Highest compression used without noticable quality loss?

[h8sh8]

Sorry once again but looking at the calculating bitrates formula it still seems that what the average bitrate equals is dependant on the end filesize and the running time and vice versa - filesize is dependant on bitrate wanted & running time...

"The basic bitrate formula is

(Size - (Audio x Length )) / Length = Video bitrate

L = Lenght of the whole movie in seconds
S = Size you like to use in KB (note 700 MB x 1024° = 716 800 KB)
A = Audio bitrate in KB/s (note 224 kbit/s = 224 / 8° = 28 KB/s)
V = Video bitrate in KB/s, to get kbit/s multiply with 8°."

I just can't see where the percentage comes into it - couldn't two movies of the same length have different average bitrates, and therefore different filesizes, and therefore when compressed at the same percentage have worse results compared to the same two different movies compressed at different percentages to equal the same filesize and therefore the same average bitrate?

So I guess in the theory I said in earlier posts, instead of filesize (no audio) with running time taken into account being more accurate than pecentage, you could say average bitrate (which without audio is basically Size divided by Length) is more accurate than percentage. Sorry, but to me 'bitrate' just seems an easier way to represent what I was sayng about 'filesize & running time'.

Sorry if I'm becoming an annoying pain in the ass here - just can't get my head around it any other way. I hardly slept the last couple of nights - maybe that's my problem - surely I've lost the plot now haven't I?

[jeremyp969]

One other thing to consider is that not all encoders are the same. So a movie compressed to 3GB might not have the same visual quality as that same movie compressed to 3 GB with another encoder. My point is that, if you start with a 90 minute movie on a DVD-9 at 7GB and compress that down to 4GB, or if you take the same movie on a store bought(NOT COMPRESSED BY YOU), take that same 90 minute movie, but it's already at 4GB, chances are the latter will give you better quality. The reason for this is the studios who release the DVD have much more sophisticated encoders than say DVDShrink. I'm not bashing DVDShrink at all, I think it's an awesome tool and I use it religously, but it's compression techniques are a little off. Also, if you start out with a higher quality movie and compress it (in the studios before they compress it they have a VERY high digital quality movie) then your compressor will have a better prediction step and you will preserve the quality of the movie better.

Back to what I said originally though, it depends on the TV and your preference. WIth a tube TV or a plasma TV you have to be much more careful with what you do because you'll notice a degradation of quality because of how nice the picture can look on those TVs(Hi-Def ones). However with a projection, rear or front, you can compress the movie a lot more and not notice it because of the darkening issues with projection.

I have a 40inch tube, and my general rule is,
1. Never compress more than 2 hrs of movie onto one DVDR
2. Never compress to less than 85% of the original file size.

DVDRs are so cheap now($.60 for DVD+R Memorex) as well as dual DVD cases, you might as well just preserve the quality. If you have a DVD changer you don't even have to get up to change the DVD.

[h8sh8]

Ok, my brain just kicked in

I think I'm confusing the crap out of myself. Though I think the theory still works...

Where my thoughts originally evolved from was that I was coming from the P.O.V. of trying the compress the movie to a minimum before becoming visibly bad quality to allow more extras/menus on the disc. This is where I think my theory would work, and is what I thought the original post was about.

for example, following my way of thinking, don't go below 3gigs (without audio) for a 90min movie at whatever percentage (or alternatively don't go below a certain average bitrate - 4562kit/s ???), as 60% (for example) may or may not go below 3gbs therefore being either good or bad quality.

Ok, I'm banning myself from this thread now - my apologies

[h8sh8]

One other thing to consider is that not all encoders are the same. So a movie compressed to 3GB might not have the same visual quality as that same movie compressed to 3 GB with another encoder. My point is that, if you start with a 90 minute movie on a DVD-9 at 7GB and compress that down to 4GB, or if you take the same movie on a store bought(NOT COMPRESSED BY YOU), take that same 90 minute movie, but it's already at 4GB, chances are the latter will give you better quality.

I know I said no more posts by me, but you got this as I was typing the last one.

I think you just shot my idea to hell!!
but... does it also mess with the percentage way?
would just simply judging by eye be the best in that situation?

[aggies11]

re h8sh8:

I think the stumbling block here is that you don't get to *choose* the %. Ie it's not an arbitrary value.

A nice way to think of it is that the shrinking is done "just enough" so that it can fit on a single DVDR. So by definition, the % is infact related to the original size. To fit a bigger movie onto a single DVDR you have to, obviously, shrink it more (lower percentage). While a smaller movie, requires less shrinking (higher %) to fit on.

So % in general is merely another way to relate the original movie size.


There can be a small snag to the above though, when you take into count bitrates. More to the point, you can have a long (eg 3hr) movie encoded at a low bitrate, and a shorter (eg. 1.5hr) encoded at a high bitrate.

These movies can both be the same size (eg 7GB etc..). Each movie would then get a pretty similiar shrink % (eg. 60%). However the effects of a 60% shrink on a low bitrate original vs high bitrate, could be quite different. Namely higher bitrate originals handle the "shrinking" better, while the lower ones would fair alot worse. (eg. The 8Mb/s original gets reduced to say 5Mb/s, while the 4Mb/s original ends up being 2.3Mb/s).

The lower the final bitrate, the more quality issues you have.

So in general, %Quality is an indication of how much shrinking is done, which is directly related to the original movie size. You only run into problems comparing these figures on special cases when you have movies of substantially different bitrates (often the two extremes of movie length).

For Reference, I just did a CloneDVD2(1.9.2.4SneakPreview) of Fellowship of the ring EXT version. (1:1 copy per disc). It was a full backup, except for french audio and subtitles(with all 6audio tracks + extras). Resulting in a quality rating of about 60%. In normal watching of the film everything looks fine, and I was quite surprised. You can't really guess that it's not an original. All perceived differences are likely subjective/psychological . Only on 8x zoom, using the hi-res computer monitor, can you see differences, in particular, that text is somewhat blurred. But on regular viewing (especially on a tv) the picture is extremely crisp.

So I'd say that transcoding has reached a level where, to all but the most scrutinizing enthusiast, the differences between original and backup are largely indistinguishable. Which ofcourse, is the perfect time to introduce dual layer recordable discs...

Hope that was of use

Aggies