Testing DVD2ONE...Result..not good..

[DVD__GR]

Hello guys i tested blade2 encoding of dvd2one,
this movie was chosen not for comparing image quality
because as we know its compressibility is tremendous,
but to see how this program works..
my results.. were not good (for me at least,
everything you can see in the graph i attach the time the moderators let it be posted.
LEGEND:yellow line(bitrate)lower line->dvd2one..

PS:I have nothing with this program,i still believe its a solution
if you dont have the time to encode,but look at the graph guys.its really a question the curve...Waiting for future releases...

[mpucoder]

I approved the graph, but I'm sure I'm not the only one who doesn't get it. The bitrate is lower, what else would you expect? Or is it the sudden rise in bitrate toward the end, indicating that DVD2One was too conservative initially. Well, 2 passes might improve that, but that's not the nature of the algorithm used.

[DVD__GR]

Hello mpucoder thanks for letting it,I just want to get the answer that everyone asks,I dont want neither to insult the author neither you nor anyone.
The think is that its like you said,
in the middle part of the movie the bitrate is really too bad,
the problem is not the bitrate but its distribution which is totally wrong in my decent opinion,a bit of global bitrate estimation would be great.

[jdobbs]

@mpucoder

I thought the bit distribution in these types of transcoding algorithms exactly matched the original allocations only at a lower rate?

[mpucoder]

Not exactly sure what it does, but I am sure it does this in one pass. I know of several different strategies, some of which are described in the links here http://forum.doom9.org/showthread.php?s=&threadid=44321
None of them produce a strict ratiometric bitrate reduction. They depend on perceptual models to determine what can most safely be decimated. Motion is one important part, high motion can get by with less detail (blurring).
Another trick is reducing one (Cb) of the color components, as the human eye is less able to detect detail in blue. This has been used in NTSC since the beginning, where the color space is a close relative of YUV, called YIQ. The I vector is basically red, and has a bandwidth of 1500KHz, whereas Q (blue) is reduced to 500KHz.

[atreides93]

So like what is this graph supposed to be showing???

what are the other colors representing?? you have two yellow lines, you say the lower one is dvd2one? what's the upper one? cce??

[Nic]

@DVD__GR: Just of interest what was your opinion of the visual quality of the dvd2one outputed movie ?

-Nic

[mb1]

I reported that negative bitrate behaviour already 2 weeks ago
http://www.dvd-svcd-hilfe.de/forum/mb1/dvd2one/dvd2one.htm

First curve is LOTR SEE PAL (first 500 seconds).
Second graph is DVD2one (look at the first minute where the movie is nearly black only and look at the last minute where the bitrate drops down)
Third graph shows CCE.

Because of that bitrate behaviour DVD2one unnecessarily decreases quality in some scenes too much.

[DVD__GR]

Originally posted by Nic
@DVD__GR: Just of interest what was your opinion of the visual quality of the dvd2one outputed movie ?

-Nic

I can say that the output quality was acceptable but
My purpose was to find how this program works..meaning how it
makes the adjustment of bitrate to fit a dvd,with this graph
i saw that the distribution was really really bad (in my opinion)..
Look at the middle of the movie,the bitrate was flatted out,look at
the end and the beginning,it is almost source bitrate!!
@atreides:The other lines are the quantizers,the up yellow is the source vobs and the low yellow is was dvd2one outputted.
Summing up I could say that the author should check it,i am surely he
could do something to fix that.
Waiting for next release...:)

[vorbis]

lets not go nuts here, these are backups we are creating and the only important factor is visual quality. If its acceptable for you then its doing a good job.

[DVD__GR]

its a matter of choice,if you put that on a projector you should see
nothing but the truth and believe you would not want to compare with
cce..