logo Sign In

Disney's Beauty and the Beast [spoRv] BD-25 (Released) — Page 2

Author
Time

LOL, hardly!  I actually know next to nothing about this stuff.

"Right now the coffees are doing their final work." (Airi, Masked Rider Den-o episode 1)

Author
Time

So, are you only a Pokemon expert?!? (^^,)

Sadly my projects are lost due to an HDD crash… 😦 | [Fundamental Collection] thread | blog.spoRv.com | fan preservation forum: fanres.com

Author
Time

I'm not an expert on anything :P

I'm just good at recall of trivia.

"Right now the coffees are doing their final work." (Airi, Masked Rider Den-o episode 1)

Author
Time

Molly, I believe the MKV version is not a re-encoding but the mpeg stream repackaged in an MKV container.

I just got to looking at the test4-analoclean+magicup2 version and don't like it as well as the first version.  The original preserved some grain and looks more natural with less digital artifacts.

How do those versions relate to the ones tagged as deghost?  Is that just the analoclean+magicup and analoclean+magicup2 again or something new?

Dr. M

Author
Time

... there's something that I can't understand... before, you wrote:

Doctor M said:

My thoughts on the sample clips.

Best is the Analoclean with magicup.  The Magiclean turns the noise into a frozen noise pattern that looks like your watching it through a screen door.

...

Seriously though, if the analoclean+magicup was your final release I'd be pretty damn happy with it.  It's really clean but looks like there is still some grain.

then now you wrote:

Doctor M said:

I just got to looking at the test4-analoclean+magicup2 version and don't like it as well as the first version.  The original preserved some grain and looks more natural with less digital artifacts.

Could you explain me, please?!?

How do those versions relate to the ones tagged as deghost?  Is that just the analoclean+magicup and analoclean+magicup2 again or something new?

analoclean+magicup and less grain = analoclean+magicup2

the first two attempts uses the same analoclean+magicup2 scripts, plus two different approaches at deghost; the last one - laserclean+magicup, is a slightly modified analoclean script, incorporating a third deghosting method, that seems to work better than the previous two. Note: I decided to change the name of the script from laserclean to haloclean, because when a laserdisc capture has no halos, it erase a lot of details, so the previously-called laserclean script actually cleans laser image a bit too much... (^^,)

Sadly my projects are lost due to an HDD crash… 😦 | [Fundamental Collection] thread | blog.spoRv.com | fan preservation forum: fanres.com

Author
Time

Sorry if I wasn't clear before.  What I meant was the magiclean froze the grain in clumps and it made the picture look like you were viewing it through a dirty screen.

The analoclean didn't produce the same effect.  It left grain as grain, that is small moving particles.  That is more desirable.

The analoclean+magicup2 seemed to scrub all grain, and I'm just not a fan of that.

Can you calarify, is this correct:

generic deghost script 1 + AnaloClean + MagicUp = analoclean+magicup(1)
generic deghost script 2 + AnaloClean + MagicUp = analoclean+magicup2

Or is the deghost a new addition to the previous scripts?

 

Dr. M

Author
Time

Doctor M said:

generic deghost script 1 + AnaloClean + MagicUp = analoclean+magicup(1)
generic deghost script 2 + AnaloClean + MagicUp = analoclean+magicup2

Or is the deghost a new addition to the previous scripts?

no, it's an addition to the previous script; I managed to clean up the grain, as you asked into your post; well, I could also manage to leave it as before, or reduce it a bit - as in the last scripts they are all cleaned up.

Now I'm waiting your opinions about the last three deghosted test clip; about grain, we'll think about it later, ok?

...sometimes is quite difficult to understand each other, if there are an ocean, a language and 7000km+ between us, right? (^^,)

Sadly my projects are lost due to an HDD crash… 😦 | [Fundamental Collection] thread | blog.spoRv.com | fan preservation forum: fanres.com

Author
Time

Grain is good (done right).

I'll look a bit later at the new clips, but I'm more worried that there is little comment on these clips besides myself.

Dr. M

Author
Time

Well, it's a little bit sad to see almost no one comments here; but you know, people is lazy - me too! - and sometimes waits until the final release is ready to comment - or, to grab it and fly away! (^^,)

OK, here you the last test #5:

http://www.sendspace.com/file/44stux

this has some grain, but not too much; it is basically the modified analoclean plus deghosting; and the usual magiup as upscaler at the end.

Sadly my projects are lost due to an HDD crash… 😦 | [Fundamental Collection] thread | blog.spoRv.com | fan preservation forum: fanres.com

Author
Time
 (Edited)

My dilemma is I don't know what I can offer in terms of comments as I have no expertise in this area whatsoever.

This latest sample looks good to me - I'd definitely be more than satisfied if the final product looked like this.  I didn't notice any color bleeding, and what haloing I saw was minimal.  My only issue was there appeared to be a fair amount of macroblocking, but I chalked it up to being a low-res clip.

A Goon in a Gaggle of 'em

Author
Time

bkev, any comment (if constructive and not destructive) is more than welcome! Thanks for spending some time to do it.

I used x264 @16mbps, and that should be enough to avoid macroblocking at all... I didn't notice it; could you please be more specific? Like, take a screenshot and make, let's say, a red circle around the worst offender, could be useful.

Sadly my projects are lost due to an HDD crash… 😦 | [Fundamental Collection] thread | blog.spoRv.com | fan preservation forum: fanres.com

Author
Time
 (Edited)

It's at about 4 seconds.  The color of his fur is all over the place, but strangely evens out right before he turns around.  It looked like compression to me, but I don't have the best eye for these things; so, if this isn't macroblocking but something I'm imagining, I do apologize for causing any worry.  For the record I'm also just on an uncalibrated macbook pro monitor.

http://i4.photobucket.com/albums/y137/bkev/Screenshot2013-11-15at11026AM_zpsf25db1cb.png

A Goon in a Gaggle of 'em

Author
Time

Well, I use a 22" calibrated Dell monitor, and I do not see macro blocking.

On the 4th test clip (LaserClean) there was a slight bleed on the beast's fur, which is perhaps what you saw, but it doesn't appear on the HaloClean, which is smoother.

The very slight amount of analog noise leaves me satisfied that no extra detail is lost either, so I see it as the best option yet.

Author
Time

I agree; it should also be pointed out that the final bitrate of the release will be higher; for the test clip was set at 16mbps, while the final bitrate will be around 23 or 24mbps - too much to note macroblocking IMHO...

Sadly my projects are lost due to an HDD crash… 😦 | [Fundamental Collection] thread | blog.spoRv.com | fan preservation forum: fanres.com

Author
Time

NEW TEST CLIP AVAILABLE!

Here you are the latest test clip; this time I changed scene, because I was tired to watch and rewatch always the same one...

This is made with the latest avisynth scripts, haloclean+magicup, plus a little bit of sharpening; a possible candidate to the final video quality.

http://www.sendspace.com/file/slq7le

please post your comment; your opinion is important!

***   ***   ***

I was wasting some time (when I was waiting for the clip to be processed and uploaded) making some comparisons between the BD and the LD, and I said myself "why don't you try to ColourMatch the BD, to see what will happen?"

So, here you are the results; top BD (from bluray.com), middle ColourMatch'ed BD, bottom enhanced LD (click for larger version):

results vary in quality; BD resolution is obviously better, but upscaled LD is quite good, and colors IMHO are always better than BD.

Sadly my projects are lost due to an HDD crash… 😦 | [Fundamental Collection] thread | blog.spoRv.com | fan preservation forum: fanres.com

Author
Time

Yikes, looks like ColourMatch is a pretty mixed bag here.  Belle's skin is practically purple in that middle shot.

Also, the first one looks to have really strong reds to a distracting agree.  That's a shame, but hey - the LD cap looks great!

A Goon in a Gaggle of 'em

Author
Time
 (Edited)

I agree with bkev.  The third shot is the best; the other two are not very good at all.  In fact, I would say that they are the worst results I have seen from your normally excellent ColourMatch filter, Andrea.

Author
Time

Hey, I posted them to show you that infact the BD colors are so different from the LD, that neither ColourMatch could fix them... (^^,) the fact is probably due, apart the color differences, to different cropping - LD is 1.66:1 while BD is 1.78:1; plus, I've done it using the BD jpg's and not the actual video...

This are not excuses, just suppositions... infact, ColourMatch worked well in more than 95% of frames in The Matrix, that was the most difficult movie until now... but it seems that BATB is thougher!

Well, let's forget about those screenshot comparisons; what about the test clip? Anybody?

Sadly my projects are lost due to an HDD crash… 😦 | [Fundamental Collection] thread | blog.spoRv.com | fan preservation forum: fanres.com

Author
Time

_,,,^..^,,,_ said:the fact is probably due, apart the color differences, to different cropping - LD is 1.66:1 while BD is 1.78:1

Obviously, you could crop the LD to match the BD.  I take your point though.  Your results make me wonder if the 3D would fare any better.

Author
Time

No, I could crop the BD to match the LD; but, as the LD image is smaller than the BD, I could not crop it... infact, poor results *may* be also due to the fact that the script tries to match two images that are basically quite differents between them... I don't think the 3D BD would be better...

But, did you see the last test clip? It's OK? Too much blurred? Too much sharpened?

Sadly my projects are lost due to an HDD crash… 😦 | [Fundamental Collection] thread | blog.spoRv.com | fan preservation forum: fanres.com

Author
Time

_,,,^..^,,,_ said:No, I could crop the BD to match the LD

I thought it would be the other way round, as the LD is 1.66:1 and the BD is cropped to 1.78:1.  "Crop and resize" is what I meant to type, but I assumed that the latter was implied.  I suppose whichever source has more picture information varies from shot to shot.

Author
Time

What I intended to say is that the LD has less information on the sides, in respect to the BD; so, the ColourMatch tries to match colors for the BD that has, let's say, 100% of the image, using the LD as color reference that has about 92% of the image; then, it has to guess what color the missing 8% should have... got the point? Add to this the fact that the jpg images are compressed more than the BD, they have the "blu-ray.com" overlaid on them and color palette is completely different... those things contribute to NOT match perfectly the colors... maybe using ColourLike (or ColourLikeFBF, or RGBMatch) the BD colors could be better, but color matching the BD is NOT the mission of this project; it was just a waste-some-time joke to fill the empty moments - and they are rare this week - between a project work and the other...

Please apologize if I seemed a bit rude, it was not my intention; I just wanted to explain my point of view. Are we still friends? (^^,)

***   ***   ***

I'm processing the scene when Belle discovers the rose; it should be online in less than an hour.

Sadly my projects are lost due to an HDD crash… 😦 | [Fundamental Collection] thread | blog.spoRv.com | fan preservation forum: fanres.com

Author
Time

_,,,^..^,,,_ said:Please apologize if I seemed a bit rude, it was not my intention; I just wanted to explain my point of view. Are we still friends?

That's all right, and we are.  I realize that the BD color-correction is just a tangent and not the point of the thread.  The scene in which Belle discovers the rose sounds like a perfect choice for a sample.

Author
Time

The latest clip looks very clean.  Did you increase the level of noise reduction?  I hate commenting on this sort of thing.  Msycamore and g-force are the ones to ask when it comes to spotting the side-effects of noise reduction.  All I can is that Belle's face looks quite smooth and I think the clip looks a bit blurry.  Sorry, Andrea. I'm little use when it comes to this sort of thing.

I am excited at the prospect of seeing the LD colors again and pleased that you're now putting h.264 clips into MKVs rather than AVIs.  The clips sounds good.  I hope you're using PCM for your release and not AC-3.