DVD9 vs DVDRB with CCE/HCE vs DVDFab

[Th3KiNG]

Source DVD9
http://img651.imageshack.us/img651/8107/dvd9i.png

DVD5 using DVDFab
http://img195.imageshack.us/img195/7/dvd5fab.png

DVD5 using DVDRB/CCE | Encode 3pass
http://img109.imageshack.us/img109/5329/dvd5cce3pass.png

DVD5 using DVDRB/HCE | Encode 2pass
http://img831.imageshack.us/img831/8228/dvd5hce2pass.png

[SuLyMaN]

What are the differences? Seems pretty much the same to my eyes?

[Ghitulescu]

The indian movies are generally heavily denoised, so practically any "shrinker" will perform more or less identical, in particular if the size reduction is not that much (say ~90%, 5GB->DVD5). And comparing static images are not relevant.

[steptoe]

It would have been better to post the images as a lossless image, sunch as a BMP or lossless TIFF file that way you're not losing any data to compare the results

Save the images and zoom in to 400% then you'll see the differences

Personally, the DVFab results appears better but probably because it has more smoothing, but the CCE and HCE versions are close but a bit more blocky, but that's because there is less smoothing so to me the CCE/HCE are better as the DVD Fab is applying much more smoothing to reduce the size effect of blocking

I've also noticed that Indian movies have terrible noise so need strong denoisers, which also means more smoothing so need to apply some sharpening

[jdobbs]

Too many factors at work -- that's why I always ignore these comparisons. If the original was encoded at an exceptionally high bitrate, any mediocre package could give you decent results. The sun even shines on a dog's ass sometimes, but that doesn't mean you should try to use it as a tanning booth.

It's when the going gets tough that the better packages stand out.

I know that DVD-RB gives me the best results for any given source... and that's all that matters to me.

[Th3KiNG]

update... tested CCE SP2

DVD9
http://img832.imageshack.us/img832/3319/dvd9e.png

DVD5 using DVDRB/CCE SP1 2.50.01.00 | Encode 3pass
http://img90.imageshack.us/img90/9383/dvd5ccesp1v2503pass.png

DVD5 using DVDRB/CCE SP2 1.0.1.9 | Encode 3pass < 2x faster :D... and a lot better quality....same as DVD9
http://img39.imageshack.us/img39/589/dvd5ccesp23pass.png

If DVDRB adds CCE SP3 it's will be awesome... after seeing in CCE SP2 speed/quality improved can't wait to find out how fast and quality i get in CCE SP3 :D... (btw I found out yesterday CCE SP2 works with DVDRB :D... after seeing CCE SP1 vs CCE SP2 results... i got shocked)

[Chetwood]

Too many factors at work -- that's why I always ignore these comparisons.
Especially when you get still images to judge the quality of moving pictures.

[Sparktank]

images as a lossless image, sunch as a BMP or lossless TIFF file

Isn't PNG a lossless picture format? :confused:

[jdobbs]

I believe so. But, why would that matter? Using a still picture to evaluate a video stream is like looking at a glass of water to evaluate a waterfall.

Example: In early computer animation each picture was created as if it were snapshot with slight changes where the motion takes place... but it looked terrible when played back. They realized that motion had to purposely be blurred in order to give a natural effect. The lesson learned: You can't treat a motion picture like a group of independent snapshots -- they are two entirely different things.

[Sparktank]

Perhaps a custom made DVD of short samples to be played back on a rewritable would serve a better function of comparisons.
Mpeg files in muxed mpeg (non-elementary) might do, but then your perception could change depending on your codecs and settings.

I did a short test DVD when I first learned about SplineRisezers larger than 64 (100/144).
I added (forced)subtitles to each title that would indicate what spline was used.

Perhaps a sample DVD? (>400mb)