Endless Paradigm

I made a Blu-ray comparison for work to showcase the resolution difference from DVD to Blu-ray.

Both images are taken from the same trailer, one 480p (digitally upscaled to 1080p) and one 1080p.

I have not modified the images in anyway.



I Encourage Firefox users to open each image in a different tab and switch between them.


Babylon A.D.

DVD (480p)
Click to download the Original .TIF File (4.45mb)


Blu-ray (1080p)
Click to download the Original .TIF File (4.45mb)





Max Payne

DVD (480p)
Download the Original .TIF file (4.45mb)


Blu-ray (1080p)
Click to download the Original .TIF File (4.45mb)






Pineapple Express

DVD (480p)
Download the Original .TIF file (4.45mb)


Blu-ray (1080p)
Click to download the Original .TIF File (4.45mb)

I thought it would be a greater difference, considering the size difference, but i guess on a good TV or monitor you could notice the difference. I just watch on the lappy, so i just download the 720p version because the regular version is like 624×352, so it looks like spoon full screen. Anyway, thanks for the comparison =p

blu-ray for the win!

you can't make out 100% difference if your monitor does not support high resolutions. Yet, the difference is still apparent in form of clarity and sharpness.

Well, it's not surprising the native HD 1080p will look much sharper than the upsized.

I always choose to download the HD stuff (720p generally) when I have the option, unless it's absolutely massive (like 8GB for a movie = learn to encode).

Kuu Wrote:Max Payne

Blu-ray (1080p)
Click to download the Original .TIF File (4.45mb)

The 1080p is a lie. It's 1080i.

What Deinterlace/IVTC filter did you use? Way too many combing artifacts.

Anyways, you probably want to resize the 1440x1080 to 1920x1080 (if it's 1440x1080 on the BluRay, that's because it's anamorphic). Because then if you crop the top and bottom black boarders, you get the correct movie aspect of 2.39:1.

^ Too technical for me @_@

All I did was download the 1080p and 480p trailers from Apple and Yahoo, opened with Adobe Premiere, HDV 24p

General
Editing mode: MainConcept MPEG Pro
Timebase: 23.976 fps

Video Settings
Frame size: 1440h 1080v (1.333)
Frame rate: 23.976 frames/second
Pixel Aspect Ratio: HD Anamorphic 1080 (1.333)
Fields: No Fields (Progressive Scan)

Audio Settings
Sample rate: 48000 samples/second

Capture Format
MainConcept MPEG Pro

Video Rendering
Maximum Bit Depth: Off
Preview File Format: MainConcept MPEG Video
Compressor: MainConcept MPEG Video
Color depth: Millions of colors

Default Sequence
Total video tracks: 3
Master track type: Stereo
Mono tracks: 0
Stereo tracks: 3
5.1 tracks: 0
Submix mono tracks: 0
Submix stereo tracks: 0
Submix 5.1 tracks: 0



Then I frame capped it, exported as .tif. Opened it in Photoshop and Saved As .PNG.

---

This is what Premiere says it is:

Type: QuickTime Movie
File Size: 153.9 MB
Image Size: 1920 x 800
Pixel Depth: 1920
Frame Rate: 23.976
Source Audio Format: 44100 Hz - 16 bit - Stereo
Project Audio Format: 48000 Hz - 32 bit floating point - Stereo
Total Duration: 00:02:22:01
Average Data Rate: 1.1 MB / second
Pixel Aspect Ratio: 1.0

Kuu Wrote:^ Too technical for me @_@

All I did was download the 1080p and 480p trailers from Apple and Yahoo, opened with Adobe Premiere, HDV 24p

General
Editing mode: MainConcept MPEG Pro
Timebase: 23.976 fps

Video Settings
Frame size: 1440h 1080v (1.333)
Frame rate: 23.976 frames/second
Pixel Aspect Ratio: HD Anamorphic 1080 (1.333)
Fields: No Fields (Progressive Scan)

Audio Settings
Sample rate: 48000 samples/second

Capture Format
MainConcept MPEG Pro

Video Rendering
Maximum Bit Depth: Off
Preview File Format: MainConcept MPEG Video
Compressor: MainConcept MPEG Video
Color depth: Millions of colors

Default Sequence
Total video tracks: 3
Master track type: Stereo
Mono tracks: 0
Stereo tracks: 3
5.1 tracks: 0
Submix mono tracks: 0
Submix stereo tracks: 0
Submix 5.1 tracks: 0



Then I frame capped it, exported as .tif. Opened it in Photoshop and Saved As .PNG.

I see.

As for what I was talking about, well, if you look closely at the image, for example the place I circled, you can see slight horizontal line artifacts (combing artifacts). That's caused by bad de-interlacing (converting 1080i to 1080p) not removing all the interlacing. If you don't remember deinterlacing it yourself, then premier probably auto-deinterlaced it for you?



EDIT:

Kuu Wrote:Video Settings
Frame size: 1440h 1080v (1.333)
Frame rate: 23.976 frames/second
Pixel Aspect Ratio: HD Anamorphic 1080 (1.333)
Fields: No Fields (Progressive Scan)

Lolwut..?

It says it's fieldless (progressive)... Does that mean the source came like that? If so then shame on Apple/Yahoo/Whatever you got the source from.

Ok, look at the source directly. if it has lots of lines across it like this
then it's interlaced (1080i). If it's clean, then it's progressive.

---

Kuu Wrote:This is what Premiere says it is:

Type: QuickTime Movie
File Size: 153.9 MB
Image Size: 1920 x 800
Pixel Depth: 1920
Frame Rate: 23.976
Source Audio Format: 44100 Hz - 16 bit - Stereo
Project Audio Format: 48000 Hz - 32 bit floating point - Stereo
Total Duration: 00:02:22:01
Average Data Rate: 1.1 MB / second
Pixel Aspect Ratio: 1.0

This is the second part of what I was talking about previously. The correct frame size for that picture should be 1920x1080, rather than 1440x1080. And after you crop top and bottom it becomes 1920x800, which is the correct aspect ratio for the movie. But that's not important for comparison purposes (is only important if you're going to encode it), so I'll shut up about it.

Assassinator is EP's encoding Guro

You've got used to this s^&% haven't you Assassinator haha.

Yeah the comparison is quite clear.. I want to see how Toshiba are claiming their Upscaling TV magics and how its better than the standard upscale...

Correct me if I'm wrong but standard upscaling is Nearest Neighbor..?

Senseito URΩBΩROS Wrote:I want to see how Toshiba are claiming their Upscaling TV magics and how its better than the standard upscale...

Correct me if I'm wrong but standard upscaling is Nearest Neighbor..?

Typically, I find most hardware scalers use bilinear (pretty much a blur) or sometimes, bicubic.  Nearest neighbour/pixel resize gives horrible jagged edges most of the time.

Better resampling techniques include Lanczos and Spline64.  But the type of material does also affect which is really "better" (sometimes, I find a bilinear looks better than a Lanczos/bicubic).
As for claims, well, you could really claim anything there.