Why does HD need 25GB+ (Another Con Trick)

[DVD Maniac]

OK so I have now done 3 BD Rebuilder runs to DVD-5's on movies less than 2 hour run times. Watched them all and been impressed with quality so much I ran some basic visual tests. On my Pioneer Plasma I had the original running in one window and the copy in another next to each other and watched for 10 minutes. Result = I cannot see even a fraction of a difference.

So here is my question (and if I am missing something here please put me out of my misery / annoyance / confusion). The powers that be decide to release the all new improved digital video standard which we are told is 5 times more detailed than DVD and requires a new optical disc storage solution = 20-25GB.

Sounds reasonable.....

Then within no more of a year of said release a clever chap creates a tool which can create an (identical to my eyes) result in roughly a fifth of that storage space. Remembering that compressing a DVD beyond 50% usually results in obvious decay in visual quality, we are now talking roughly a 25% "compression" (I don't want to get into arguments about encoders / transcoders! - you get my drift).

SOOOO.... What the hell do we need 25GB+ storage space for when the same result can be achieved with the industry standard encoder thats being used to create (more expensive than DVD) 25GB+ retail creations? We go out and spend our money on upgraded disc players and screwed by overpriced blanks that work out more expensive than storing the same stuff on a hard drive only to find the same result can be achieved on the same size storage media one year later.....

Is it me, or is this one big massive con trick thats been played on consumers ........YET AGAIN! :devil::devil::devil::devil:

[jdobbs]

Probably a little con and a little coincidence. The design of the higher density drives were a great idea -- it just so happened that the H.264 standard was also developing that was ~4x more efficient. The result? You can choose between needing 5x as much space with the old codecs or using the existing media with the 4x efficiency. Interesting is how many commercial BD discs are encoded in MPEG-2 -- because with that much space it's fine for HD.

You know which way I prefer. ;)

[Sharc]

Most disappointing is poor quality, i.e. when a "Blu-ray" looks more like an upscaled and possibly postprocessed DVD.

[DVD Maniac]

Probably a little con and a little coincidence. The design of the higher density drives were a great idea -- it just so happened that the H.264 standard was also developing that was ~4x more efficient. The result? You can choose between needing 5x as much space with the old codecs or using the existing media with the 4x efficiency. Interesting is how many commercial BD discs are encoded in MPEG-2 -- because with that much space it's fine for HD.

You know which way I prefer. ;)

Wasn't aware that MPEG-2 was being used to create HD format - I thought HD could only be achieved with MPEG-4?

Either way, it still seems to me that the industry leaped ahead with the commercial baseline of selling an HD movie format which would generate revenue from hardware, pressed media and blank media sales. Like I said, blank HD format media prices is still a joke (HDD storage GB for GB is far cheaper), so even if you argue that the blue laser development works for increased data storage, the argument only holds good for portable storage. Even that is not very compelling given the recent developments in USB external drive size / portability and price.

The same result could have been achieved with existing red laser / DVD technology and like you say, an improved encoder, but then that would have meant far less revenue potential before the next generation of "HD+' arrived.

Seems to me that Blu-ray / Blu laser is a technical turkey!

[jdobbs]

All they really needed to do was to add AVC support to DVD players.

[laserfan]

The same result could have been achieved with existing red laser / DVD technology...Seems to me that Blu-ray / Blu laser is a technical turkey!Altho I agree with this last comment, I do not agree w/the former.

The 25Gb+ storage capacity allows the providers to load-up their discs with "value-added" features i.e. many multiple soundtracks and subtitles and extras which of course many of us don't need or use. But the capacity gives the studios to build discs that meet all needs.

While it seems clear that they might have made a market selling red laser discs, not only would these be limited in the number/quality of soundtracks, extras, et. al., there would always be some significant number of people (including many here on this bboard!) who would complain too about quality & bitrates!!!

[jdobbs]

Fair enough. There's also the advantage of the higher peak bitrate. But you could get around that with faster spin speeds. The bottom line is that new technology has to keep coming and Blu-ray did. But HD space and flash memory are sure coming up in size faster than optical... I wonder how long it will be before flash drive prices make optical obsolete? You can get a flash drive with space equivalent to a dual-layer DVD for under $30 these days.

[blutach]

I wonder how long it will be before flash drive prices make optical obsolete? You can get a flash drive with space equivalent to a dual-layer DVD for under $30 these days.Can't be too long. BD is not garnering market share fast enough to be viable on its own as a technology (even they admit that they hope to have only have 50% market share by 2012). I think flash drives will leap frog BD.

Regards

[chadamir]

Blow it up to 120" on a projector and see if there's no difference. What size TV are you watching it on?

[tetsuo55]

I wonder..

the rule of the thumb should be h264 has the same PQ at 4x less bitrate.
I guess this assumes straight from source and with the highest settings.

As transcoding would be more lossy we should up the bitrate a bit. Lets look at SD:
A single layer DVD is 4480MB this means that if the entire disc was mpeg2-video only(no audio or menu's) the target size for the h264 file would be 1120MB
If we take an average of 2 hours per movie that gives us a bitrate of 1299

It would probably be best to go VBR and up that to 1500(or is that bitwaste already?) because we don't have the original source
Extrapolate that to 720p and 1080p and you get:
480P :1500
720P :3000
1080P:3500

Now it would be interesting to see at which screen size you start seeing the difference between the source and the transcode(Insane DXVA settings + b-pyramids and max ref frames for that resolution used)

[MadMonkey57]

Blow it up to 120" on a projector and see if there's no difference. What size TV are you watching it on?

I wonder..

the rule of the thumb should be h264 has the same PQ at 4x less bitrate.
I guess this assumes straight from source and with the highest settings.

As transcoding would be more lossy we should up the bitrate a bit. Lets look at SD:
A single layer DVD is 4480MB this means that if the entire disc was mpeg2-video only(no audio or menu's) the target size for the h264 file would be 1120MB
If we take an average of 2 hours per movie that gives us a bitrate of 1299

It would probably be best to go VBR and up that to 1500(or is that bitwaste already?) because we don't have the original source
Extrapolate that to 720p and 1080p and you get:
480P :1500
720P :3000
1080P:3500

Now it would be interesting to see at which screen size you start seeing the difference between the source and the transcode(Insane DXVA settings + b-pyramids and max ref frames for that resolution used)

I like to use tools (http://forum.doom9.org/showthread.php?p=1227047#post1227047) to help me get an idea of the quality of my backup.

I know that many of you (all of you ?) disagree with me. It's OK, you can yell at me... :p

[jdobbs]

I wonder..

the rule of the thumb should be h264 has the same PQ at 4x less bitrate.
I guess this assumes straight from source and with the highest settings.

As transcoding would be more lossy we should up the bitrate a bit. Lets look at SD:
A single layer DVD is 4480MB this means that if the entire disc was mpeg2-video only(no audio or menu's) the target size for the h264 file would be 1120MB
If we take an average of 2 hours per movie that gives us a bitrate of 1299

It would probably be best to go VBR and up that to 1500(or is that bitwaste already?) because we don't have the original source
Extrapolate that to 720p and 1080p and you get:
480P :1500
720P :3000
1080P:3500

Now it would be interesting to see at which screen size you start seeing the difference between the source and the transcode(Insane DXVA settings + b-pyramids and max ref frames for that resolution used) I sure don't understand this logic... just because the H.264 encoder is 4x more efficient it doesn't mean the disc size shrinks??? Think of it this way: If a DVD-5 holds two hours of video in MPEG-2 it uses an average bitrate of a little over 5Mbs. Now -- if H.264 is 4x more efficient, that means that you are using exactly the same average bitrate, but you would get picture quality equal to MPEG-2 running at 20Mbs.

I watch all of my video on a high-quality Sony 52" LCD screen.

[Atak_Snajpera]

I would never say that x264 is 4 times more efficient than qood MPEG-2 encoder. Perhaps 2-3 times...

[laserfan]

I like to use tools (http://forum.doom9.org/showthread.php?p=1227047#post1227047) to help me get an idea of the quality of my backup.

I know that many of you (all of you ?) disagree with me. It's OK, you can yell at me... :pOK I will! Well, only that I tried the VQMT from MSI and it didn't work at all, or I should say, after 9 hours of cranking on two videos the program crashed & the output was meaningless to me. MM why don't you start a new thread describing how you use the tool(s) including interpreting of results. I would like to see... :)

[MadMonkey57]

OK I will! Well, only that I tried the VQMT from MSI and it didn't work at all, or I should say, after 9 hours of cranking on two videos the program crashed & the output was meaningless to me. MM why don't you start a new thread describing how you use the tool(s) including interpreting of results. I would like to see... :)

Choose whatever metrics (try SSIM for a start), load 1 AVS (original movie) and 2 more AVS (the ones you want to compare to the original) with different encoding settings (say different bitrates). The tool produces charts that show the metric for each frame (you can export the metrics to excel). You also get the average metric. And then compare both charts... peaks (downwards)... and check the frames that got a bad score (high motion mostly)... IMO, the key thing is to not interpret the results in absolute, but compare 2 (or more) results.

This is not a real scientific approach here. I use the tool as part of a semi-automatic analysis. It is still my eyes and my liking that decide but the tool helps in pointing where my attention should focus.

[GaPony]

If your eyes can't see a difference between the conversion and the original, is there a difference? I think not.

[MadMonkey57]

If your eyes can't see a difference between the conversion and the original, is there a difference? I think not.

All right, here we go again...

Can your eyes see a difference on a 32" LCD ? on a 52" LCD ? on a 70" LCD ? on a 100"+ projector ? I mean, what your eyes can see depend a lot on what you use to watch the material. Furthermore, don't forget to watch your backups in either dark and bright environments... I can't afford to try several encodings for several backups on several devices in several viewing environments, can you ?
Oh yeah, you can make a couple of checks of an encoding or two on your own LCD in your favorite dark environment... But what happens when you win the lottery and get a huge projector ??? What happens when you move to a new place and can't ensure proper darkness ???

You see, I use a quality measurement tool as a pragmatic mean of comparing several encodings... I'm simply putting on my side all the best I can get.

[Rectal Prolapse]

There is no reason why the entire LOTR trilogy cannot be encoded at 1080p on a double-density floppy disc! I'm sure this can be done when the new VC2 codec from Microsoft comes out.

BD Rebuilder should have no problem accomplishing the above.

[Guest]

@Rectal

If you do not have anything intelligent to say, refrain from posting, else rule 11 strikes can follow. If you need to debate this, contact me by PM, don't follow up here.

[GaPony]

All right, here we go again...

Can your eyes see a difference on a 32" LCD ? on a 52" LCD ? on a 70" LCD ? on a 100"+ projector ? I mean, what your eyes can see depend a lot on what you use to watch the material. Furthermore, don't forget to watch your backups in either dark and bright environments... I can't afford to try several encodings for several backups on several devices in several viewing environments, can you ?
Oh yeah, you can make a couple of checks of an encoding or two on your own LCD in your favorite dark environment... But what happens when you win the lottery and get a huge projector ??? What happens when you move to a new place and can't ensure proper darkness ???

You see, I use a quality measurement tool as a pragmatic mean of comparing several encodings... I'm simply putting on my side all the best I can get.

You can't really be serious, are you? :) What do you do when your "tools" tell you everything is great, but it doesn't look good on screen. Do you think your eyes are failing you because the tools said it was good? I don't think so. You, yourself said something to the effect that your own eyes were the final arbitor.

Whether something looks good on a 30", 50" or 100" inch screen is a different question. Your tools can't possibly tell you that. There are way too many variables that only begin with the quality of the copy. This is a red herring argument that has been around since people started copying DVDs to SL media and is purely subjective.

I applaud your goal of attaining the best possible result...it just seems like alot of extra work. :)

[MadMonkey57]

... What do you do when your "tools" tell you everything is great, but it doesn't look good on screen. Do you think your eyes are failing you because the tools said it was good?...

So far, the tool I use never failed me that way.

...You, yourself said something to the effect that your own eyes were the final arbitor...

Yes, Indeed. I mainly use the tool to help me comparing different encodings and to point to potential "tough scenes".

... Whether something looks good on a 30", 50" or 100" inch screen is a different question. Your tools can't possibly tell you that....

Of course it can't. But it helps in comparing encodings, therefore optimizing bitrate, media, picture size, ... eventually maximizing the chances my backup will look good on large screens.

... I applaud your goal of attaining the best possible result...it just seems like alot of extra work. :)

Well, thank you! It is indeed a lot of extra work. But I only bother going that far when I suspect a "risky" encoding.

[tetsuo55]

I sure don't understand this logic... just because the H.264 encoder is 4x more efficient it doesn't mean the disc size shrinks??? Think of it this way: If a DVD-5 holds two hours of video in MPEG-2 it uses an average bitrate of a little over 5Mbs. Now -- if H.264 is 4x more efficient, that means that you are using exactly the same average bitrate, but you would get picture quality equal to MPEG-2 running at 20Mbs.

I watch all of my video on a high-quality Sony 52" LCD screen.

What i meant was: mpeg 2 uses the full disc, while h264 could give the same quality in 1/4th of the disc, or give 4x higher quality on the full disc.
since we end users do not have access to the master, converting a dvd to 1/4th its size should be transparent(assuming x264 is really 4x as efficient)

[jdobbs]

Well... I can't swear by the 4x as that would be another one of those arguments that ends up with "what is quality?" -- but from what I've read that's a general consensus (although it seems it may be more pronounced at low bitrates).

I've done some 720x480 encodes at bitrates in the 1500Kbs range -- and frankly they look good. But having not done any extensive testing/comparisons at that range I don't think I could make any statements as to absolutes. I have also done a lot of 1080p at the 3500Kbs range (VBR with a maximum bitrate of 17500) using X264 -- and those also look very good...

No miracles reported or expected -- but my experience with X264 and high compression ratios has been very good.

[tetsuo55]

So you're experience basically confirms my theory.

Maybe someone with a 120" screen will disagree

1080P can be stored with a bitrate of 3500 VBR with level 4.1 peak limit. A DVD9 actually allows a bitrate of 12645(not counter muxing overhead and audio/menu/subtitles)

On to TrueHD/DTS-HD-Master, how big are these soundtracks(all hardware should be capable of extracting the AC3/DTS core tracks(this could have been decided when creating the new system)) and does the DVD drive spin fast enough to read both the video and the audio

As soon as open-source software becomes completely stable i'm going to use the results from this thread to transcode all current media at the discussed bitrates to MKV with h264video, TrueHD or DTS-HD-Master soundtrack and subtitles(converted to ass if possible). This should be completely transpartent on my 40" full-hd display(I have already done some testing with Corpse Bride BD and i see no difference)

[ron spencer]

I did 5 BRs here over the holidays..or holidaze...(95 mins to 152 min movies)

Well...on my 50" Plasma I noticed the following using movie only (I do not care about menus, special features, etc.).

-even with the 95 min. movie on single layer dvd-r (bd5), it mostly looks good...however, dark scenes do suffer...very noticeable to myself and guests over holidays. What I would consider an "ok" backup

-with movie 110 mins and over I do see a quality drop over original...quite noticeable

-any movie on a dual layer dvd (bd10) it looks VERY, VERY, VERY, VERY good...impressed


So for me I think movie only for "best" quality on bd9 is the way to go....of course YMMV and everyone has different tastes.

Once BD-R becomes cheap (I have some and treat them like gold), then compressing to bd-r for movies oversized or FULL backups will be real cool.

[tetsuo55]

What was the average bitrate for the DVD5 and i assume you kept the resolution at 1080p?

Did you're friends get to compare with the original bluray?

[prettyboy85712]

I did 5 BRs here over the holidays..or holidaze...(95 mins to 152 min movies)

Well...on my 50" Plasma I noticed the following using movie only (I do not care about menus, special features, etc.).

-even with the 95 min. movie on single layer dvd-r (bd5), it mostly looks good...however, dark scenes do suffer...very noticeable to myself and guests over holidays. What I would consider an "ok" backup

-with movie 110 mins and over I do see a quality drop over original...quite noticeable

-any movie on a dual layer dvd (bd10) it looks VERY, VERY, VERY, VERY good...impressed


So for me I think movie only for "best" quality on bd9 is the way to go....of course YMMV and everyone has different tastes.

Once BD-R becomes cheap (I have some and treat them like gold), then compressing to bd-r for movies oversized or FULL backups will be real cool.

Did you maintain the original resolution?

[gizzin]

Are you guys actually watching you encodes on a 1080p capable tv? 1080p -> DVD5 don't look good at all. The loss of detail is bad. 1080p -> DVD9 is better sure because of the bitrate increase but still a noticeable loss in quality. Also keep in mind if your doing 720p encodes thats less detail, so those will probably look good on a DVD9 transparent to the original in most cases. I've done alot of standard dvd encodes with x264 and I can say that 1/4 the size of the dvd results in a transparent quality encode (most cases). So saying the x264 is 4x more is efficients is a good statement to be making. We will ignore the guy thats uses tools to tell quality differences. :)

[ron spencer]

Yes...all encodes were original resolution...I did not want to look at the ini file too much yet so I kept original resolution....no I do not have logs...these were done a while ago.

My family members and friends did look at original blu ray, but I did not tell them which was original and which was bd5 or bd10. In all cases of bd5 vs original you could see difference...not sure why some say they cannot but it is pretty evident. For bd10, you can get pretty close to original, but not original quality...again this is on samsung plasma 1080p.

There is no way you can do FULL blu ray with superlative results I think...at least not on my panel

of course YMMV

[ron spencer]

I have 3 BD-rs left....maybe I should try a longer movie on put on that and see quality???

If so...what would anyone suggest? Dark Knight?

[prettyboy85712]

For BD5 I've found you really need to drop the resolution down to 720P in order to main quality close to the original.

[laserfan]

...on my 50" Plasma I noticed the following using movie only (I do not care about menus, special features, etc.).

-even with the 95 min. movie on single layer dvd-r (bd5), it mostly looks good...however, dark scenes do suffer...very noticeable to myself and guests over holidays. What I would consider an "ok" backupThe very first attempt I ever made at a BD backup was Live Free or Die Hard to a DVD-5, using jdobbs' first .bat file for x264. On my 10ft screen w/1080p projector zoomed to 2.40:1 it looked perfect. Amazing in fact. No problems at all with the many dark scenes therein. Video bitrate was something around/below 4000kbps iirc.

Of course YMMV (obviously it Has) but I wonder if there may be a quality difference between BR players--mine's a LG BH200.

[tetsuo55]

Obviously encoder settings are going to have a huge impact here. Also x264 has come a long way and with psyRDO and psytrellis.

with the most exhaustive settings enabled the results should be different today than when compared to a year ago, or even 6 months.

sure transcoding a single movie will take 48-72 hours on a quadcore but it should be transparent at a VBR bitrate of 3500(for 1080P)

[Sharc]

I have found it crucial that the display device uses the source resolution and does not apply any automatic rescaling on its own, as this can really spoil the quality of the source.

[ron spencer]

I have a bh200 as well...my panel does not resize anything....yah I used ripbot before and downsized...worked ok.

I have used the latest x264...again Ithink movie only at bd9 is really the only option if you like really good quality....again YMMV, but bd5 is out for me for any movie; quality is no there for my eyes

[ron spencer]

how can bd rebuilder change resoultion? IS this done in ini file?

[jdobbs]

As of today it doesn't -- that's a future feature.

[ron spencer]

hey that is good!!! Will be interesting to see quality of those encodes

[GaPony]

I have 3 BD-rs left....maybe I should try a longer movie on put on that and see quality???

If so...what would anyone suggest? Dark Knight?

Dark Knight would be a good one, Traitor would also be good for testing. Both have some very dark scenes and that's where the PQ seems to suffer the most, the soonest. I much prefer the Movie Only conversions to BD9 (DL media)... they look very, very good. BD5... not so much, at least not via my PS3 and Samsung 52" 1080P, 120hz TV.

[ron spencer]

not sure it is worth doing DK as bd-r....just doing movie only with DD makes a size of 27.4 meg....bdr is 25 meg. So I suspect result would be very good. Would a more worthwhile test be bd-9?

[GaPony]

DK shows in my log as a Movie Only input size of 32.66gb, but I think you'll find it to be an excellant copy of DK, if you use BD-9. Its certainly a better test if you want to see the results possible from BD_Rebuilder.

[ron spencer]

I think you chose audio of true HD 5.1...I chose dolby digital 664 bitrate, which is smaller and no subs. Is there any movie only (movie + 1 audio stream) and no subs that is say, 35 gig in total size? This would be fun to try to BD-r

[GaPony]

From what I have on the drive right now...

Pirates of the Caribbean 2 = 32gb

Pirates of the Caribbean 3 = 39.5gb

Disney's The Rookie (Dennis Quaid) = 37.8gb

We Were Soldiers = 33gb

I didn't take subs into account, but they would add a bit more size if you include them.

[Sharc]

I did 5 BRs here over the holidays..or holidaze...(95 mins to 152 min movies)

Well...on my 50" Plasma I noticed the following using movie only (I do not care about menus, special features, etc.).

-even with the 95 min. movie on single layer dvd-r (bd5), it mostly looks good...however, dark scenes do suffer...very noticeable to myself and guests over holidays. What I would consider an "ok" backup.
Dark, fading and flat scenes are a challenge to any encoder at low bitrates. There has been a discussion on similar effects here
http://forum.doom9.org/showthread.php?p=1228109#post1228109
Perhaps BD-RB may in a future version offer more encoder tweaking options, e.g. as hidden settings.
Viewers often tend to rate encoder A inferior compared to another encoder B if one scene (1 second) exhibits noticeable distortion / artefacts, although the rest of the movie is almost perfect and better than encoder B.

[jdobbs]

You also have to be very careful as to what you are really checking -- I ran into some of that on a couple discs, only to find in checking that the problem with the dark scenes were also in the original (e.g. "Lonesome Dove"). X264 can do a great job -- but garbage-in always results in garbage-out.

[ron spencer]

I checked the sources... No garbage. Pristine. Maybe will try pirates for bdr.

@sharc

not trying to rate anything...just trying to see if it is reasonable to backup bluray to bd 5...and since we had lots over for holidays I thought I would ask others what they thought on our panel.

I am getting convinced that original 1080p resolution cannot be put on dvd-5....too many people saw issues throughout the movie..

Right now I any doing Dark Knight onto BD9...this really dark movie will provide a good test as to what can be done I think.

I ordered 2 BD-REs last night and will try some BD-25 encodes later, which is really where I think we all will end up anyway (provided prices fall on bd-r)

[Sharc]

Perhaps a little off topic:
I found anamorphic encoding to be a good compromise between file size and quality when I do BD movie-only backups on a BD5. Needs a little manual work though:
Many 1920x1080 BD come with black borders. I usually crop such pictures to remove the black borders (often 132 pixels on top and bottom), then resize the remaining active picture of 1920x816 to 1280x720 and set the --sar accordingly for anamorphic encoding (45:34 in this case). This preserves the original resolution / details as much as possible despite resizing, because the full 1280x720 has become the active picture area.
Anamorphic encoding is included in the BluRay spec IIRC - so there should be no basic compliance problem. Playback and aspect ratio are flawless on the PC/projector, I could not yet test the playback compatibility on a standalone player though.

Now, don't try this with Dark Knight, because that one changes the format (black border width) dynamically.

[tetsuo55]

Ron what settings are you using for x264? and are you cropping the video where possible?

[ron spencer]

settings? the default ones in BD rebuilder (highest quality)...no resizing...although I have used ripbot before to resize to 1280 and worked ok

[Sharc]

..... Right now I any doing Dark Knight onto BD9...this really dark movie will provide a good test as to what can be done I think.
I did that one before as movie-only, resized to 1280x720p in view of its duration. I am very pleased with the result. I really had to analyze it in detail to discover deficiencies in difficult scenes.