Digital TV Banter

Digital TV Banter (https://www.digitaltvbanter.co.uk/forum.php)
-   uk.tech.digital-tv (Digital TV - General) (https://www.digitaltvbanter.co.uk/forumdisplay.php?f=4)
-   -   4k TV on Freesat or Freeview? (https://www.digitaltvbanter.co.uk/showthread.php?t=34439)

Java Jive[_2_] August 13th 15 05:31 PM

4k TV on Freesat or Freeview?
 
Bearing in mind the pictures that I have just posted elsewhere in this
thread, can you explain what test you have done?

I ask because I can see no evidence at all that my, admittedly old and
first generation, LCD is doing any de-interlacing.

On Thu, 06 Aug 2015 23:43:43 +0100, Andy Furniss [email protected] wrote:

I do know that my TV de-interlaces as I can test it with a computer.

--
================================================== ======
Please always reply to ng as the email in this post's
header does not exist. Or use a contact address at:
http://www.macfh.co.uk/JavaJive/JavaJive.html
http://www.macfh.co.uk/Macfarlane/Macfarlane.html

Andy Furniss[_3_] August 13th 15 10:24 PM

4k TV on Freesat or Freeview?
 
Java Jive wrote:
Bearing in mind the pictures that I have just posted elsewhere in
this thread, can you explain what test you have done?


It's connected via hdmi and advertises interlaced and progressive modes
so I can see the difference between playing the same test sequences.

When in interlaced mode I can see that it weaves static = I can see a
full res test pattern. If something moves eg. I move a mouse over the
pattern I can see that it is blending the mouse cursor and an area
around the motion.

Using test videos like fast scrolling text I can see that the TV, when
in interlaced mode is de-interlacing them to field rate. I can do the
same thing in s/w when in progressive mode on TV or monitor so I know
what it looks like de-interlaced.

The TV, because it is motion adaptive turns the vid back to full vert
res weaved if I pause it.

It takes a bit of messing around to actually use the TV correctly in an
interlaced mode as computers don't tend to do field sync. I use linux
and an ffmpeg filter via mpv that works around this - all you need is
vsync at field rate (which is normal). You also need to convert to 422
chroma in an interlaced aware way - but unless anyone actually needs all
the detail I won't go as far as giving full command lines and
justifications.

I ask because I can see no evidence at all that my, admittedly old
and first generation, LCD is doing any de-interlacing.


I don't see any evidence it doesn't - but then you paused to get your
pics and I don't know whether that changes anything on your setup(s).
I don't know what motion was going on in that shot or what res to expect
- but it looks progressive and assuming the source was truly interlaced
- which for sport is highly likely then that means it got de-interlaced.

IIRC from a previous thread your panel was higher res than SD so it must
be scaling. If it is just scaling one field at a time that counts as
deinterlacing - though it would likely do more than that or it would bob
up and down.

As for LCD motion artifacts you mention in your other post - early LCDs
were poor and may have had poor de-interlacers, without seeing it
running it's hard to know what you mean.

Comparing progressive modes/content I count my 120Hz "capable" LCD
monitor as quite crap (in a blurry/not fast enough way) when compared to
my plasma TV. If I drag around the window I am typing this into, it
doesn't take much speed to make the text unreadable which doesn't happen
on the plasma - or on a CRT monitor.

On Thu, 06 Aug 2015 23:43:43 +0100, Andy Furniss [email protected] wrote:

I do know that my TV de-interlaces as I can test it with a
computer.


Java Jive[_2_] August 13th 15 11:51 PM

4k TV on Freesat or Freeview?
 
On Thu, 13 Aug 2015 23:24:52 +0100, Andy Furniss [email protected] wrote:

Java Jive wrote:
Bearing in mind the pictures that I have just posted elsewhere in
this thread, can you explain what test you have done?


It's connected via hdmi and advertises interlaced and progressive modes
so I can see the difference between playing the same test sequences.

[snip]


Thanks for the explanation.

I ask because I can see no evidence at all that my, admittedly old
and first generation, LCD is doing any de-interlacing.


I don't see any evidence it doesn't


I beg to differ. I go into some detail on the subject, referencing a
particular image of an athlete's arm in the CRT vs LCD web-page. I am
convinced that neither that nor any other video is de-interlaced by my
TV. It's a very early model Panasonic with analogue tuners.

- but then you paused to get your
pics and I don't know whether that changes anything on your setup(s).
I don't know what motion was going on in that shot or what res to expect
- but it looks progressive and assuming the source was truly interlaced
- which for sport is highly likely then that means it got de-interlaced.


The main motion would be the forerunner ski-racer on the piste - a
forerunner is a not-quite-yet-top or a formerly-top-but-now aging
ski-racer who descends the slope at near full speed to check it's safe
for the main racers who will follow when the competition starts. It's
a bit difficult to estimate how fast the skier would be travelling
from being stationary at the start gate, but a back-of-an-envelope
calc suggests probably about 50-55 fps, or around 35mph (on some
slopes they can reach 100mph). At the scale of the picture 10px ~
1ft, that's about 10px per interlaced field. I would have thought
that if there were motion artifacts to be seen, we would be able to
see them in this picture, but we can't.

So why can't we? Well ...

IIRC from a previous thread your panel was higher res than SD so it must
be scaling.


.... it's a little bigger than SD horizonally, but a little less
vertically, so yes, it's certainly doing some scaling, but ...

If it is just scaling one field at a time that counts as
deinterlacing - though it would likely do more than that or it would bob
up and down.


.... well it's not really de-interlacing, as that would involve working
with more than one field at a time, and the picture of the athlete's
arm in the website pages suggests that it is not.

So it's certainly buffering, and it's certainly scaling, it HAS to be
doing both, but de-interlacing, I'm pretty sure not.

So why no artifacts? Well this LCD doesn't have the vertical
resolution it should, 492 vs 576, so perhaps they are being lost in
being scaled, or perhaps they are doing the same trick as with CRTs,
and making each scan line overlap over its neighbours in the previous
field.

As for LCD motion artifacts you mention in your other post - early LCDs
were poor and may have had poor de-interlacers, without seeing it
running it's hard to know what you mean.


Well, I wasn't saying that my LCD shows artifacts, in fact I've never
seen it show the combing effects claimed on interlaced input, only
dot-crawl from using CV as the input source. Merely I was suggesting
why in general LCDs might legitimately show such motion artifacts if
fed an interlaced signal, because they their technology allows them to
display what they are fed more faithfully than CRTs.
--
================================================== ======
Please always reply to ng as the email in this post's
header does not exist. Or use a contact address at:
http://www.macfh.co.uk/JavaJive/JavaJive.html
http://www.macfh.co.uk/Macfarlane/Macfarlane.html

R. Mark Clayton[_2_] August 14th 15 10:09 AM

4k TV on Freesat or Freeview?
 
On Wednesday, 12 August 2015 21:42:29 UTC+1, _Unknown_Freelancer_ wrote:
"Brian Gregory" wrote in message
...
On 03/08/2015 12:07, _Unknown_Freelancer_ wrote:
"Michael Chare" wrote in message
...
4K TVs are becoming available and more reasonable prices and there are
some 4k internet streaming channels.

So when, if ever, will any of the Freesat or Freeview channels convert
to
4K?

--
Michael Chare


4K ....over Freeview?
One hopes you are 'having a laugh'. Seriously.

Consider that this stuff originates at 12Gb/s.
Tweleve gigabits of data, per second.
In comparison, present HD originates at just under 1.5Gb/s.

And to compress this enough so that its transmittable over freeview, you
would have to dispense with so much information that you would render
'UHD'
pointless.


You could use a whole DVB-T2 multiplex and the H.265 codec.

It would arguably be a better redition of 4K than the current HD Freeview
is of 2K.


Yes, you could.

But it would contain f.all detail.


I don't think you understand how video compression works. In a [typical] video stream there will be areas of the pictures which are the same as neighbouring ones and / or the same from frame to frame. It is these that are compressed, NOT the fine detail.

Even simple [and lossless] run length encoding will result in a substantial reduction in the file size for an image.

MP3 compression of audio from original CD to 128kbps results in compression of around 91%. Purists say you can tell the difference and so for a while I abjured it, however even though I have quite good ears, I can't tell the difference.

Obviously if one is able to compress in two dimensions and time one will achieve much higher compression with relatively little loss. So the method described above of using a [single] full [satellite] transponder and H.265 compression will result in a very good result at 4k resolution for most all content.


Yes, you're transmitting the information for four times more dots, but there
just is not enough bandwidth available over DTT to allow it to carry any
level of Ultra High _Definition_. ....with particulat emphasis on the word
'Definition'.

All you're doing is making the picture bigger.
Not improving the quality.





i.e. The loss in quality would be so bad that it would be comparable to
existing HD. Therefore, just what is the point?

The only way to transmit 'acceptable' UHD is either usings serveral
channel
spaces over satellite, or proper* broadband.
* meaning something better than BTs standard twisted pair phone line.
Either
the co-ax that Virgin is installing, or fibre from both BT and Virgin.

4K over Freeview?
Hopefully, never.



--

Brian Gregory (in the UK).
To email me please remove all the letter vee from my email address.



_Unknown_Freelancer_ August 14th 15 11:03 AM

4k TV on Freesat or Freeview?
 
Snippity snip snip delete (well, seems thats what the cool kids do)

4K ....over Freeview?
One hopes you are 'having a laugh'. Seriously.

Consider that this stuff originates at 12Gb/s.
Tweleve gigabits of data, per second.
In comparison, present HD originates at just under 1.5Gb/s.

And to compress this enough so that its transmittable over freeview,
you
would have to dispense with so much information that you would render
'UHD'
pointless.

You could use a whole DVB-T2 multiplex and the H.265 codec.

It would arguably be a better redition of 4K than the current HD
Freeview
is of 2K.


Yes, you could.

But it would contain f.all detail.


I don't think you understand how video compression works. In a [typical]
video stream there will be areas of the pictures which are the same as
neighbouring ones and / or the same from frame to frame. It is these that

are compressed, NOT the fine detail.

Even simple [and lossless] run length encoding will result in a substantial
reduction in the file size for an image.

MP3 compression of audio from original CD to 128kbps results in compression
of around 91%. Purists say you can tell the difference and so for a while
I abjured it, however even though I have quite good ears, I can't tell the
difference.

Obviously if one is able to compress in two dimensions and time one will
achieve much higher compression with relatively little loss. So the method
described above of using a [single] full [satellite] transponder and H.265

compression will result in a very good result at 4k resolution for most all
content.



I actually laughed out loud at my monitor when I read your first sentence!

sarcasm
n decades* working in live television, and n+1 decades programming computers
in up to sixteen different laguages.
I really do not have the slightest clue.
/sarcasm


My dear chap, I am very well aware of the constructs employed in standards
deployed by the moving picture experts group.
Ground zero for it was the technique employed in a jpg picture, and then
applying that over a time period.

Yes, when turning down the bitrate for a video stream, the first items to be
sacrificed ARE the non-moving areas*.
Keep turning down the bit rate, as has to be done for DTT, and very soon you
start to afffect the detail contained within anything moving.
In fact, whilst an object is moving, its detail is lost.
Only once the 'thing' stops moving is its detail restored. (Again, this is
detail lost because of compression, NOT motion blur.) Its restored quickly.
So much so that if you're not paying particulat attention to the moving
this, you miss the instant when a vauge fuzzy shape becomes a pitch, with
its skid marks and divots.
Thus, grass on any field based sport shown on DTT turns to VHS mush whilst
the camera is tracking play.
Only when the camera becomes still is any detail added to the pictch.

* What goes against your suggestion was a 'reality' programme on ITV32 a
couple of years back.
Which showed a QR code at the top of the screen constantly. This would link
to chat rooms and a website.
The QR code was stationary, and on screen during the entire programme.
i.e. It was present on every frame and did not move.
No-ones phone could recognise it because the transmission chain crushed all
the detail out of it.
It was a stationary item, compressed to grey mush.
HD?
No 'definition'.

I digress.
If you saw such programmes at source quality, and then compared the DTT off
air signal, you would understand just how much detail Arqiva are removing
from the national viewing pleasure.

To get a 12Gb/s signal out over DTT, you are going to have to drop a vast
amount of detail.

Take a football match being covered in 4K. (Ive seen some.)
Now get your main gantry camera to frame on a stationary wide angle* facing
across the pitch.
*As wide as the lens will go. Not pointing at anything in particular.
Such is the detail in that picture you can make out individual facial
expressions of people in the crowd in the opposite stand (usually around
100metres away).
Its a fair wager that detail will _never_ be present in any "Freeview 4K".


Instead of wasting so much bandwidth on such guff, why not just turn up the
bandwidth for present HD channels, and make Freeview better quality than Sky
satellite or BT TV???
i.e. Make HD.... the best HD. FOR FREE.




Bear in mind CD audio is sampled at 44.1KHz. The entire broadcast world uses
48KHz sample rates for everything. So that CD has already lost the very high
frequencies. Anything above 22KHz to be precise.
Then compressing that in to a 128Kb/s mp3 file..... Any sound supervisor
worth their salt CAN recognise an mp3.
There are several who refuse to use mp3 files in their programmes, insisting
on .wav source files.

If you take a 96Kb/s mp3 file, its easy to hear how bad it is.
128Kb/s is only just a little above that.
i.e. That file is only just above the point where audio is defeningly
obvious how bad it is.



H265 is H264 tweaked.
The macroblock size is increased (the grid which the encoder chops the
picture in to), and they rejigged the colour space. Thats it.
Its really not worth getting all moist about h265 as being the best thing
since the Altair 8800.
Its not some amazing soloution that will allow cinema quality pictures to
pushed down a dial up connection.

Just like that 128Kb/s mp3 file, its h264, but just above the point where
the masses can tell its crap.




Roderick Stewart[_3_] August 14th 15 04:37 PM

4k TV on Freesat or Freeview?
 
On Fri, 14 Aug 2015 12:03:38 +0100, "_Unknown_Freelancer_" /dev/null
wrote:

Instead of wasting so much bandwidth on such guff, why not just turn up the
bandwidth for present HD channels, and make Freeview better quality than Sky
satellite or BT TV???
i.e. Make HD.... the best HD. FOR FREE.


I've encountered people with Freeview HD receivers selecting channels
1 to 4 even when I've pointed out to them that the same programmes are
available in HD on channels 101 to 104, and they are apparently quite
happy with what they are watching. Maybe they just can't be bothered
to type the extra digits, or don't see any advantage.

Rod.

Vir Campestris August 14th 15 08:57 PM

4k TV on Freesat or Freeview?
 
On 14/08/2015 12:03, _Unknown_Freelancer_ wrote:
snip
Instead of wasting so much bandwidth on such guff, why not just turn up the
bandwidth for present HD channels, and make Freeview better quality than Sky
satellite or BT TV???
i.e. Make HD.... the best HD. FOR FREE.


Onik. Flap. Oink. Flap.


Bear in mind CD audio is sampled at 44.1KHz. The entire broadcast world uses
48KHz sample rates for everything. So that CD has already lost the very high
frequencies. Anything above 22KHz to be precise.
Then compressing that in to a 128Kb/s mp3 file..... Any sound supervisor
worth their salt CAN recognise an mp3.
There are several who refuse to use mp3 files in their programmes, insisting
on .wav source files.

If you take a 96Kb/s mp3 file, its easy to hear how bad it is.
128Kb/s is only just a little above that.
i.e. That file is only just above the point where audio is defeningly
obvious how bad it is.

H265 is H264 tweaked.
The macroblock size is increased (the grid which the encoder chops the
picture in to), and they rejigged the colour space. Thats it.
Its really not worth getting all moist about h265 as being the best thing
since the Altair 8800.
Its not some amazing soloution that will allow cinema quality pictures to
pushed down a dial up connection.

Just like that 128Kb/s mp3 file, its h264, but just above the point where
the masses can tell its crap.


I can hear the defects in 128k MP3, but not in 256k MP3 nor in 128k WMA.
I haven't tried on AAC. DAB, OTOH? 80k MP2? Yuck.

And as for Freeview? My techie son and I have learned not to mention
the faults when my wife is there. She just gets cross.

Andy

Andy Furniss[_3_] August 14th 15 11:17 PM

4k TV on Freesat or Freeview?
 
Java Jive wrote:

I ask because I can see no evidence at all that my, admittedly
old and first generation, LCD is doing any de-interlacing.


I don't see any evidence it doesn't


I beg to differ. I go into some detail on the subject, referencing
a particular image of an athlete's arm in the CRT vs LCD web-page. I
am convinced that neither that nor any other video is de-interlaced
by my TV. It's a very early model Panasonic with analogue tuners.


Looking at that image and reading your site I think what I would call
de-interlacing differs from what you would.

I consider any processing to display interlaced content on a progressive
display as de-interlacing. It may be that it is very simple like line
doubling fields and a bit of filtering to compensate for the different
spatial position of the fields, but it's still processing that is needed
and wouldn't be if a native interlaced display were being driven. Doing
line doubling hurts resolution just weaving fields together is good for
res with static portions of the frame but artifacts on motion. Being
clever and adding edge interpolation/motion detection to get the best of
both is what "advanced" de-interlacers do.

Judging by the arm shot your TV looks like it's doing fields - I accept
what you write about the pairs not matching, but that could be
additional processing that's nothing to do with interlacing.

I don't know the details of your screen, but it could be, like the
monitor I am currently looking at, only 6 bit and using spatial and or
temporal dithering to fake 8 bit colour, which could explain why the
lines differ 1:1 chroma but the brighter edge steps look paired.

As well as colour LCDs also have to pull tricks to get motion to "work"
- just holding pixels on for the whole frame/field period is great for
flicker, but is really bad for making motion look blurred, so they
"soomehow" have to try and work around that.


- but then you paused to get your pics and I don't know whether
that changes anything on your setup(s). I don't know what motion
was going on in that shot or what res to expect - but it looks
progressive and assuming the source was truly interlaced - which
for sport is highly likely then that means it got de-interlaced.


The main motion would be the forerunner ski-racer on the piste - a
forerunner is a not-quite-yet-top or a formerly-top-but-now aging
ski-racer who descends the slope at near full speed to check it's
safe for the main racers who will follow when the competition starts.
It's a bit difficult to estimate how fast the skier would be
travelling from being stationary at the start gate, but a
back-of-an-envelope calc suggests probably about 50-55 fps, or around
35mph (on some slopes they can reach 100mph). At the scale of the
picture 10px ~ 1ft, that's about 10px per interlaced field. I would
have thought that if there were motion artifacts to be seen, we would
be able to see them in this picture, but we can't.

So why can't we? Well ...

IIRC from a previous thread your panel was higher res than SD so it
must be scaling.


... it's a little bigger than SD horizonally, but a little less
vertically, so yes, it's certainly doing some scaling, but ...


I notice some of your pics are from tiny screens (OK the 22" isn't so
tiny) I wonder given how much portables used to overscan whether they
could be doing 1:1 and loosing lots off top & bottom - just a thought.

If it is just scaling one field at a time that counts as
deinterlacing - though it would likely do more than that or it
would bob up and down.


... well it's not really de-interlacing, as that would involve
working with more than one field at a time, and the picture of the
athlete's arm in the website pages suggests that it is not.


Multi field deinterlacing is "advanced" - it's possible but sub optimal
to call processing one field de-interlacing. processing may involve more
than just line doubling eg. edge detection and smoothing to hide the
half res steps on diagonals.

So it's certainly buffering, and it's certainly scaling, it HAS to
be doing both, but de-interlacing, I'm pretty sure not.

So why no artifacts? Well this LCD doesn't have the vertical
resolution it should, 492 vs 576, so perhaps they are being lost in
being scaled, or perhaps they are doing the same trick as with CRTs,
and making each scan line overlap over its neighbours in the
previous field.


I doubt weave would be lost in scaling - IME it looks far worse when scaled.

As for overlapping - well yes, though I would say totally overwrite
rather that overlap. I wonder whether a pic of a full size CRT would
look the same as your portable. Saying that, though, I recall reading
years ago that rf output ist gen game consoles/early home computers used
to pull a trick on interlaced CRTs to make them progressive (IIRC there
is a pulse/something to mark top/bottom field and they just repeated one
rather than alternating). It did say this worked on most but not all TVs
and I recall wondering "why no gaps", so perhaps all CRTs have a fat
spot/spots.


As for LCD motion artifacts you mention in your other post - early
LCDs were poor and may have had poor de-interlacers, without seeing
it running it's hard to know what you mean.


Well, I wasn't saying that my LCD shows artifacts, in fact I've
never seen it show the combing effects claimed on interlaced input,
only dot-crawl from using CV as the input source. Merely I was
suggesting why in general LCDs might legitimately show such motion
artifacts if fed an interlaced signal, because they their technology
allows them to display what they are fed more faithfully than CRTs.


Perhaps it's to do with things other than interlacing - like it's hard
to get slow/always on LCDs to do motion properly.



Andy Furniss[_3_] August 14th 15 11:47 PM

4k TV on Freesat or Freeview?
 
_Unknown_Freelancer_ wrote:

Take a football match being covered in 4K. (Ive seen some.) Now get
your main gantry camera to frame on a stationary wide angle* facing
across the pitch. *As wide as the lens will go. Not pointing at
anything in particular. Such is the detail in that picture you can
make out individual facial expressions of people in the crowd in the
opposite stand (usually around 100metres away). Its a fair wager that
detail will _never_ be present in any "Freeview 4K".


Instead of wasting so much bandwidth on such guff, why not just turn
up the bandwidth for present HD channels, and make Freeview better
quality than Sky satellite or BT TV???


I am curious what bitrate you think is enough for HD?

AIUI BT HD is max 10mbit for their (BTW) premium offering or 7.5 1440
standard. For UHD they require a connection of 44mbit (I really hope
they are not allowing for record one and watch another in that!).

I don't know what Sky HD uses, though I have "come across" some SD
transport stream motorsport rips that don't seem any higher than the
same content from the BBC.

i.e. Make HD.... the best HD. FOR FREE.


I agree that should happen - but what Brian wrote was -

"You could use a whole DVB-T2 multiplex and the H.265 codec.
It would arguably be a better redition of 4K than the current HD
Freeview is of 2K."

I think 40mbit hevc 2160p50 should surely be better than the current HD
offering - maybe in the case of 1080i25 high motion bits even better
than "quality/raw HD" deinterlaced/scaled up for a UHD TV.

I guess it should really go to HD - but there's a T2 mux spare on my
transmitter with just nulls and QVC.

Hmm, maybe I should find the keys to my lab and test -

Would grainy 65mm 2160p50 film scans with professionally made (for VQEG)
1080i25 derivatives do :)


_Unknown_Freelancer_ August 15th 15 12:02 AM

4k TV on Freesat or Freeview?
 
"Roderick Stewart" wrote in message
...
On Fri, 14 Aug 2015 12:03:38 +0100, "_Unknown_Freelancer_" /dev/null
wrote:

Instead of wasting so much bandwidth on such guff, why not just turn up
the
bandwidth for present HD channels, and make Freeview better quality than
Sky
satellite or BT TV???
i.e. Make HD.... the best HD. FOR FREE.


I've encountered people with Freeview HD receivers selecting channels
1 to 4 even when I've pointed out to them that the same programmes are
available in HD on channels 101 to 104, and they are apparently quite
happy with what they are watching. Maybe they just can't be bothered
to type the extra digits, or don't see any advantage.

Rod.


Either blissfully oblivious/thick, or cant be arsed to press two more
buttons.
.....one digit memory




All times are GMT. The time now is 05:57 PM.

Powered by vBulletin® Version 3.6.4
Copyright ©2000 - 2020, Jelsoft Enterprises Ltd.
SEO by vBSEO 2.4.0
Copyright 2004-2006 DigitalTVBanter.co.uk