Plea to PAL land HD providers

[Asmodian]

I think he is just mad that the new equipment going in is 8-bit 4:2:0 MPEG-4, not the ideal 10-bit 4:4:4.

Ah the sad real world where most people still think native 1080p 8-bit 4:2:0 is amazing quality (most of the channels on my cable are 720p). I grew up with VCRs being the norm.. it is amazing trying to watch a tape now; 1080p 8-bit 4:2:0 is great.

Quality is relative, maybe when most of the consumers haven't grown up with analog SD being the norm people will start to be able to notice the improvement that comes from 10-bit 4:4:4 enough for some content to be delivered in it. Right now the cost/benefit ratio is terrible to any business looking at the question.

[paradoxical]

I think he is just mad that the new equipment going in is 8-bit 4:2:0 MPEG-4, not the ideal 10-bit 4:4:4.

And I share his frustration. I roll my eyes every time I talk to someone who tells me that Blu-Ray has no real benefit over DVD (and I still hear that from time to time even now). He just fails to realize that the real-world never lives up to what we would want in an ideal. Compromises are always going to have to be made simply because of cost, bandwidth, etc.

Right now the cost/benefit ratio is terrible to any business looking at the question.

Bingo. There's just too much to have to upgrade for anyone to realistically do it en masse. Same reason why the IPv6 rollout so far has been so mediocre.

[jq963152]

I think he is just mad that the new equipment going in is 8-bit 4:2:0 MPEG-4, not the ideal 10-bit 4:4:4.

Yeah, but it's not only about the 4:4:4 10bit thing. It's also about temporal resolution.

Today 1080 basically is capped at 30p or 60i, which is very unfortunate.

60p would be much better (and could be displayed natively on virtually all HDTV displays today).

In an optimal world
in an ideal world

In an ideal world we would have 15360x8640p @ 1200 fps @ 16 bit @ 4:4:4 @ HDR @ huge color space @ lossless compression and so on... or something like that... :D.

[paradoxical]

Yeah, but it's not only about the 4:4:4 10bit thing. It's also about temporal resolution.

Today 1080 basically is capped at 30p or 60i, which is very unfortunate.

Yes, because 1080p60 requires much more bandwidth than 1080p30 or 1080i60. Which means that you can fit less channels in the available bandwidth or you sacrifice quality to fit the same amount of channels in. We do not have infinite bandwidth. Hence why the compromise of 720p60.

[wonkey_monkey]

Just upconvert all old 24p, 25p, 30p, 50i and 60i footage to 60p

Anyone who really cared about video quality would never suggest such a thing!

[jq963152]

Yes, because 1080p60 requires much more bandwidth than 1080p30 or 1080i60. Which means that you can fit less channels in the available bandwidth or you sacrifice quality to fit the same amount of channels in. We do not have infinite bandwidth. Hence why the compromise of 720p60.

In some various threads here the last days, there were various reports that digital video broadcasting via cable or satellite would basically be limited at around 18 - 20 Mbps (per channel).

I can not imagine that doubling that would be such an insoluble issue...

Why would it be such an issue to increase the bandwith from 20 Mbps to 40 Mbps?

Also, there are things like H.265 which might be able to reduce the necessary bandwith...

PS:

IMHO 20 Mbps are not enough for encoding 1080p30 or 1080i60 in H.264, especially when the encoding has to be done in real-time...

So 1080p60 should rather get more bitrate than 20 Mbps * 2 IMHO ;).

62.5 Mbps (H.264 HP@L4.2) would be better.

Anyone who really cared about video quality would never suggest such a thing.

In this thread we learned that it is not possible to switch framerates on a TV channel because it's a continuous stream.

So it can only be one framerate.

And actually 24p, 30p and 60i basically can be converted to 60p without any issue AFAIK.

Apart from that, it's already happening already anyway.

AFAIK, NTSC TV channels already are broadcasting 24p, 25p, 30p, 50i and 60i footage at 60i (1080) and 60p (720).

So it would be nothing new. The only new thing about the suggestion would be that it would be the full 1920x1080p @ 60fps and that "PAL land" would switch from 50 Hz to 60 Hz.

[nhakobian]

Why would it be such an issue to increase the bandwith from 20 Mbps to 40 Mbps?

Also, there are things like H.265 which might be able to reduce the necessary bandwith...


For a satellite, yes. Assuming the satellite network is running at or near capacity, suddenly needing 2x the bandwidth could mean building and launching all new satellites.

Thats not even considering the increased microwave spectrum that would be needed to broadcast the signals, and potentially re-working the ground uplink stations to handle the increased bandwidth. This could easily cost 10's to 100's of millions of dollars.

Assuming that the satellites are just simple signal relays (which I would hope they are), then it could be very worth while to adopt a new video format (as you suggested) since it's probably a fraction of that cost to replace all the set top boxes in customers' houses.

For cable transmission, this might not be possible. Increasing the bandwidth available in a cable signal would definitely require changing all the transmission hardware. This assumes that all the copper cable is of high enough quality to transmit information at the higher frequencies without significant loss.

This is essentially what held back cable internet services in the late 90's and early 2000's. The transmission hardware had to be upgraded in order to handle the increased bandwidth (and the bi-directional data flow), and this was often done on a neighborhood basis. I don't think my parents could even get any high speed internet until 2004 or 2005 (and they live in a large city). For the companies, this was very worth the money invested since it was a new service they could start charging people for (and get $60/month per household). But if they were told that they needed to upgrade the equipment again, but couldn't charge any higher prices, they would most likely be reluctant to spend the money. (Again, adopting a new video format is a possibility, but there is still a large expense in swapping out transmission side encoders/hardware.)

Now if there was a competitor which offered all these services and people were leaving the big cable companies, then maybe they would think about upgrading their infrastructure.

Eventually the equipment is going to have to be upgraded. New transmission standards will be developed and people will slowly have to swap out their equipment (as in the slow initial adoption of HDTV's), and then we'll see the benefits. Unfortunately by then, there probably will be an even newer technology that beats the pants off of that.

This isn't the only area you see these types of issues. It occurs in literally everything, especially science. It often takes 10+ years before a technology goes from production to perfection and is considered ubiquitous. And in many, many cases, you re-use existing infrastructure for as long as you possibly can, in many cases you make exceptions so equipment can be used for 30+ years because replacing it is just not cost effective.

I think another limiting factor is backwards compatibility. People would get really mad if their cable/satellite company were to suddenly tell them their TV's wont work anymore unless they go and buy new ones. Making sure there is backwards compatibility, at least for a reasonable amount of time, has contributed to many of the strange issues and requirements the current formats have.

I think I've somewhat lost the thread of what I originally intended to say...

It sucks, I agree.

[jq963152]

I think another limiting factor is backwards compatibility. People would get really mad if their cable/satellite company were to suddenly tell them their TV's wont work anymore unless they go and buy new ones. Making sure there is backwards compatibility, at least for a reasonable amount of time, has contributed to many of the strange issues and requirements the current formats have.

For TVs (displays) that would be a non issue AFAIK.

Because, if the channels would broadcast at 1080p60 and your TV would not support 1080p60 input, then you simply could set the STB to output 1080i60 or 720p60 or 480p60 or 480i60.

Simple.

It would be the same for 4:4:4 or 10bit or UHDTV and so on.

If your TV doesn't support that as input, then simply let your STB scale it down for a compatible output.

The STB could make it backward-compatible.

[TheSkiller]

and that "PAL land" would switch from 50 Hz to 60 Hz.And break compatibility (or negatively affect the quality) with 60+ years worth of legacy material, sure. :rolleyes:

[jq963152]

And break compatibility (or negatively affect the quality) with 60+ years worth of legacy material, sure. :rolleyes:

Somehow i knew that this would come up as an argument...

You have a point.

But on the other hand: it would make legacy material worse but at the same time would make all future material better.

It would be a compromise.

The only way to not make a compromise would be to go for... 1200 fps, correct?

1200 fps would be compatible with 24 fps, 25 fps, 30 fps, 50 fps and 60 fps. It would even be compatible with 48 fps ("The Hobbit").

But hey... 1200 fps... even to me that sounds like:




fantasy land


:p

So, a compromise would probably be needed, don't you think ;)?

[filler56789]

An even better option:
invent the Variable Refresh-Rate :devil: devices.

But I know, I keep dreaming :D

[smok3]

Actually what should have happened is the dead of the i part of the equation. I shoot a lot of 25p for example (because I know that post will be fast enough and generally non-problematic, everything is tied to deadlines), which then at the final stage gets animated titles, which are 50i and everything packed into 50i stream, sick.

p.s. Converting 25p to 60p would be the same or bigger sin, i'am not even considering converting that mixed 50i/25p to 60p.

[jq963152]

Actually what should have happened is the dead of the i part of the equation. I shoot a lot of 25p for example (because I know that post will be fast enough and generally non-problematic, everything is tied to deadlines), which then at the final stage gets animated titles, which are 50i and everything packed into 50i stream, sick.

p.s. Converting 25p to 60p would be the same or bigger sin, i'am not even considering converting that mixed 50i/25p to 60p.

What do you mean with that?

Are you suggesting that instead of going for 50p or 60p only, there instead should only be 25p and 30p for broadcasting and no 50i and no 60i?

If yes, then there would be the problem that you would have to convert legacy 50i material (with 50 Hz temporal resolution) to 25p (with only 25 Hz temporal resolution) and legacy 60i material (with 60 Hz temporal resolution) to 30p (with only 30 Hz temporal resolution).

Not really perfect either...

Created a thread regarding that just a few days ago, see:

"60i to 30p (or 50i to 25p) without artifacts? (http://forum.doom9.org/showthread.php?t=168142)"

;)

[smok3]

What do you mean with that?
Thank you for your question, I will think deeply about your words put in my mouth and will try to figure out a replay, please give me a year or two to really think about it.

[paradoxical]

I can not imagine that doubling that would be such an insoluble issue...

Yes, I can imagine you wouldn't see the issue since you are, frankly, clueless.


Why would it be such an issue to increase the bandwith from 20 Mbps to 40 Mbps?

Increasing the bandwidth requires using more spectrum to do so. Spectrum is a finite resource and buying more of it is definitely not cheap (even slim bands of spectrum can cost billions of dollars especially in already crowded frequency blocks). You might want to read up on both physics and information theory.

[jq963152]

This is such an invalid argument IMHO.Your opinion means little.Maybe that should be reported as rule #4 violation :)?Go ahead.
Yes, I can imagine you wouldn't see the issue since you are, frankly, clueless.


Reported as violation of rule #4 :o ;).

[paradoxical]

Reported as violation of rule #4 :o ;).

Ooooooooooh, I'm quaking. It's the truth nonetheless. Acting like doubling bandwidth for broadcasting is some sort of trivial problem? That pretty much sums up how clueless you are. The only other answer is like nev said, you're simply trolling.

[jq963152]

That pretty much sums up how clueless you are.

You are probably mixing up cluelessness with something like visionariness there ;)?

Reported as rule #4 violation by the way :).

The only other answer is like nev said, you're simply trolling.

Uuups..., thanks for the reminder ;), forgot to report his post.

So:


You should really stop trolling every thread that even remotely touches the topic of frame rates or resolutions, its getting quite boring.


Reported as rule #4 violation :).

[Groucho2004]

Reported as rule #4 violation :).
Is that your new thing besides asking 20 questions per day? I have not noticed any profanity or insults in paradoxical's or nevcairiel's posts, just simple observations.

While we're at it, have a look at rules 1 and 11.

[jq963152]

Is that your new thing besides asking 20 questions per day?

From my user profile statistics:

Posts Per Day: 1.23

So what you're saying is far from being true.

And even if i would ask twenty questions per day (which i don't), then that would be perfectly fine AFAIK (as long as it would be proper questions).

I have not noticed any profanity or insults in paradoxical's or nevcairiel's posts, just simple observations.

IMHO they are repeatedly not nice to me with what they posted (which is what rule #4 requires). They accuse me of fictitious things and are acting derogatory towards me IMHO.

Yes, it could have been worse, but still.

While we're at it, have a look at rules 1 and 11.

And?

I don't see how i infringe any of those rules. IMHO my posts are valuable.

Besides, quite a few people are replying to my posts (in a reasonable way).

If you feel my posts are not worth being replied to or something like that, why do you even reply to them ;)?

Just a hint:

http://forum.doom9.org/profile.php?do=ignorelist

:)

And now maybe back to topic ;)?

[Wilbert]

Please be polite to each other and stay on topic. Oh, and also try to understand each other points and learn from them. Can't heart.

[jq963152]

Please be polite to each other and stay on topic. Oh, and also try to understand each other points and learn from them. Can't heart.

Thanks for taking care.

Why would it be such an issue to increase the bandwith from 20 Mbps to 40 Mbps?

Increasing the bandwidth requires using more spectrum to do so. Spectrum is a finite resource and buying more of it is definitely not cheap (even slim bands of spectrum can cost billions of dollars especially in already crowded frequency blocks). You might want to read up on both physics and information theory.

Came across this:




http://www.eutelsat.com/home/news/communiques-de-presse/press-list-container/eutelsat-launches-europes-first.html

Eutelsat launches Europe’s first dedicated Ultra HD (4K) channel

[...]

The new channel will operate in progressive mode at 50 frames per second. It will be encoded in MPEG-4 and transmitted at 40 Mbit/s in four Quad HD streams. Eutelsat is partnering with ATEME, a video compression solution provider to the broadcast industry, for the transmissions that will be uplinked to the EUTELSAT 10A satellite from its teleport in Rambouillet, near Paris.

[...]



;)

[wonkey_monkey]

I can not imagine that doubling that would be such an insoluble issue...

You can't imagine that it might cost twice as much and halve the number of broadcastable channels?

IMHO 20 Mbps are not enough for encoding 1080p30 or 1080i60 in H.264, especially when the encoding has to be done in real-time...

IMHO, it's plenty.

62.5 Mbps (H.264 HP@L4.2) would be better.

No it wouldn't. There'd be less space for other channels.

In this thread we learned that it is not possible to switch framerates on a TV channel because it's a continuous stream.

So it can only be one framerate.

Your conclusion does not follow from your statement! There can only be one framerate per channel (or as better fits the current situation, per territory), and we've managed just fine with that situation for years.

And actually 24p, 30p and 60i basically can be converted to 60p without any issue AFAIK.

Agreed on 24p to 30p - though it's a little bit wasteful - but converting from interlaced to progressive - when all TVs can already deinterlace - is a complete waste of bits. You're creating extra information which must then be encoded - information which was never present in the first place.

Apart from that, it's already happening already anyway.

AFAIK, NTSC TV channels already are broadcasting 24p, 25p, 30p, 50i and 60i footage at 60i (1080) and 60p (720).

Because they have to - as you've pointed out, channels can't switch framerate! It's not because they think 60i/60p is the best, and everything must be forced into it.

So it would be nothing new. The only new thing about the suggestion would be that it would be the full 1920x1080p @ 60fps and that "PAL land" would switch from 50 Hz to 60 Hz.

You may as well suggest the everyone in the UK starts driving on the right just because everyone in Europe already does it, despite the fact that all of our roads and signs are laid out for left-side driving, and all of our cars are made with right-hand drive. There's some inconvenience for anyone making the switch, but it's minor compared to the alternative.

David

[jq963152]

IMHO 20 Mbps are not enough for encoding 1080p30 or 1080i60 in H.264, especially when the encoding has to be done in real-time...IMHO, it's plenty.
The 1080i DVB-S2 H.264 broadcasts i know of have ridiculous picture quality.

Some of them even use anamorphic 1440x1080i, instead of the full 1920x1080i :(.

Apart from that, you can try it yourself: just try encoding 1080p25 or 1080p30 or 1080i50 or 1080i60 with x264 with --vbv-maxrate 20000 at real-time speed...

... the result won't be pretty (IMHO).

[TheSkiller]

In addition to the problems a switch from 50 Hz to 60 Hz would cause I don't see any possible benefits for future recordings in all 50 Hz terretories switching to 60 Hz.

It's not like anybody can tell the difference between a 50 fps video displayed at 50 Hz and a 60 fps video displayed at 60 Hz. If you think otherwise you have probably never tested it.
So what would be the point besides eliminating all other frame rates? We would no longer have to convert between frame rates but according to you this is not an issue anyway since you suggest to convert everything to 60 fps.

[wonkey_monkey]

Apart from that, you can try it yourself:

Running x264 on a PC has little relation to using a broadcast-quality professional encoder (which probably costs an order of magnitude more than the PC).

[kieranrk]

Running x264 on a PC has little relation to using a broadcast-quality professional encoder (which probably costs an order of magnitude more than the PC).

I would beg to differ. Have you ever evaluated x264 vs a "professional" encoder?

[jq963152]

In addition to the problems a switch from 50 Hz to 60 Hz would cause I don't see any possible benefits for future recordings in all 50 Hz terretories switching to 60 Hz.

There would be two huge benefits.

The first one being:

60 Hz (60 fps) is the native refresh rate of almost all (modern) displays today.

Even some HDTVs targeted for the european market still have (visible) issues with 50 Hz input signals.

So, worldwide adoption of 60 Hz (60p i.e. 60 frames per second) video would be much better.

So what would be the point besides eliminating all other frame rates? We would no longer have to convert between frame rates

That would be the second huge benefit.

but according to you this is not an issue anyway since you suggest to convert everything to 60 fps.

I did not suggest that. I suggested to do that for legacy material which has not been recorded at 60 fps.

All future recordings should then be done at 60 fps of course (not just TV productions, but also movies and so on).

And i did not say that 50 Hz to 60 Hz conversion would not be an issue. Of course it is an issue.

But AFAIK it would not work without such a compromise.

[foxyshadis]

Is there an actual point to this thread? It's basically a combination wishlist and argument that no one involved in any standards is ever going to hear.

jq963152, submit your wishlist to the relevant ETSI, ISO, and MPEG committees. Let us know how that goes. Beyond that, this thread should probably be closed.

[Blue_MiSfit]

This is silly. I'm closing this thread.