logo Sign In

Letterboxed Widescreen vs. Anamorphic Widescreen Discussion — Page 4

Author
Time
Post link
Just trying to squeeze in

Moth3r,

I just looked at the sample vob and it looks very good. IMO the upsampling does not lose much (any) detail because of the already quite low resolution on LDs. I also think that at least all crt users with 16:9 mode will benefit much from anamorphic even if it is STS "faked".

BTW I noticed a linedrop in frame# 1030 in the sample, it's right in the middle of the picture, maybe you already fixed it...

Author
Time
Post link
"Squeeze in", lol.

"I also think that at least all crt users with 16:9 mode will benefit much from anamorphic even if it is STS "faked".

Moreso than an "anamorphic" transfer from NTSC laserdisc, which is little better than a zoomed-in letterbox image - the lowest common denominator of NTSC video.

<span class=“Italics”>MeBeJedi: Sadly, I believe the prequels are beyond repair.
<span class=“Bold”>JediRandy: They’re certainly beyond any repair you’re capable of making.</span></span>

<span class=“Italics”>MeBeJedi: You aren’t one of us.
<span class=“Bold”>Go-Mer-Tonic: I can’t say I find that very disappointing.</span></span>

<span class=“Italics”>JediRandy: I won’t suck as much as a fan edit.</span>

Author
Time
Post link
Originally posted by: eros
changing ld ripped video to anamorphic will degrade picture detail and sharpness. Let me present to you images A and B.

A:
http://img146.exs.cx/img146/4241/a5tx.jpg

B:
http://img146.exs.cx/img146/2779/b9go.jpg

Image A is the raw capture from the laserdisc, borders cropped, with noise reduction and sharpening filters applied.

Image B is the same frame, noise filtered, sharpening filter, upsampled to anamorphic resolution, encoded to MPEG-2 (this frame is taken from the first pass, so is not even the finished encode), borders cropped and finally resized back down to it's original resolution.

Now tell me how much the picture is degraded.

Originally posted by: Warp99
BTW I noticed a linedrop in frame# 1030 in the sample, it's right in the middle of the picture, maybe you already fixed it...
No, I left that in to give the DVD an authentic laserdisc feel...

Seriously, there are literally hundreds of little spots and lines on the capture; some are off the actual film when it was transferred to the laserdisc, others are caused by dust or scratches on the disc surface when I did the capture. I considered using a despot filter to try and remove some of it, but I was worried that it might have a detrimental effect on other aspects of the video, e.g taking out twinkling stars because the filter thinks they're laserdisc defects. (I'm very proud of my beautiful starfield, I even think it looks better than the official retail DVD). And I have neither the time nor the software to go through frame-by-frame and remove the defects manually...

Guidelines for post content and general behaviour: read announcement here

Max. allowable image sizes in signatures: reminder here

Author
Time
Post link
doing the anamorphic conversion without using filters would be a fairer comparison

also talking about resolution differences on dvd is irrelevent because this is always dependent on what type of display your viewing the video on.
Author
Time
Post link
Quote

--------------------------------------------------------------------------------
Originally posted by: Warp99
BTW I noticed a linedrop in frame# 1030 in the sample, it's right in the middle of the picture, maybe you already fixed it...
--------------------------------------------------------------------------------

No, I left that in to give the DVD an authentic laserdisc feel...

Well, it was a whole line(s) that missed so I just thought maybe you wanted to remove them, they do not look very movie like.

But I agree that the different spots and line removing filters sometimes remove more than you want...



Originally posted by: eros
also talking about resolution differences on dvd is irrelevent because this is always dependent on what type of display your viewing the video on.

What displays are you using?

My computermonitor can show 1600x1200, my projector can show 1200x700, I estimate that my pal crt tv can show 640x576.

DVD is 720x576 or 720x480, I estimate the best OT LDs are something like 440x576 or 420x480.

So I guess that my displays (except for maybe my tv) can at least show all the resolution from a LD or todays DVD...


Author
Time
Post link
"doing the anamorphic conversion without using filters would be a fairer comparison "

They've both received the same treatment, so your point is moot.

"also talking about resolution differences on dvd is irrelevent because this is always dependent on what type of display your viewing the video on."

Funny that you didn't mention these details to begin with. I guess your statement should read "changing ld ripped video to anamorphic will degrade picture detail and sharpness [depending] on what type of display your viewing the video on.."?

<span class=“Italics”>MeBeJedi: Sadly, I believe the prequels are beyond repair.
<span class=“Bold”>JediRandy: They’re certainly beyond any repair you’re capable of making.</span></span>

<span class=“Italics”>MeBeJedi: You aren’t one of us.
<span class=“Bold”>Go-Mer-Tonic: I can’t say I find that very disappointing.</span></span>

<span class=“Italics”>JediRandy: I won’t suck as much as a fan edit.</span>

Author
Time
Post link
Originally posted by: Warp99
Well, it was a whole line(s) that missed so I just thought maybe you wanted to remove them, they do not look very movie like.
Now you've mentioned it, I'm going to have to change my script in an attempt to remove them, or they're just gonna bug me every time I see them now...

Guidelines for post content and general behaviour: read announcement here

Max. allowable image sizes in signatures: reminder here

Author
Time
Post link
Now you've mentioned it, I'm going to have to change my script in an attempt to remove them, or they're just gonna bug me every time I see them now...

Sorry, I guess working with old laserdiscs have given me an eye for spots an flaws

It's strange if you have many whole linedrops in your capture, the cld925 is usually very good at filtering them out...


Author
Time
Post link
This is off topic but since it worked very well on the sample vob, anyway, I used this to remove the line:

ConvertToYV12
turnleft()
descratch(minlen=702,asym=15,mindif=3,maxgap=3,blurlen=8,keep=0,border=0,maxangle=0,modeY=3,mark=false)
turnright()

The good thing is that it should not remove anything shorter than 702 pixels so your starfield is quite safe. I could not get a single detection error in the sample anyway, use mark=true for debug mode.

http://bag.hotmail.ru/descratch/descratch.dhtml

Author
Time
Post link
Originally posted by: eros
doing the anamorphic conversion without using filters would be a fairer comparison.


No, it wouldn't, because the statement didn't say anything about filters -- it only said, "changing ld ripped video to anamorphic will degrade picture detail and sharpness," which just isn't true in all cases. As has been proven. QED.


Author
Time
Post link
Originally posted by: MeBeJedi
"doing the anamorphic conversion without using filters would be a fairer comparison "

They've both received the same treatment, so your point is moot.

"also talking about resolution differences on dvd is irrelevent because this is always dependent on what type of display your viewing the video on."

Funny that you didn't mention these details to begin with. I guess your statement should read "changing ld ripped video to anamorphic will degrade picture detail and sharpness [dependending] on what type of display your viewing the video on.."?


detail and resolution are not the same i can have a video of a black and white chequer board and convert it to 1024x768 but it would still be a bunch of black and white squares, no extra detail.
Author
Time
Post link
Originally posted by: Warp99
This is off topic but since it worked very well on the sample vob, anyway, I used this to remove the line:

ConvertToYV12
turnleft()
descratch(minlen=702,asym=15,mindif=3,maxgap=3,blurlen=8,keep=0,border=0,maxangle=0,modeY=3,mark=false)
turnright()

The good thing is that it should not remove anything shorter than 702 pixels so your starfield is quite safe. I could not get a single detection error in the sample anyway, use mark=true for debug mode.

http://bag.hotmail.ru/descratch/descratch.dhtml
Thanks, I'd already identified the descratch plugin but hadn't had a chance to test it. I'll give your vlaues a try.

Guidelines for post content and general behaviour: read announcement here

Max. allowable image sizes in signatures: reminder here

Author
Time
Post link
Greetings, fellows. I'm a long time reader, first time poster. I own the original TR47 set and enjoy reading about the technicalities of laserdisc transfers. I wanted to throw an idea out that pertains to one of the bigger problems with doing laserdisc transfers- low resolution and aliasing artifacts. I imagine some of you have heard of DCDI technology by the Faroudja company. They put motion adaptive scaling chips in displays and in DVD players that intelligently eliminate interlacing and "jagged edge" artifacts. If one was to pass the output from a laserdisc through a display with DCDI and then somehow output the resulting signal to a capture card, would that gain a significant advantage over simply applying various "anti-alias" filters to an already lower-resolution signal as has been done already? I say this because the laserdisc transfer I have seen seems to double up on aliasing flicker- there are interlace/jagged artifacts from the laserdisc source and additional interlace artifacts from the DVD encoding. It can be downright distracting. Hypothetically, I would think somehow harnessing this advanced de-interlacer chip would yield a more refined and smooth picture than we've seen before. There is a decent explanation of this algorithm near the bottom of this page: DCDI I know most Sharp LCD screens make use of this chip, for instance.

Let me know if I'm crazy or maybe onto something. Aside, I'm in awe of all the hard work being done here. Thank you!


Author
Time
Post link
I also have thought about this.

If one would take the laserdisc player and put it in one of these (iscan) for example and take the output into the Capture Card on the computer, would the signal be Progressive? And would it be an improvement over the direct signal into the comp?
  • do or do not, there is no try -
Author
Time
Post link
Sooooooooo, who wants to buy one for us?

<span class=“Italics”>MeBeJedi: Sadly, I believe the prequels are beyond repair.
<span class=“Bold”>JediRandy: They’re certainly beyond any repair you’re capable of making.</span></span>

<span class=“Italics”>MeBeJedi: You aren’t one of us.
<span class=“Bold”>Go-Mer-Tonic: I can’t say I find that very disappointing.</span></span>

<span class=“Italics”>JediRandy: I won’t suck as much as a fan edit.</span>

Author
Time
Post link
How 'bout this instead:

1. Capture original LD signal from high-quality player, as exactly as possible (at least 9-bit ADC, to lossless codec).
2. IVTC, not in real time, with best possible algorithm in AviSynth (free, and testable/tunable) to get full-resolution progressive frames.
3. Scale, not in real time, with best possible scaling algorithm in AviSynth (free, and testable/tunable).

Wouldn't that be a radical idea?
Author
Time
Post link
But what Senator would have the courage to propose such a radical amendment?

<span class=“Italics”>MeBeJedi: Sadly, I believe the prequels are beyond repair.
<span class=“Bold”>JediRandy: They’re certainly beyond any repair you’re capable of making.</span></span>

<span class=“Italics”>MeBeJedi: You aren’t one of us.
<span class=“Bold”>Go-Mer-Tonic: I can’t say I find that very disappointing.</span></span>

<span class=“Italics”>JediRandy: I won’t suck as much as a fan edit.</span>

Author
Time
Post link
You know what, a professionally scaled image does look better than the source.
If any of you live near a really high end AV store, go check out a terranex scaler. We ran laserdisc through a terranex and the results are absolutely mindblowing.

Also, if you have a fixed panel display, (like plasma, LCD, projectors and so on) having an image scaled professionally to a 1:1 pixel match for your display looks much much much better than the job the crummy internal scaler in your fixed pixel display does.

If you took the time to do it offline with the latest technology and then play *that* rsult back - you wet your pants.

Now the argument goes that you can't get back information that isn't there in the first place, but sometimes you can.
If you have a sine wave that has been samples at a low bitrate, you get a blocky representation of a sine wave. If it was audio, it won't sound like the original.

If you take that blocky sinewave, and convert it to splines projected through it, and sample that at a really high bitrate you will end up with a smooth exact match for the original sine wave and it will sound just like the original.

With video it is a lot harder, but some new algorithms presented at Siggraph last year do an absolutely amazing job, and if you have a HD device, a pro scaled image definately looks a LOT better than feeding it 480i or 480P from your DVD player (unless you own a terranex)
Author
Time
Post link
Having read all the above, I am much more informed than I was about DVD technology - thank you. However, I am still slightly confused about DVD regions vs. video standards and I'm sure you guys know all about this, so...

1) Are region 1 DVDs in NTSC resolution and region 2 in PAL resolution (in which case why the new names)?
2) If so, do multi-region players therefore scale between these standards in real-time?

From what I previously knew and what I have learned here, it seems as if:

a) the best possible LD>DVD transfer should start with a PAL LD, regardless of the eventual region/standard of viewing;
b) from there, an anamorphic DVD would be best for all 16:9 TV owners, a letterboxed DVD for most 4:3 owners;

Please correct me if I am wrong. Another question that then arises is:

3) Which is better for NTSC viewers, PAL disc played on multi-region player with player's real-time scaling or NTSC disc made from PAL transfer?

If the answer is, as I suspect, the latter, would this mean that any transfer would ideally exist in 4 versions (PAL AM, PAL LB, NTSC AM, NTSC LB)?

Once again, thank you all for sharing your technical know-how.
Author
Time
Post link
Region number and video system type are two different issues.

Anamorphic is only beneficial on widescreen crt's.
Author
Time
Post link
Does anyone have access to the prints in the Library of Congress?
All three of the complete, true original editions are preserved there (so it's said) in great quality. The Star Wars print is even the "Pre-ANH" version.


Surely gaining access to these would be a simple matter of waiting for a gala to take place at the library, causing an alarm related to the prints to seem to malfunction by using a laser built into a handycam so they would be removed from high security for investigation, gaining access to the gala by posing as a workman with an ID created in photoshop from a blown-up photo of a real workman, getting the librarian in charge's thumbprint on a glass and transferring this to a plastic thumb cover to gain access to the secure elevator?
Author
Time
Post link
Unfortunately, we can't buy souvenir reels of film at the gift shop to replace the real things with, so we'd probably get caught.

That's why I haven't done it, anyway...
Author
Time
Post link
Originally posted by: THX
1) Are region 1 DVDs in NTSC resolution and region 2 in PAL resolution (in which case why the new names)? Most (virtually all?) R1 titles are NTSC, since that's the region code for North America. R2 are split between PAL & NTSC, because a lot of R2 is in Europe (PAL), and at least some more R2 is in Japan (NTSC). As another poster mentioned, these two things are quite independent.

2) If so, do multi-region players therefore scale between these standards in real-time? Typically they do, but that's sort of a lucky fluke. I'd say more players can do PAL <--> NTSC conversions than can be made region-free, so by the time you're able to solve the region-freeness problem, the chances of the format conversion being done are pretty good.

a) the best possible LD>DVD transfer should start with a PAL LD, regardless of the eventual region/standard of viewing;
True, if the following two things are also true:

1) The PAL telecine session is as good as or better than the NTSC session; and
2) A PAL LD player exists that is as good as or better than the best available NTSC one.

I can tell you from direct experience that it is unclear that (1) is true, and that (2) is (unfortunately) almost certainly not true. We're trying, though...

b) from there, an anamorphic DVD would be best for all 16:9 TV owners, a letterboxed DVD for most 4:3 owners;
I'm in the camp that says anamorphic is better for everybody overall, but I guess technically an argument could be made that you're right. I know MBJ -- whose opinion I respect -- would say you're right, for example. I would argue that since everything's moving to 16:9 HDTV eventually, the tiny compromise in picture quality made for those aging 4:3 sets out there is worth it. And let me assure you, this is coming from an owner of a mid-range, not-too-aging 4:3 set. (Of course, I figured out how to access my set's service mode, and now I can make it do native 16:9...)

3) Which is better for NTSC viewers, PAL disc played on multi-region player with player's real-time scaling or NTSC disc made from PAL transfer?

I say an NTSC disc from a PAL transfer. If you play a PAL disc, it'll be 4% too fast (25/23.976 = 1.04, or thereabouts). With a properly-made NTSC-from-PAL effort, the material would benefit from both the better starting resolution of PAL, and the correct speed of NTSC.
Author
Time
Post link
"I would argue that since everything's moving to 16:9 HDTV eventually, the tiny compromise in picture quality made for those aging 4:3 sets out there is worth it."

While I do think both versions are fine for now (since most people still have 4x3 sets), I would agree that the difference is not always noticeable. My neighbor's Sony 4x3 has anamorphic squeeze, so I was able to easily compare the anamorphic and non- transfers on the same size screen. Fine details were fantastic in the anamorphic version, but the overall picture looked almost identical.

Mind you, however, that this was Gladiator, and a true anamorphic transfer. When it comes to making anamorphic DVDs from the LD transfers, we simply don't have the means (see Laserman's post) to really add the extra information to the extent that I would like to make up for the inevitable loss in resolution when downconverted to 4x3. I would much rather watch a 4x3 transfer on my 4x3 tv, because in my mind, it's the best picture I can get at that point in time. (I also don't like unnecessary filtering/downconverting, etc.)

In other words, my preference for 4x3/16x9 really depends on the source material.

That being said, I ran across a really cool PAL->NTSC conversion procedure that I'm dying to try out.

<span class=“Italics”>MeBeJedi: Sadly, I believe the prequels are beyond repair.
<span class=“Bold”>JediRandy: They’re certainly beyond any repair you’re capable of making.</span></span>

<span class=“Italics”>MeBeJedi: You aren’t one of us.
<span class=“Bold”>Go-Mer-Tonic: I can’t say I find that very disappointing.</span></span>

<span class=“Italics”>JediRandy: I won’t suck as much as a fan edit.</span>

Author
Time
Post link


That being said, I ran across a really cool PAL->NTSC conversion procedure that I'm dying to try out.


care to share a link to this procedure? I'd love to play Moth3r's transfer on my denon1910