1080 vs 720, i vs p

[Sareesh Sudhakaran]

Originally Posted by Jonnie Lewis

Most televisions have settings of 1080p and 1080i to select from. If broadcasters are sending signals in 1080i, what is the television doing when I set it to 1080p? Presumably it doesn't do anything and it's reserved for things like BluRay players?

It depends on the panel. My panel changes automatically depending on the input. Some panels upsample SD to make it bearable (mine does). Whatever the case, I have manually set my television to show 1080p60, but obviously it has a mind of its own. When I play games, it shows 1080p60. When I watch PAL DVDs, it upsamples SD and displays 720p50.

Similarly, my friend has his PC hooked up to his TV for use as a monitor. Despite living in the UK, I noticed that his TV was displaying in 60HZ. I'm assuming this is because PCs work differently from broadcast signals and it probably has a frame rate of 30fps+?

This is similar to what I have, except my computer is an HTPC. Same thing, really. Modern panels mostly go up to 60p, and the higher end ones go to 120 Hz (120p) - which 3D gamers use.

I think I can feel my brain melting - does anyone know of any good books that I could pick up regarding the different video formats and their uses? I think it'd be nice to have as a reference - reading posts across various different forums often gets me more confused! So much terminology...

Thanks again.

You can start by reading my blog! To answer your previous questions:

I've read that TV stations only broadcast in interlaced, so what happens if I give them a 1080p file?

You can't. The BBC, e.g., only officially accepts interlaced masters. Of course, if they wanted it, they will accept progressive. I've read they are going to shoot their new mega-series in 4K, for which there is no interlacing standard.

Where is it decided whether an image is interlaced or progressive?

It can be 'decided' in many places. The only way to know for sure is to study each signal as it passes a device for changes. Professional devices are used because they claim to tell you like it is, though in reality some compromises are always made.

Also, how does the resolution relate to the size of a screen? If I was to buy a 60" TV, surely a 1920x1080 image would have to be stretched up to fit it and it would look crap?

Yes, it would. At the same resolution, the bigger the TV, the further back you have to go, but after a point you lose the advantage of resolution. The same happens when you get closer. Every size-resolution combination has a zone of best 'viewability'. You can read What is 4K Television? | wolfcrow for starters. Then I recommend reading the THX specifications for cinema, which is written lucidly enough. A little trigonometry, and you'll be gold.

Would there be any adverse effect on the quality of the footage if I was to film in 1080p, and place it into a 720p sequence within Premiere Pro? It would obviously have to be shrunk - would that have a negative effect on image quality?

That depends on what you call 'image quality'. Will it make that beautiful flower worse? No, but it will lose some resolution. Most people wouldn't notice, simply because most people don't sit at the right distance from their television sets. This is one reason why standard definition is dying a slow death (among other important factors like lack of bandwidth, money, etc).

Also, 1080p is compressed more than 720p, even on the web. This 'extra' compression reduces perceptible resolution. It's all about perception.

If so... How is it that 1080p looks just as good on a 42" television as it does on a 27" television? Surely the 27" television has to shrink the image to fit it on the screen?

No, the resolution is the same. Each pixel is a box. The number of boxes remain the same. The size of the box changes. In fact, all things being equal, a 27" panel at 1080p has a higher resolution than a 40" panel at 1080p, and is a better television. Simple test, walk into an electronics store and view them side by side. See which ones let you get closer, and at what point they look the same.

No rocket science, just basic trigonometry. Hope this helps.

[Jonnie Lewis]

That was a wonderful answer, Sareesh. Really appreciate it!

I'll check out your blog now and add it to my RSS feeds :)

[David Heath]

Originally Posted by Ervin Farkas

Here's a solid article to document my point: on your average TV, there is no visible difference between 720P and 1080i.

Read the numbers.

Why 4K TVs are stupid | TV and Home Theater - CNET Reviews

A short quote: "A 720p, 50-inch TV has pixels roughly 0.034 inch wide. As in, at a distance of 10 feet, even 720p TVs have pixels too small for your eye to see. That's right, at 10 feet, your eye can't resolve the difference between otherwise identical 1080p and 720p televisions."

Read the entire article... fascinating numbers about the physics of the human eye.

Sorry, but those numbers are incorrect, and whilst much of the rest of his article is correct, that error leads him to completely false conclusions.

It's easy enough to show. For the 720p, 50" TV he refers to, that will be 50" across the diagonal, so for 16:9 aspect ratio, that gives a height of 30" and a width of 40". (By Pythagorus theorem.) Now if it's 720p, there must be 720 pixels vertically in those 30", or 24 per inch, which equates to a pixel spacing of 0.042" - NOT 0.034".

I agree with his other figures about the limiting resolution of a typical human eye (1 minute of arc, or 0.035" at 10 foot viewing distance), but that means the eye is capable of resolving more detail than a 720p screen is capable of giving at that distance. Redo the figures with 1080p, and 1080 in 30" means 36 per inch, or a spacing of 0.028". So better than human vision.

And that's the maths and science behind why 720p is not good enough for 50" screens at normal viewing distances, whilst 1080p is generally regarded as OK up to about 60".

In ball-park terms, at 10 foot viewing, 720p is normally regarded as good enough up to about 37-40" screens, but you'll see an improvement with 1080 with bigger screens. 1080p should be good enough up to about 50-55" - bigger than that, and there will be a noticeable difference with 4k.

Practically, it's better to oversample, rather than be right on the resolution limit, for reasons so 4k may start paying dividends even around the 50" mark. Certainly the CNET conclusion (".....at 10 feet, your eye can't resolve the difference between otherwise identical 1080p and 720p televisions") is simply not true.

[Alister Chapman]

Originally Posted by Sareesh Sudhakaran

You can't. The BBC, e.g., only officially accepts interlaced masters. Of course, if they wanted it, they will accept progressive. I've read they are going to shoot their new mega-series in 4K, for which there is no interlacing standard.

That's a little misleading. The BBC only accepts masters that are recorded/ encoded using an interlaced signal structure. That includes both conventional interlace and PsF. PsF is Progressive wrapped in an interlace frame. PsF is progressive. A huge percentage of BBC HD production is progressive, the same for Discovery and Nat Geo etc. All of the BBC, Nat Geo and Discovery productions that I've worked on in the last 3 years have been progressive.
As web delivery of content by broadcasters becomes more and more significant (and in the future may well overtake traditional broadcasting over the air) less and less interlace will be used.
The majority of HD content is watched on progressive display devices that have to use tricks like bob-deinterlacing. Many productions are edited on computer systems that only have progressive computer monitors for display. I think that interlace is something that will gradually disappear.

[Sareesh Sudhakaran]

Originally Posted by Alister Chapman

That's a little misleading. The BBC only accepts masters that are recorded/ encoded using an interlaced signal structure.

From the horses mouth (http://www.bbc.co.uk/guidelines/dq/p...rdsBBCv3.pdf):

All material delivered for UK HD TV transmission must be:
1920 x 1080 pixels in an aspect ratio of 16:9
25 frames per second (50 fields) interlaced - now known as 1080i/25.
colour sub-sampled at a ratio of 4:2:2

The HD format is fully specified in ITU-R BT.709-5 Part 2.

2.1.1 Origination

Material may be originated with either interlaced or progressive scan.
Interlaced and progressive scan material may be mixed within a programme if it is required for editorial
reasons or the nature of the programme requires material from varied sources.

2.1.2 Post-production

Electronically generated moving graphics and effects (such as rollers, DVE moves, wipes, fades and
dissolves) must be generated and added as interlaced to prevent unacceptable judder.

2.1.3 Film motion or ‘film effect’

It is not acceptable to shoot in 1080i/25 and add a film motion effect in post-production. Most High
Definition cameras can capture in either 1080i/25 or 1080p/25. Where film motion is a requirement,
progressive capture is the only acceptable method.

You can shoot progressive or interlaced, but only interlaced masters are officially accepted.

I agree with you about intelacing disappearing. The UHDTV standards make no room for it.

[Ron Evans]

I wonder when we will actually move to true1920x1080 50/60P since almost all the consumer cameras can now do this as well as my latest GoPro !!! There is a difference and it would make display a lot easier.

Ron Evans

[Alister Chapman]

Sareesh, Paragraph 2:1:1 Material may be originated with either progressive or interlace scan.

If you shoot progressive and place it in an interlaced stream what do you get? PsF. An interlaced master that contains progressive material is PsF, the content on the tape is still progressive, it hasn't become interlaced, its still progressive, only the single progressive frame is now split into odd and even lines. The temporal motion of the footage is still the same, there is no temporal difference between the fields. The delivery master may well be a tape or file that has 50 fields per second, but if the two fields contain odd and then even lines from material that was originated progressively, then this is PsF.

Most BBC documentary production is progressive, almost all drama is progressive.

[Sareesh Sudhakaran]

Originally Posted by Alister Chapman

Sareesh, Paragraph 2:1:1 Material may be originated with either progressive or interlace scan.

If you shoot progressive and place it in an interlaced stream what do you get? PsF. An interlaced master that contains progressive material is PsF, the content on the tape is still progressive, it hasn't become interlaced, its still progressive, only the single progressive frame is now split into odd and even lines. The temporal motion of the footage is still the same, there is no temporal difference between the fields. The delivery master may well be a tape or file that has 50 fields per second, but if the two fields contain odd and then even lines from material that was originated progressively, then this is PsF.

Most BBC documentary production is progressive, almost all drama is progressive.

Agreed! Are we disagreeing or agreeing? :)

[Alister Chapman]

Agreeing, if we both agree that the BBC do accept productions that are shot, edit and produced in progressive. It is only the delivery file or tape that must use PsF. As PsF is still progressive, then the BBC do accept progressive files, you just can't send them a straight from the camera 25p frame only file.

To me your original statement sounded like you we're saying the BBC don't accept progressive programmes.

[Sareesh Sudhakaran]

Not at all. We agree 100%!

[Petter Flink]

Originally Posted by David Heath

Sorry, but those numbers are incorrect, and whilst much of the rest of his article is correct, that error leads him to completely false conclusions.

It's easy enough to show. For the 720p, 50" TV he refers to, that will be 50" across the diagonal, so for 16:9 aspect ratio, that gives a height of 30" and a width of 40". (By Pythagorus theorem.) Now if it's 720p, there must be 720 pixels vertically in those 30", or 24 per inch, which equates to a pixel spacing of 0.042" - NOT 0.034".

A small mistake in your calc.
A 16:9 ratio display the sides can not be 3:4 with a hypotenuse of 5.
@16:9 ratio the sides will be 24.5" and 43.6" with a diagonal of 50".

And 720 pixels over 24,5" gives a pixel size of 0.034"

I agree with his other figures about the limiting resolution of a typical human eye (1 minute of arc, or 0.035" at 10 foot viewing distance), but that means the eye is capable of resolving more detail than a 720p screen is capable of giving at that distance. Redo the figures with 1080p, and 1080 in 30" means 36 per inch, or a spacing of 0.028". So better than human vision.

And that's the maths and science behind why 720p is not good enough for 50" screens at normal viewing distances, whilst 1080p is generally regarded as OK up to about 60".

In ball-park terms, at 10 foot viewing, 720p is normally regarded as good enough up to about 37-40" screens, but you'll see an improvement with 1080 with bigger screens. 1080p should be good enough up to about 50-55" - bigger than that, and there will be a noticeable difference with 4k.

Practically, it's better to oversample, rather than be right on the resolution limit, for reasons so 4k may start paying dividends even around the 50" mark. Certainly the CNET conclusion (".....at 10 feet, your eye can't resolve the difference between otherwise identical 1080p and 720p televisions") is simply not true.

[David Heath]

Originally Posted by Petter Flink

A small mistake in your calc.
A 16:9 ratio display the sides can not be 3:4 with a hypotenuse of 5.
@16:9 ratio the sides will be 24.5" and 43.6" with a diagonal of 50".

And 720 pixels over 24,5" gives a pixel size of 0.034"

Hmmm. The only thing I can disagree with there is the word "small"! :-) Having had it pointed out, I can only say I'm guilty as charged, and it's a pretty big mistake by me! My apologies to all.

If I have any excuse, it's that my (incorrect) working seemed to agree with conclusions that have been generally found in practice, so maybe I didn't check the figures as closely as if they had seemed to show something surprising. As illustration, the best research I've come across was done a while back by BBC R&D and is online at http://downloads.bbc.co.uk/rd/pubs/w...les/WHP092.pdf .

Then, they came to the conclusion that 720p should be good enough for general broadcast - but were basing that on the assumption that screen sizes for the home were likely to be around 37" typically, up to about 42" max. In practice, that was very pessimistic, home screens are often larger than 42" now and the conclusion was out of date almost before the document was published (though not the test results).

Look at figure 9 and it shows that up to a 40" screen only 15% of observers could see a benefit to 1080 over 720. That increases to nearly 30% for a screen size of 42" and nearly 50% for a 50" screen.

[Alister Chapman]

I was lucky enough to spend a weekend using an 84" 4K Sony Bravia to show and demo some 4K footage and HD footage.

Up close (within 4m) the difference between the 4K and HD was clear to see, but further away the difference became harder to see. 4m is pretty close to such a big screen.

However, even when at the back of the room, about 8m to 10m away there was something incredibly "real" about the 4K image that wasn't there with HD material. I would not say the footage looked sharper, or that I could see more detail, but many in the room could sense something different about the footage when it was shown as 4K compared to HD. It really was like looking out of a window. Perhaps it was just the scale, but I've never had such a distinct sensation of looking out of a window with 4K projection. I get to play with the TV again next week at CabSat.