - when did it switch from 60Hz to 120Hz? The demos I saw at IBC and the BBC were still 60fps.


I know that people were suggesting that motion blur at even very low levels was an issue with the resolution they were running at with a 60Hz frame rate, and questioning whether 60Hz was applicable, but equally sensitivity will be an issue as they increase the frame rate (and drop the exposure time). The system I saw demo-ed at 60Hz had significant blurring on even very slow camera pans - and this was using the 600Mbs MPEG2 circuits (live fibre from London to Amsterdam using IP technology)

I think they're using larger sensors to cope with the requirement to have 16x as many sample sites as 1080p HD as it is!

It was over IP, but only if we look up to level 3 OSI. In order to work over long distances (such as London to Amsterdam) it's mapped into TDM (SDH tech - for Europe, SONET for USA, Japan - both virtually the same) and, after that, into DWDM. That's the 'standard' way to do it. Just to mention, protection schemes use in TDM are far more advanced and reliable than what IP is offering
In the beginnings, IP wasn't supposed to 'act' as a transport layer and that's why there are all sort of difficulties when is used such as. There's all kinds of 'improvements' and 'gadgets' making vendors to recommend it as a cheap alternative but, eventually, you'll end up just like with SLRs, thinking you'll get cheaper for shooting video.

I forgot to mention that Wowow uses MPEG2 at 20Mbps CBR. At least that what it was last time I checked.

IMHO The best HD PQ I saw in the world is YES Network which is local sports channel for baseball NY Yankees. Their PQ is always top notch, no matter if they broadcast home games or away.

I don't think they've switched yet, but I think that is about what they are recommending (or somebody from NRK was I think) as the frame rate (or one of them) for the standard. ie. after the SHV tests/research (where they had .PDF documents mostly about the spatial resolution and what could be resolved etc.) they said they were going to look at more tests/research into what frame rate(s) would be best.

Most sites/papers say 60 fps, a few others say 120 fps (or 120 as an option).

eg.
http://www.nhk.or.jp/strl/open2011/tenji/04_e.html


Also there's a BBC paper called "Challenges and solutions in broadcast archives" with a section that says "New Formats" and "Super Hi-Vision" (SHV), "50, 60 or 120 fps".

So they're not currently using it. Maybe it's more "at some time in the future". It might just end up being just 59.94 fps.

@joebloggs - yes. Hadn't seen it anywhere stated as fact other than wikipedia, which always worries me.

The IBC at which I saw SHV for the first time was also where the BBC were discussing whether 50fps progressive was enough (let alone 25fps) for future TV systems, with demos of 100fps and higher frame rate acquisition (though display maxed out at 100fps the material was acquired at 300fps I think)
(Obviously from a European perspective, though similar arguments hold for 24/30 and 60fps when it comes to motion blur etc.)

I wonder what is written in SMPTE 2036 ...

Is that the 2009 document (or 2008/2010)? If you have around $50 spare you could find out

Though they're still working on UHDTV standards this year.

eg.
http://www.linkedin.com/groups/What-...nition-3814729

Posted 3 months ago:

Posted 1 month ago:

120Hz will be interesting in 50Hz land - but 300Hz was never really going to fly, and wouldn't have been great for 24Hz sources.

600Hz anyone...

100 frames/fields per second is the minimum frame rate for sports television. Of course, more is better, if you can take the hit in sensitivity.

I still laugh when people say that progressive 50p or 60p is "better for sports" than the comparable interlaced frame rate. A motion-blurred image is still blurry regardless of whether it is 1280x720 or 1920x540.

Yes, I think they should use 600 Hz

But even that would be a problem because they use 59.94Hz and 50.0 Hz so they'd need to do speed-up of 59.94 Hz content (or pull-down for live content) if the TV was 600.0 Hz. And if the TV was 599.4 Hz, 50Hz content would need pull-down.

And we still wouldn't be able to watch The Hobbit films (around 48 fps) properly (without pull-down conversions) .

Yes 120Hz will be a problem for 50Hz shooting and content. They should allow more than one format in the standard.

I suspect that by this point they'll find a way of dealing with 59.94 vs 60. Speed-up for upconversion of non-live 59.94->60Hz - though live is trickier. (Sony's 60Hz HiVision ->59.94Hz SD converter tried to frame drop on cuts didn't it?)

As for 48Hz content - I suspect speed-up would be a solution if frame interpolation hasn't improved enough!

Yep - suspect there would be a lot of pressure for this. Be nice if they could find a way of handling multiple frame rate transmission neatly.

Well, it's true that in theory neither of 50 or 60 Hz aren't high enough to provide good motion image ... heck, one expert in video said that if 720p format didn't retain 4:2:2 sampling and it used 4:2:0 sampling, the beneficial option of a 75 Hz frame rate could have been added with no increase in bit rate.
Due to the poor temporal response, the human eye translate (at some point) those frequencies in motion blur, when eyes are 'fixed' viewing an moving object. But if the eyes are following the object (which they do), than temp. frequencies become zero and details appear. That's the trick that when the camera is fast panning following the action, let's say football, you don't notice that grass looks like sh*t.

Takin' into account current motion compensation algorithms (and deinterlacing for interlaced signals) inside tv's, do you think if that they shoot at 100 fps (talkin'bout 1920x1080 and 1280x720 signals resolution), you might be able to see some benefits?

What I think you're talking about is watching the TV in real time. If so, 100 Hz would only make a difference on really big screens (like the ones being suggested for UHDTV) so that when your eye tracks an object on the screen it doesn't appear to jump.

What I'm talking about is when the action is slowed down for a replay. The magic number for "super slo mo" was 3x the normal frame rate. At that point you had good replay motion portrayal and a good balance with sensitivity. Higher frame rates would "stop" the motion better, but require a more sensitive camera.

ooohh somehow I've fail to grasp it from previous post ... ma' bad

You'd be surprised. The demos I saw at IBC were on small-ish projectors (only DLP technology allowed a 100p signal to be displayed on readily available kit at the time ISTR)

The sequences that really showed the differences were where the camera panned to follow a model train. The train was clearly visible with lots of detail at 25p, 50p and 100p. The big difference was the level of motion blur in the background which was moving much more quickly. At 100p you could clearly read the station sign on the platform as the train was followed through the station. At 50p you could see there was a sign but reading the text was tricky. At 25p you could see there was a station, but probably didn't really see the sign enough to recognise it as a sign! I suspect this would have been visible on even smaller displays - as it's detail that is simply not captured clearly at low frame rates, but is at higher frame rates.

There's a link to the white paper here : http://www.bbc.co.uk/rd/publications...paper169.shtml

(The screenshot of the train shows the opposite effect - where the train is moving through shot rather than being followed by the camera)

Yes, motion blur is a problem whenever the camera moves. The "film" people have known this forever, and have even created a whole class of people (cinematographers and the like) whose job, or at least part of it, is to keep the director from moving the camera in ways that slow frame rates make look bad.

The background blurring is somewhat unavoidable, because the director is using the camera as a proxy for your eye, deciding for you what you should be paying attention to. The camera tracks the train so you don't have to. You don't get a choice, except in the lab. The sign isn't what the director wants you to watch. High frame rates and high resolution let you decide what you want to pay attention to. In crowd scenes you can look for Waldo where you want to, not where the cameraman or director wants you to look. In effect, higher frame rates would take control of your eye away from the director, but that's an argument for a different thread.

imho, talkin' specifically movies, it's not about who gets to decide who's watching what. The camera is telling a story. Point. As simple as that.
Even if you'll get a choice, it still won't be yours, 'cause it's the framing of the shoot who's tricking you, take Citizen Kane for example, where deep focus (double exposure technique) "let you choose". Sure, higher frame rate in shooting movies has benefits, that's why Cameron is preaching towards 48 or 60 fps and Peter Jackson is shooting "The Hobbit" in 48 fps (with a 270 degree shutter angle), saying that it gives "a lovely silky look". But 48 is for 3D ... for now. Will they shoot 2D movies like that? Well, I think time will tell ...