Shrink better in low bitrate scenes ?

[apfraats]

First of all I do not want to suggest that DVDSHRINK is any better dan DVD-RB-PRO in general, but just want to mention 1 aspect only.


If I do Star Wars III with DVDSHRINK and with DVD-RB-PRO I notice that action scenes get's worse in SHRINK, but the low-bitrate , non action scenes looks better in this particular case.

Result with DVD-RB-PRO and CCE:

You see clearly GIBBS-artifacts, as musquito noise around sharp contrasted area's in low-bitrate and more static scenes.
When using bitrate viewer there is a much lower Q.-level.

Result with DVD-SCHRINK:

No noticeable BIBBS-artifacts around sharp constasted area's, much better result. The Q-level is higher.



How can I improve this ????

That musquito noise around sharp contrasted area's is annoying and with about the SAME BITRATE it isn't there with SHRINK.

Of course I don't gonna compare the action scenes, but it seems DVD-SHRINK does do a better job regarding to the GIBBS-artifacts.


Can I improve this in the DVD-RB environment without loosing high-action scene, high bitrate quality ?

Maybe a way forcing to do more compression at the same bitrate would solve this as SHRINK results in a higher Q-graphic with BITRATE viewer ????

Thanx

[Jeffster]

Maybe adjusting the quality_prec setting can help?

This is a parameter for the balance of bit allocation between
simple and complicated part. The range you can set is 0 to 100.
As the value becomes closer to 0, a higher bit amount is allocated
to complicated part of the image. As the value becomes
closer to 100, a higher bit amount is allocated to flat part of the
image. When the value is close to 0, the mosquito noise at the
edges (noise causing hazy part along the edges, looking like flying
mosquitoes) is less noticeable, but the contouring noise (noise
which looks like contour line patterns, which appear in flat and
wide areas, such as a dark background) is more noticeable. The
opposite occurs when the value is closer to 100.

The optimal setting depends on the footage, but roughly speaking,
16 to 40 is recommended. If the bitrate is relatively high,
greater value may bring better result.

[happycase]

I have my qualiy_prec at 24 for over 4 mbps. For less than 4 mbps, I set it down to 20. If the film is 3 hours like Alexander, I'll try encodes at 16 and 20 and see which one has less mosquito noise.

[writersblock29]

My own CCE settings for most projects involve a Bias of 10, Quality Prec of 16. There are times I'll tweak this, but for DVD projects this is the norm. Usually 2 passes are fine, but on Revenge of the Sith I beleive I used 3 (total running time was above my rule-of-thumb 2 hour mark). On this project, I remember I also pre-processed with DVD Remake Pro, removing foreign language tracks (and deleting the menu buttons that would have called them) and angles (which were also foreign-language associated). I may also have used DVD Shrink to knock down the menus a bit... but I don't recall if I did or not for this one. Long-winded, yes... but the result pleased me enough to burn it and move on.

DVD Shrink doesn't shave an equal amount off of the entire video stream (like many other transcoders will), so it's entirely possible that the "better" footage was simply closer to the original than the re-encoded copy. Shrink-vs-Rebuilder comparisons will, unfortunately, always depend on the source material. If Shrink can make a juicer cut in a bitrate-bloated section of a given project, it won't have to scoop away at a section that's on thin ice to begin with... leaving that section closer to the original than an encode will. It should also be mentioned that some DVDs have quite a surplus of bitrate, and removing some of it won't affect scenes that really didn't need it in the first place... like those areas that had a higher Q but better percieved quality that you noticed. (This is actually what DVD Shrink is "deciding on" while doing the Deep Analysis pass.) BUT there aren't many full-disk projects that Shrink wouldn't have to slice thin -- and in places that WILL make a difference -- in order to make it fit, which is where encoders shine compared to transcoders. You also can't apply filters to a transcoder, so you'll still have a noisy copy of a noisy original -- perhaps MORE so, since severe transcoding will add macroblocks and other artifacts of its own.

It'll always be a toss-up, since quality's subjective. I've read posts from a well-respected member here that state most encodes can't be distinguished from transcodes. While I personally disagree that this applies in all cases, I'll admit that DVD Shrink had me abandoning the old methods involving DVD Maestro, Chapter Extractor, CCE, ect ect ect ect for some time. Gaining High Definition playback equipment made me glad Jdobbs was creating DVD Rebuilder, though... and I've certainly re-done many old DVD Shrink backups to gain the extra quality I perceive exists with encoding.

[manono]

apfraats-
Result with DVD-RB-PRO and CCE:
You see clearly GIBBS-artifacts, as musquito noise around sharp contrasted area's in low-bitrate and more static scenes.
When using bitrate viewer there is a much lower Q.-level.
and:
Result with DVD-SCHRINK:
No noticeable BIBBS-artifacts around sharp constasted area's, much better result. The Q-level is higher.
Do you understand that a higher Q-Level in Bitrate Viewer means generally lower quality, and that more detail is removed?

Having said that, it's understandable that Shrink might give a higher Q-Level, and maybe look better in those static scenes at the same time. Shrink keeps the same matrix, and I have no idea what matrix you used with DVD-RB. If the Standard Matrix, then maybe redo it with the same matrix as the original (and Shrink) for an apples to apples comparison. Lower bitrate matrices, such as the Standard Matrix, display mosquito noise much more quickly than do better matrices. I don't have that DVD, but quite a few recent ones are using the Standard Matrix with AQ turned on (something not possible with DVD-RB these days), meaning that the static scenes will use a very good matrix. In addition, the encoding for the retail may have had different settings than you used for your encode. That is, it may have given relatively more bits to the static scenes and less to the complex scenes. If that's the case, then the complex scenes transcoded by Shrink should look worse than the DVD-RB/CCE encode. It's a tradeoff. You only have so many bits to go around, and if you improve the quality of the static scenes to get rid of the mosquito noise, those bits will have to come out of the complex scenes. As mentioned earlier, adjusting the settings to give more bits to the static scenes may help. And I, for one, prefer to use just about any matrix other than CCE's Standard Matrix.

Nice post, writersblock29.

[writersblock29]

@manono

:thanks:

[apfraats]

First of all, thanks for the info so far, everybody !

I have to set Q-prec lower I think for first.

Do not forget I DON'T talk about OVERALL QUALITY, as you can throw DVD-SHRINK away. Star Wars 3 seems a typical movie that's 'DVD-SHRINK' friendly.

However it blocks more in high action high motion scenes, but that's only noticed when you press pause and do picture to picture stepping. I even than have to count the number of blocks to make sure which is the best...

I use DVD-RB(-PRO) since I was convinced it overall quality is a lot better.

The only reason I can think off, is that decoding introduces the GIBBS artifacts and reencoding it again including these GIBBS artificacts gives somewhat more muisquito noise around sharp edges.

Because Shrink doen't have to suffer from extra decoding artificats, maybe that's why CERTAIN static scenes show up less of this kind of noise around sharp contrasted areas (A robot head in front of a somewhat darker background).

Because there are a lot of CGI in star wars III, edges can be 'unnatuarally' sharp.

I just noticed it, and setting Q-level lower and using a higher bitrate matrix helped a lot.

When you use a constant bitrate source, DVDSHRINK messes up totally.

In general results with DVD-RB-PRO are always better, and I also redo a lot of SHRINKED dvd's, especially those made without AEC, they look really bad.... With AEC enabled there is much more room, but still not comparable to DVD-RB-PRO and CCE. DVDSHRINK is really source dependent and DVD-RB a lot less. But I have made perfect result with some movies at below 60% compression with SHRINK and AEC.

So generally seen I prefer DVD-RB-PRO because given a certain length of movie and average bitrate I always know what to expext.

But still it would be nice if you could mess with the Q-factor.

Some original movies have over 2 hour DTS video (with 5.1) and fit perfectly on a standard 4.36 GB DVD.... I always notice higher Q values here.

But as far as I know, you can't infuence the Q-factor as BITRATE-VIEWER uses and shows it........