What is the real visible resolution of a Blu-ray video? (800 * 700?)

[Stears555]

Hello!

I've searched many test pattern which are typically located on most Blu-ray disc. I've uploaded one Test pattern file, but it is only in mpeg 2 on all of my blu-rays.

As you can see it, it had only around 800*700 resolution

Have you got a h264 version of blu-ray test image? Can you upload it?


Thank you for your reply!

[LMotlow]

The encoded resolution of the test pattern m2ts you posted is 1920x1080.
Many commercial BluRays are encoded with the HD version of MPEG2. What difference would h264 make? MPEG and h264 aren't the only tech used for BluRay encoding.

Allowed BluRay resolutions, encoders, and other stuff are listed here: https://www.videohelp.com/hd#tech

[Brad]

You uploaded color bars, not a resolution test pattern. And it isn't aligned on chroma boundaries, so it's softer than it could be.

The real visible resolution possible with Blu-ray is the same as what the spec says: 1920x1080 for luma and 960x540 for chroma. The free AVS HD 709 download includes resolution patterns.

[kieranjol]

You uploaded color bars, not a resolution test pattern. And it isn't aligned on chroma boundaries, so it's softer than it could be.

The real visible resolution possible with Blu-ray is the same as what the spec says: 1920x1080 for luma and 960x540 for chroma. The free AVS HD 709 download includes resolution patterns.

The colour bars are only for the first 15 seconds. I don't understand where the OP is getting 800x700 from. How is that clip not 1920*1080?

[poisondeathray]

I don't understand where the OP is getting 800x700 from. How is that clip not 1920*1080?

He's implying the concept of measureable "lines of resolution" , one measure of "detail", such as LP/PH

A given frame size does not necessarily detail. e.g. If I upscale an SD image to 1920x1080, it doesn't mean it has 1920x1080 lines of actual visible detail - it will look very soft compared to one that has been downscaled properly from UHD

Test patterns like lines, zone plates - they are low level measures to indicate maximum theoretical resolution. When a pattern shows aliasing, or no more lines are clearly distingishable, that is the mark where the resolution is said to be in one axis (horizontal, vertical)

But the issue is more complicated. It also depends on the source characteristics (eg. quality, compression, grain, noise, etc...), processing and hardware, and human perception of content (you don't watch static test patterns usually, and in motion its difficult to measure or perceive detail)

As vaporeon800 said, the maximum is 1920x1080 for Y and 960x540 for CbCr - but all those other factors may work to reduce the actual visible resolution. Also a typical "movie" won't be 16:9 , but some wider AR like 2:35:1. So the maxiumum greyscale height for a 2:35:1 movie might only be ~800, not 1080, and all those other factors go towards reducing the resolution farther

You can buy a test pattern disc (e.g. Spears and Munsil or DVE) and evaluate it. Some player setups have higher chroma resolution than others because of the scaling algorithm used

[kieranjol]

I don't understand where the OP is getting 800x700 from. How is that clip not 1920*1080?

He's implying the concept of measureable "lines of resolution" , one measure of "detail", such as LP/PH

A given frame size does not necessarily detail. e.g. If I upscale an SD image to 1920x1080, it doesn't mean it has 1920x1080 lines of actual visible detail - it will look very soft compared to one that has been downscaled properly from UHD

Test patterns like lines, zone plates - they are low level measures to indicate maximum theoretical resolution. When a pattern shows aliasing, or no more lines are clearly distingishable, that is the mark where the resolution is said to be in one axis (horizontal, vertical)

But the issue is more complicated. It also depends on the source characteristics (eg. quality, compression, grain, noise, etc...), processing and hardware, and human perception of content (you don't watch static test patterns usually, and in motion its difficult to measure or perceive detail)

As vaporeon800 said, the maximum is 1920x1080 for Y and 960x540 for CbCr - but all those other factors may work to reduce the actual visible resolution. Also a typical "movie" won't be 16:9 , but some wider AR like 2:35:1. So the maxiumum greyscale height for a 2:35:1 movie might only be ~800, not 1080, and all those other factors go towards reducing the resolution farther

You can buy a test pattern disc (e.g. Spears and Munsil or DVE) and evaluate it. Some player setups have higher chroma resolution than others because of the scaling algorithm used

Thanks for clarifying. I need to learn how to interpret those test patterns. Looks really useful.

[Brad]

The colour bars are only for the first 15 seconds.

Oh. I'm so used to test pattern discs that split each one into separate video files, I didn't even bother seeking ahead.

Pattern 2 is aliased as hell. That isn't what a proper resolution pattern on Blu-ray looks like, Stears555.



AVS HD 709 overscan test with little lines and dots:

[newpball]

Pattern 2 is aliased as hell. That isn't what a proper resolution pattern on Blu-ray looks like, Stears555.

Brilliant engineering at work.

[Stears555]

The encoded resolution of the test pattern m2ts you posted is 1920x1080.
Many commercial BluRays are encoded with the HD version of MPEG2. What difference would h264 make? MPEG and h264 aren't the only tech used for BluRay encoding.

Allowed BluRay resolutions, encoders, and other stuff are listed here: https://www.videohelp.com/hd#tech

All of my blu-ray disc were encoded by H264, only the test patterns and advertisings, movie trailers were encoded in Mpeg 2 .

[Stears555]

You uploaded color bars, not a resolution test pattern. And it isn't aligned on chroma boundaries, so it's softer than it could be.

The real visible resolution possible with Blu-ray is the same as what the spec says: 1920x1080 for luma and 960x540 for chroma. The free AVS HD 709 download includes resolution patterns.

It is not true. The real visible resulotion is only around the half of the nominal resolution in every video what I've ever seen.

[Brad]

I already posted an image that disproves your statement. AVS HD 709's overscan pattern (among others) includes alternating single-pixel grey and black dots.

[Stears555]

I already posted an image that disproves your statement. AVS HD 709's overscan pattern (among others) includes alternating single-pixel grey and black dots.

You confused the real visible resolution with the nominal resolution. Please upload a test pattern video (it is only around 100-150Mbyte) from a real blu-ray disc.

What you tried to say is mathemaically impossible.
Do you know how do itself the 4:2:0 yuv conversion of the color space reduce/degrade the real resoltion? And after the color space conversion the codec compressor degrade the resolution even further...

[newpball]

Do you know how do itself the 4:2:0 yuv conversion of the color space reduce/degrade the real resoltion? And after the color space conversion the codec compressor degrade the resolution even further...

Sssst, some like to keep up appearances: SD video is already great while HD is the icing on the cake!

[Brad]

No, the visible resolution for a monochrome image really is 1,920 pixels across and 1,080 pixels down. It's obvious if you view a full-frame single-pixel checkerboard pattern, and should be obvious by inference from the smaller grid shown in the overscan test above.

Are you saying AVS HD 709 isn't a "real" Blu-ray Disc? Why?

We already mentioned the resolution of the chroma channels. I refer you to a colorless grid and you respond that color subsampling reduces the resolution of it. Do you see the problem here?

In real images, compression reduces the visible resolution, yes. Not in properly-encoded static test patterns.

[Stears555]

The colour bars are only for the first 15 seconds.

Oh. I'm so used to test pattern discs that split each one into separate video files, I didn't even bother seeking ahead.

Pattern 2 is aliased as hell. That isn't what a proper resolution pattern on Blu-ray looks like, Stears555.


[Attachment 31757 - Click to enlarge]

AVS HD 709 overscan test with little lines and dots:


[Attachment 31756 - Click to enlarge]

It is clearly visible, that your test pattern was not designed for HD resolution. It is only a 16:9 test up to 600 lw/ph. Use only official imatest patterns, with real geometric forms designed to 1080p video.

[Stears555]

The colour bars are only for the first 15 seconds.

Oh. I'm so used to test pattern discs that split each one into separate video files, I didn't even bother seeking ahead.

Pattern 2 is aliased as hell. That isn't what a proper resolution pattern on Blu-ray looks like, Stears555.


[Attachment 31757 - Click to enlarge]

AVS HD 709 overscan test with little lines and dots:


[Attachment 31756 - Click to enlarge]

It is clearly visible, that your test pattern is not designed for HD resolution. It is only a 16:9 test up to 600 lw/ph. (600 lw/ph is good only for SD videos)

Use only official imatest patterns, with real geometric forms designed to 1080p video (1900 to 2000 lw/ph resolution test lines.)

[Brad]

I wasn't talking about those areas of the image. Look at the box where the text "AVS HD 709" is.

[Stears555]

I wasn't talking about those areas of the image. Look at the box where the text "AVS HD 709" is.

IT does not contain high resolution (2000 lw/ph) test lines and curves.

[hello_hello]

Sssst, some like to keep up appearances: SD video is already great while HD is the icing on the cake!

You off your medication again?

[Stears555]

Sssst, some like to keep up appearances: SD video is already great while HD is the icing on the cake!

You off your medication again?

I've just spoke about facts. The guy represented only a 600 lw / ph test pattern. Deal with it.

[kieranjol]

Sssst, some like to keep up appearances: SD video is already great while HD is the icing on the cake!

You off your medication again?

I've just spoke about facts. The guy represented only a 600 lw / ph test pattern. Deal with it.

Pretty sure that his meds comment was directed at newpball, not you Stears555.

[pandy]

I've just spoke about facts. The guy represented only a 600 lw / ph test pattern. Deal with it.

Ok - then straight to facts - please tell me how to display full res hd sinewave where sinewave have Nyquist frequency and phase 180deg...

[newpball]

Sssst, some like to keep up appearances: SD video is already great while HD is the icing on the cake!

You off your medication again?

I've just spoke about facts. The guy represented only a 600 lw / ph test pattern. Deal with it.

Pretty sure that his meds comment was directed at newpball, not you Stears555.

Likely, as it is not the first time.

Some people think that making personal attacks at fellow members is OK.
To me that shows more about the person making the attacks than the about the one who is attacked.

[_Al_]

you destroy every single thread you step in,.., and talking and crying about attacking now ...

[jagabo]

Here's a better quality resolution test pattern. Note that the numeric labels don't directly translate to # of lines per width/height of the image. Around 11 corresponds to 1920x1080. The original image was a ~4K JPG, pretty clean, and was already cropped. I included the small text and discreet line patterns from Vapereon's image as an inset near the top center. The second half of the video has the same pattern reduced to half resolution and transferred to the chroma channels, with a flat grey luma channel. So you can plainly see the effect of 4:2:0 chroma subsampling.

[hello_hello]

Some people think that making personal attacks at fellow members is OK.
To me that shows more about the person making the attacks than the about the one who is attacked.

Have a look in the mirror. You even attacked others in this thread. Not directly, but you were having a "general" dig with your pointless post regarding keeping up appearances. I've read several posts of yours where you've made derogatory comments about "other posters preferences" in respect to video quality, but I guess none of the other posters here are supposed to take that as a "personal attack". What's really bizarre though, is those claims are generally not true. Helping someone with something they're trying to achieve doesn't define your own preference, nor does helping them without first giving them a hard time for asking.

In the recent thread on cropping you replied to my post where I said I often zoom in on video simply to say you weren't surprised. Was that supposed to be complimentary?

I asked if you were off your meds again because your rant in this thread was a little more irrelevant than usual. What's anyone's opinion on SD video quality got to do with the topic here?

[-Habanero-]

Some people think that making personal attacks at fellow members is OK.
To me that shows more about the person making the attacks than the about the one who is attacked.

Jesus christ, do you do anything but whine? Go find a Moron Preservation League to play the victim in.

As to the topic, do I understand that this test pattern is used to find the true, entropic resolution in a given (upscaled) video? I had no idea of the existence of such. Are they more accurate than doing an animated downscale, subtracting the original image with it and finding the point where details start to appear? That's what I do.
Most Blu-rays are rarely at a true 1080p resolution unless they've been mastered at 4K and some are even upscaled SD garbage.

[hello_hello]

Are they more accurate than doing an animated downscale, subtracting the original image with it and finding the point where details start to appear? That's what I do.

Could you explain that process? It sounds interesting. Something I might want to try. Cheers.

[jagabo]

I think he means something like this:

# downscale to indicated percent then upscale back to original res
function DownUpScale(clip v, int percent)
{
  # be sure to keep mod2
  width = (v.width * percent / 200) * 2
  height = (v.height * percent / 200) * 2
  v.BicubicResize(width, height).BicubicResize(v.width, v.height)
  Subtitle("width="+String(width)+" height="+String(height))
}

ffVideoSource("video.mp4") 
Subtract(last, Animate(1, 100, "DownUpScale", last,100, last,1))

A resolution test pattern will show differences even at 99 percent.

[hello_hello]

I'll give it a spin. Cheers.