|
||||||||
1920 vs 1440 the test |
![]() |
|
|
Thread Tools | Search this Thread |
|
|
#1 |
|
Forum Member
Join Date: Jan 2007
Posts: 19,460
|
1920 vs 1440 the test
I have, as promised, run a test comparing the 1920x1080 ITV1 HD Granada broadcast of the final episode of Monte Carlo or Bust with the 1440x1080 ITV1 HD London version. I'm afraid I am quite dismayed by the results, given the reassurances we have been given by the BBC that the difference is negligible (1440 is used for ITV1 HD London & Meridian, both BBC HD channels and all Freeview HD channels).
My reluctant conclusion is this: fine detail that exists in the 1920 version simply does not exist in the 1440 version, it is wiped out by the lower resolution. This has the additional effect of reducing the brilliance of objects and giving a 'soft' effect as many people have attested to noticing, particularly on the Sky HD forums where they have the advantage of comparing to a majority of 1920x1080 HD channels on the Sky platform. Below are the comparison images- the main links are jpg images, the png files are losslessly compressed versions 2-5MB in size- it should be noted you do get some more ringing/blocking artifacts in the 1920 versions due to ITV1 HD Granada being run at an insufficient bitrate, however the greater detail is still clearly visible: NOTE: all images should be saved and viewed at full (100%) magnification on an image viewer such as Windows Photo Viewer, IrfanView, or Faststone Image Viewer. In Windows Photo Viewer click the box with four small arrows around it to zoom to full size- in Vista and Windows 7 you should be able to have two images open and then hold the mouse pointer over their icon in the taskbar to bring up two preview windows which you can quickly flick between by moving the pointer back and forth over each one. ITV1 HD London (png) vs ITV1 HD Granada (png) image 1 compare Jodie Kidd's woollen poncho, in the Granada version the vertical line texture is clearly visible to the left of the buttons and the fine wool grain can be seen in the brightly lit side over her right arm, both of these details are invisible in the 1440 London version. ITV1 HD London (png) vs ITV1 HD Granada (png) image 2 compare the stubble of both men, detail can be seen in the Granada version which is smudged in the London version. Compare the centre man's purple lined shirt- in the Granada version you can see there are two purple lines with a thinner white line in between, on the London version you cannot see this at all, the two purple lines are seemingly merged into one due to the lack of resolution. ITV1 HD London (png) vs ITV1 HD Granada (png) image 3 Compare the hair of Jack Dee, Jack & Adrian's skin, Jack's glasses and Adrian's hat, detail is there in the Granada version which is clearly missing in the London version. ITV1 HD London (png) vs ITV1 HD Granada (png) image 4 Compare the woollen detail in Jodie Kidd's beret, the sharpness of the both of their hair, the detailed texture of Julian's jacket, the lines in the French flag, all missing in the London version. And this, a very significant point which ties up with the observations of a lack of brilliance & wow factor in the BBC HD channels- look at the shinyness and sharpness of the light on Julian's glasses and Jodie's earrings in the Granada version giving an almost 3D quality, which is completely blurred and lost in the 1440 London version .ITV1 HD London (png) vs ITV1 HD Granada (png) image 5 Compare Rory & Penny's skin & hair & Rory's stubble, clearly more detail in the Granada version. Compare the sharpness & shinyness of Penny's chainmail shawl, and the lines in Rory's shirt, absent in the London version due to lack of resolution. ITV1 HD London (png) vs ITV1 HD Granada (png) image 6 not a huge difference as nothing is particularly in focus, the VW van, crops and foliage is slightly sharper in the Granada version, and as in all the other images the ITV1 HD logo is more blurred in the London version. ITV1 HD London (png) vs ITV1 HD Granada (png) image 7 More detail and greater shinyness on Jack's glasses, more detail on Adrian's hat, the line detail on Adrian's jacket is almost completely lost on in the London version. ITV1 HD London (png) vs ITV1 HD Granada (png) image 8 more detail in the foliage & cars in the Granada version, you can just about make out the VW logo on the Granada version, just a blur on the London version. Obviously this is disappointment for Freesat & Freeview users who were reassured by the BBC that there was no significant difference between 1920 & 1440 HD, I think as you can see from the above images that there is a difference, not just in loss of detail, but a loss of brilliance and a 3D effect, as some call it, a wow factor. To put things in context though,1440 is certainly much, much better than SD, and in many cases, will give better results than 720p HD used in many other countries, so I can say that your investment has certainly not been a waste. I do think however, that the BBC should aim to switch to 1920, this could probably be done in a couple of years time when encoding improves, and more space is availiable on satellite with the new Astra launches. I do think though people should make clear to BBC their wishes on this issue, as there may be pressure to fit another HD channel into the Freeview HD mux which would likely keep all FTA HD at 1440 level for years. |
|
|
|
|
Please sign in or register to remove this advertisement.
|
|
|
#2 |
|
Forum Member
Join Date: Apr 2007
Posts: 37
|
Quote:
given the reassurances we have been given by the BBC that the difference is negligible ...................... between 1920 & 1440 HD,
|
|
|
|
|
|
#3 |
|
Inactive Member
Join Date: Jul 2001
Location: Berkshire
Posts: 4,686
|
To be fair to the BBC, I don't think they have ever claimed there is no difference between a 1920 and 1440 image per se, but that at normal viewing distances on typical consumer displays, the average viewer would be hard pushed to tell the difference.
Of course you can see a difference when pixel peeping a screen grab, close up on a computer monitor. |
|
|
|
|
|
#4 |
|
Forum Member
Join Date: Oct 2003
Location: Darn Sarf
Posts: 28,730
|
Quote:
it should be noted you do get some more ringing/blocking artifacts in the 1920 versions due to ITV1 HD Granada being run at an insufficient bitrate, however the greater detail is still clearly visible:
I suspect that more people than you might expect will feel the same way (or not care either way). I agree though, that for people who must have the absolute maximum horizontal resolution, 1920 should be slightly better, if they sit really close to the screen, on some material. It isn't going to be an issue though, for the vast majority of HD viewers watching at their normal viewing distance and I'm sure this is part of the reason why the BBC and some others do the 1440 thing. |
|
|
|
|
|
#5 |
|
Forum Member
Join Date: Jul 2004
Posts: 16,670
|
We have a Pace Sky HD box and a Humax Freeview PVR and there is no question that the HD pictures on ITV HD are slightly but not noticeably better on Sky.
SD pictures are another matter. The reverse is true. |
|
|
|
|
|
#6 |
|
Forum Member
Join Date: Oct 2005
Location: Retford
Posts: 20,450
|
Happily I can download the pictures and place them on my Linux Desktop and use Gwenview to flick between the pictures in fullscreen. My HD TV in the living room is connected via HDMI to my PC and so I can do a comparison from my seating position.
I can see that the London version is a bit blurrier than its Granada counterpart, and in some pictures, from a good few feet from the TV, I could tell that some detailed parts of the screen (such as strands of hair, the lined purple shirt, tarmac on the road and the ITV1 HD DOG) were less defined on the London version. Although I have to say, the difference isn't as apparent as I feared, there was very little difference in many parts of the comparisons that I saw. I suspect the broadcasters are doing what they already done to DAB and Freeview, reducing bitrate to fit more in, because we who notice these things are a small minority compared to the majority who just watch and don't notice. Although I would say, 1920x1080 should be used for HD where possible. |
|
|
|
|
#7 |
|
Forum Member
Join Date: Mar 2006
Location: Birmingham
Posts: 121
|
Full-screen?
I have just downloaded the first 2 images and using Nero photosnap viewer (the default on my PC) the difference is immediate even without displaying at full screen (19" @1280x1024). I never realised there was such a difference.
|
|
|
|
|
|
#8 |
|
Forum Member
Join Date: Oct 2002
Posts: 9,343
|
Image1 London is 1920x1080 pixels 481644 bytes
Image1 Granada is 1194x672 pixels 513158 bytes |
|
|
|
|
|
#9 |
|
Forum Member
Join Date: Feb 2004
Posts: 722
|
Bear in mind if you're looking at the images that you must compare the images whilst sat on your sofa, with the images on your TV.
Any comments from comparing them on a PC monitor 10 inches away from your face is surely irrelevant. |
|
|
|
|
|
#10 |
|
Forum Member
Join Date: Apr 2005
Location: Surrey, UK
Posts: 1,302
|
The comparison does not reflect reality because each device has a different video processor and different strengths/weaknesses. Some content works better on some devices compared to others. I know that the Sky+HD box even has different attributes compared to the Foxsat-HD even though the part comes from the same manufacturer but they are different generations.
The tests I am aware of by the BBC involved real consumer devices under ideal viewing conditions and are not based on the screen caps from a software based decoder which is paused which is totally different. Sorry. |
|
|
|
|
#11 |
|
Forum Member
Join Date: Oct 2008
Location: Worcester
Posts: 4,185
|
Quote:
Image1 London is 1920x1080 pixels 481644 bytes
Image1 Granada is 1194x672 pixels 513158 bytes The 2 png files are both 1920 x 1080 but the bit depths are different 24 Granada and 32 London also file sizes are 2.52 and 5.22 MB respectively. So even the 'like for like' aren't. |
|
|
|
|
|
#12 |
|
Forum Member
Join Date: Jan 2007
Posts: 19,460
|
Quote:
To be fair to the BBC, I don't think they have ever claimed there is no difference between a 1920 and 1440 image per se, but that at normal viewing distances on typical consumer displays, the average viewer would be hard pushed to tell the difference.
Of course you can see a difference when pixel peeping a screen grab, close up on a computer monitor. Quote:
Interesting comparisons, but as I find compression artifacts far more disturbing than absolute horizontal picture detail, which to my eyes is grossly overrated, I have no problem with broadcasters using 1440 instead of 1920, if it reduces the artifact level.
Quote:
Image1 London is 1920x1080 pixels 481644 bytes
Image1 Granada is 1194x672 pixels 513158 bytes Quote:
Bear in mind if you're looking at the images that you must compare the images whilst sat on your sofa, with the images on your TV.
Any comments from comparing them on a PC monitor 10 inches away from your face is surely irrelevant. Quote:
I have just downloaded the first 2 images and using Nero photosnap viewer (the default on my PC) the difference is immediate even without displaying at full screen (19" @1280x1024). I never realised there was such a difference.
Quote:
The tests I am aware of by the BBC involved real consumer devices under ideal viewing conditions and are not based on the screen caps from a software based decoder which is paused which is totally different.
Sorry. Quote:
Have you clicked like for like? there are four images of 'Image 1' 2 jpg, 2 png
The 2 png files are both 1920 x 1080 but the bit depths are different 24 Granada and 32 London also file sizes are 2.52 and 5.22 MB respectively. So even the 'like for like' aren't. |
|
|
|
|
|
#13 |
|
Forum Member
Join Date: Jan 2007
Posts: 19,460
|
My photobucket/mediafire stats are suggesting most people are only looking at images 1 & 2 comparison, can I suggest anyone who hasn't looks at images 3, 4, 5 as these show the difference much more clearly. I've tested with Firefox, Chrome, and IE and these should all be able to open two images in separate windows at full zoom (click + magnifying glass on image) and allow you to mouse over each image window in the taskbar with Aero on Windows Vista/7 to do an easy comparison. The best way to do it is to push the image to the corners using the scrollbars so you have the images precisely lined up.
|
|
|
|
|
|
#14 |
|
Forum Member
Join Date: Apr 2001
Location: Northern Scottish Highlands
Posts: 11,307
|
Quote:
Interesting comparisons, but as I find compression artifacts far more disturbing than absolute horizontal picture detail, which to my eyes is grossly overrated, I have no problem with broadcasters using 1440 instead of 1920, if it reduces the artifact level.
I suspect that more people than you might expect will feel the same way (or not care either way). I think compression artefacts are what "spoil" digital tv. I only have a 720 line "hd ready" tv but I've done a lot of comparisons, and for me the biggest improvement when switching from an SD channel to the HD version, is the big drop in compression artefacts. For instance, even if I set my sky box to output 576, I still think BBC1 HD is a big improvement on BBC1 SD. That's because it's now showing at "full SD" rather than "over compressed SD" Changing the sky box from 576i to 1080i (remember I only have a 720 line tv) only makes a small improvement, nowhere near as much improvement as watching the HD version just to be free of compression artefacts. So on my rather limited tv, I would just be happy if all the SD channels were given enough bandwidth to be largely free of compression artefacts. Granted people with a larger, better tv than me will see a genuine worthwhile improvement with an HD picture. |
|
|
|
|
|
#15 |
|
Forum Member
Join Date: Jan 2002
Posts: 5,187
|
Quote:
So on my rather limited tv, I would just be happy if all the SD channels were given enough bandwidth to be largely free of compression artefacts.
![]() Personally I'd prefer to see decent programme content. If the BBC delivered that I'd be happy even with 405-lines! |
|
|
|
|
|
#16 |
|
Forum Member
Join Date: Jan 2007
Posts: 19,460
|
Quote:
Granted people with a larger, better tv than me will see a genuine worthwhile improvement with an HD picture.
BTW I'd like to restate people should really look at the image 3-5 comparisons as I can tell most people haven't (these show the 1920/1440 difference more starkly)! |
|
|
|
|
|
#17 |
|
Forum Member
Join Date: Apr 2006
Location: North Derbyshire
Posts: 41,783
|
Quote:
BTW I'd like to restate people should really look at the image 3-5 comparisons as I can tell most people haven't (these show the 1920/1440 difference more starkly)!
Anyone who has BBC recordings from before the change can easily see the difference. |
|
|
|
|
|
#18 |
|
Forum Member
Join Date: Feb 2004
Location: London
Posts: 7,514
|
Quote:
I suspect you're just 'barking up the wrong tree', 1440 isn't the problem, it's the excessive compression. Before they crippled the bitrates the 1440 pictures from the BBC were the best of all the HD channels. As soon as they dropped the bitrate massively it became one of the poorest channels.
Anyone who has BBC recordings from before the change can easily see the difference. and then they went VBR .... which meant it looked better. At the end of the day the best test is the eyeball companing the source and destination pictures moving and with the integration that the eye gives ... Sadly there are not many folk who can see this ... and there is so much non BBC /Siemens kit which affects the picture between the HDSDI and your eye .... |
|
|
|
|
|
#19 |
|
Forum Member
Join Date: Jan 2007
Posts: 19,460
|
Quote:
I suspect you're just 'barking up the wrong tree', 1440 isn't the problem, it's the excessive compression. Before they crippled the bitrates the 1440 pictures from the BBC were the best of all the HD channels. As soon as they dropped the bitrate massively it became one of the poorest channels.
Anyone who has BBC recordings from before the change can easily see the difference. |
|
|
|
|
|
#20 |
|
Forum Member
Join Date: Oct 2002
Posts: 9,343
|
Both pictures in image 8 are 1194 x 672 - is this correct?
|
|
|
|
|
|
#21 |
|
Forum Member
Join Date: Jan 2007
Posts: 19,460
|
Quote:
Both pictures in image 8 are 1194 x 672 - is this correct?
|
|
|
|
|
|
#22 |
|
Forum Member
Join Date: Apr 2005
Location: Surrey, UK
Posts: 1,302
|
Quote:
I'm sorry Bob, but you clearly have an interest in claiming there is no difference since all your set top boxes at the moment only really show 1440 HD, except for one ITV1 HD variant on Freesat. Whether it is one frame is irrelevent, the difference exists in every single frame over the whole length of the programme in loss of fine detail and dulling of reflective surfaces.
* each of the products is 'different' meaning that you need to view the output of the real products not some theoretical decoder * looking at off-line de-interlacing is not the same as a real-time system You are forming conclusions which don't actually reflect the reality of the situation. What if some product in the field doesn't like 1920? I don't know to what extent that is the case but there is a balance to be had that isn't reflected in your test. Moving video is the only comparison and it should be tested on each product. |
|
|
|
|
#23 |
|
Forum Member
Join Date: May 2005
Location: Redditch Worcs
Posts: 17,289
|
If a single frame was a valid comparison then a static image grabbed from the 1920 x 1080 output of the mpeg decoder should be markedly inferior from a 1440 x 1920 source rather than a 1920 x 1080 source irrespective of bitrate.
That's not my experience at all http://www.4shared.com/photo/39YN2P-...view00006.html |
|
|
|
|
|
#24 |
|
Forum Member
Join Date: Jan 2007
Posts: 19,460
|
Quote:
I am not saying that our product is perfect, what I am saying is that your demonstration is meaningless because:
* each of the products is 'different' meaning that you need to view the output of the real products not some theoretical decoder * looking at off-line de-interlacing is not the same as a real-time system You are forming conclusions which don't actually reflect the reality of the situation. What if some product in the field doesn't like 1920? I don't know to what extent that is the case but there is a balance to be had that isn't reflected in your test. Moving video is the only comparison and it should be tested on each product. Quote:
If a single frame was a valid comparison then a static image grabbed from the 1920 x 1080 output of the mpeg decoder should be markedly inferior from a 1440 x 1920 source rather than a 1920 x 1080 source irrespective of bitrate.
That's not my experience at all http://www.4shared.com/photo/39YN2P-...view00006.html |
|
|
|
|
|
#25 |
|
Forum Member
Join Date: Apr 2006
Location: North Derbyshire
Posts: 41,783
|
Quote:
Sorry Nigel but you're completely wrong on that, compression has nothing whatsoever to do with it, in fact ITV1 HD Granada is more compressed than ITV1 HD London yet it shows more detail due to the increased resolution. You actually see very few instances of artifacts due to over compression on the BBC HD channels now VBR is enabled, the issue is that they are using 1440 to save bandwidth, but as a consequence this means both BBC HD channels are incapable of showing fine detail and will also lose brilliance and 3D appearance of reflective objects, the image is basically less lifelike.
I've no interest in what the different ITV's are doing, because they are completely different you can't directly compare them. |
|
|
|
![]() |
|
All times are GMT. The time now is 09:32.


.

