Tools to compare re-encoded DVDs?

[Skunk]

I have been practicing making backups for some time now and trying out all sorts of tools, encoders, burners etc and have got to a point where I'm pretty comfortable.
Most DVDs I simply run through DVDReMake to remove the previews, unnecesary audio and subtitles, and any extra content or menus I dont need. Then it's on to DVD Rebuilder. I like HC encoder for heavier compression, but I find that I can get great results if its under about 80% using ReJig, and boy is that alot faster. (Yes I know why, please dont go there.)
Anyways, my question is this : Are there any tools I can use to analyze the quality of 2 DVDs for comparison of overall quality? Alot of the DVDs I have made buckups for both ways I can't tell the difference. Or should I pretty much assume that a re-encode will be better quality that a transcoded copy every time?

[Rockas]

Well I usually us two different tools... my right eye and my left eye :D ... just kidding but also 'talking' seriously.
A big screen TV also helps... The bigger the best (in this case size matters) :)

[hallway]

I was going to same the same thing but I figured he wanted a "serious" answer. :D Really though, your eyes are the best tool and in fact the only tool that can decide (for you) which is better.

[Harrysmiith]

Or should I pretty much assume that a re-encode will be better quality that a transcoded copy every time?

My understanding is that transcoders basically squash a movie into the available space. So everything gets sqaushed ( more or less) by the same amount.

Re-encoding on the other hand steals bits from one part of the movie to give to another part - hoping that no one will notice that some bits have gone missing and hoping that everyone will notice that some bits have been added.

With movies requiring little squashing or only being viewed on modest equipment one method may be said to be "better quality" than another but if it can't be seen then it can't be seen.

[kumi]

Have you played with avisynth (http://www.avisynth.org/)? I use it to compare different rebuilder and encoder settings.

At the risk of posting in the wrong forum, I'll paste part of a thread I wrote in another forum for someone who asked "how do I use avisynth to compare?" All of the functions are from other people, I just collated them into an .avsi file (see below). Some of the examples compare filtered vs. non-filtered; but you can use them to compare 2 different .d2v files as well.

Relevant avisynth threads
Set your monitor levels (http://forum.doom9.org/showthread.php?t=90663) <--- Do it
Is there any way to remove Mpeg2 blocking? (http://forum.doom9.org/showthread.php?t=98550)
Undot and Deen on interlaced source (http://forum.doom9.org/showthread.php?t=97628)
Alternate -- a way to compare the effects of other filters. (http://forum.doom9.org/showthread.php?t=97059)

Download MyComparisonScripts.avsi (http://www.piracy.se/~kumi/adc/avisynth_filtering_guide/) to your avisynth plugin folder.

Using Artifacts() to better show blocking and mosquito noise for a NON-INTERLACED segment
trim(6012,24791)
Artifacts()
ConvertToYV12()

Using InterlacedArtifacts() to better show blocking and mosquito noise for an INTERLACED segment
LoadPlugin("C:\Program Files\DVD-RB PRO 1.03\DGDecode.dll")
mpeg2source("G:\SUNTACK\DVDRB TEMP\D2VAVS\V02.D2V",idct=7)
trim(6012,24791)
InterlacedArtifacts()
ConvertToYV12()

Another way to see blocking...
mpeg2source("VTS_02_2.d2v").Histogram("luma")

Once you have 1 or more encoded .m2v files, load them in DGIndex and "save project" to save the D2V file(s).

Comparison AVS scripts:

Compare 2 mpeg2 clips by interleaving
a = mpeg2source("VTS_02_2.d2v").Artifacts() # or InterlacedArtifacts() if your clip is interlaced
b = mpeg2source("VTS_02_2_filtered.d2v").Artifacts() # or InterlacedArtifacts() if your clip is interlaced
Interleave(a.subtitle("original"),b.subtitle("filtered")

Compare 2 clips (splitscreened vertically)
a = mpeg2source("VTS_02_2.d2v").Artifacts() # or InterlacedArtifacts() if your clip is interlaced
b = mpeg2source("VTS_02_2_filtered.d2v").Artifacts()
BeforeAfterClip(a,b) # see function definition for more settings

[jdobbs]

Comparing and making any meaningful conclusions is almost impossible. I've seen a lot of folks try here on Doom9 -- but the conclusions often fall in the face of logic.

Comparing transcoded output to encoded output is even more difficult -- this is especially true when doing frame comparisons. Transcoders aren't equal opportunity employers... they steal most and sometimes all of their reduction from B-Frames -- often without touching I-Frames. So some tests that average out differences between the original and the output give them inaccurate scores (a "perfect match" gets a lot of weight, while a grossly distorted frame might not).

I'd agree with the earlier comments. The eye is the only comparison that counts.

[Skunk]

Thanks for the input everyone. I sort of figured if I can't SEE the difference then it doesn't really matter much. I'm just trying to compare two disks to see if maybe one will be better in the longrun if I get a nicer setup, thought the one i got now is pretty decent.
Im just a perfectionist. It can't be helped.

[apfraats]

My method (oh boy I know what's comming)


1) BIG BIG TV.
2) Search for an heavy action scene with a lot, I mean really a lot picture to picture differences.
3) Press PAUSE on the start of such a sequence.
4) Activate if not already so: ZOOM 1600% on youre TV or DVD player.
5) Inspect the image at detail level using shift over the picture; if blocks are seen goto 7
6) Go 1 step forward (1 picture) and goto 4)
7) Search for the same picture on the source.
8) Press ZOOM 1600% (like 4)
9) Is it already present on source ?, if yes then bad source or bad equipment, continue 6)
10) If NO: You have quality loss due to macroblocking.

You can extend this with comparing source details with re-encoded details to see if you have lost quality in details, but that you can do with static scenes also and is easier to see.

However the first sign is the macro-blocking in a sequence of heavy different pictures one after the other, because the difference between two pictures has to be encoded....

And Especially for transcoders:

Play the movie, and watch for PULSING NOISE.

If you see a PULSING NOISE, it's because the original movie contains noise and that is left in the often untouched I frames. However noise is also DEATIL, that's getting lost in B,P pictures as transcoders gets rid of details here to compress more.

So you will be left with a movie that has a PULSING noise over it, depending on GOP length. As we have 25 pictures/sec here that's about twice a second when using GOP-length of 12.

[dragongodz]

1) BIG BIG TV.
2) Search for an heavy action scene with a lot, I mean really a lot picture to picture differences.
3) Press PAUSE on the start of such a sequence.
4) Activate if not already so: ZOOM 1600% on youre TV or DVD player.
5) Inspect the image at detail level using shift over the picture; if blocks are seen goto 7
6) Go 1 step forward (1 picture) and goto 4)
HAHAHAHA oh this is classic. if you need to go to these extremes to see blocking or any other artifacts or differences you really need to get a life. :D

My method (oh boy I know what's comming)
well atleast we know now how you can see differences. dont expect much notice to be given to observations based on such anal-retentive behaviour. ;)

I sort of figured if I can't SEE the difference then it doesn't really matter much. I'm just trying to compare two disks to see if maybe one will be better in the longrun if I get a nicer setup, thought the one i got now is pretty decent.
you are right, if you cant see a difference then thats all that really matters. some dvds will easily benefit from encoding while others will be fine for transcoding etc. its up to you to be happy with the output of your choice and not others. if in the future you decide to get a setup where you are no longer fully happy with some of the dvds you have done then you can always redo those of course. that doesnt mean you will have to or that they will all need it, far from it.

[apfraats]

I just knew it.......:D

But still you can tell it's a MPEG-2 stream this way, and how bad/good it is.

For example if you already see blocks at 4* youre near to see blocking at 1*, depending on the quality of youre eyes, distance to youre screen, and size of youre screen.

Anaolgue stations came out much better in this way, seeing much less garbage (a little noise, but at a lot less level then the macro blocking).

So just blowing things up, is what justified, as TV screens are getting bigger and bigger and have more resolution, you WILL get to see MPEG-2 artifacts......

If you look at 4* zoom and it still looks reasonable, sure youre quality is high enough. Enlarging at 4* it not so big a deal, it's like using a projector TV.

The other way around:

Looking at a 4 inch screen from 4 meters viewing distance delivers exeptional results having just 2000 kbps average !!! Fantastic picture !!! (you just can't read subtitling, how pitty)

So yep, it's subjective to youre viewing environment and of course blowing up 16* is like trying to make a cinema screen wide viewing spectacle at home, but I just want to point out that I got into the bussiness of DVD-RB just after having bought a new BIG screen TV that make me feel sick looking at my old SHRINK backup's, while uptill then I considered them 'Oh So Well' ; man I was happy with them, yep, untill with three people we installed that bloody BIG picture tube, and started looking DVD's on it..... Oh man, that changed my quality standards drastically.

So in fact it is mainly depending on sceensize/viewing distance ratio and native picture quality of the components used, and yep of course youre eyes, and the attitude one has to get 'perfect results'.

And with the tendency to make screens a lot bigger with higher resolution, you can have to redo backups in about 2-4 years from now......

But hey ?, who matters ?

Then we use blue ray with all it's copy protection schemes that cannot be broken anyway............ and this forum would no longer exists.....

It will be like comparing vinyl records to cd's (and even now some say the vinyl was sounding better :D :D )

By the way: If youre TV has a zoom function that is implemented badly, it causes blocking as per deafult, not being Mpeg-2 artifacts, but seeing the limitations of the picture stored in digital image memory. That is most of the time also the case using zooming if present on youre DVD-recorder.

Still if we want to know more about something, we enlarge it.

Biochemics use microscopes, even electron microscopes, and a collector uses a loop to examine his stamps for imperfections.

So what is wrong with it ????



My PHILIPS has done this properly, by storing the picture in high resolution in memory, so blowing it up, makes you not see the memory resolution but the picture details in fact.

Even the white/black noise (normally supressed with a 'blue' screen, but not if comming from an extrernal source), you'll get the noise enlarged on screen.

Also very nice to adjust antenna signal for optimal picture quality concerning analogue stations, which are very good here.
Just blow up to 16* and try to get the lessest noise, simple but effective, especially while test signals are transmitted.

[dragongodz]

But still you can tell it's a MPEG-2 stream this way, and how bad/good it is.
no you cant. your method fails to be truly representative of what you will see with new hardware for several reasons.
1. you can not be sure that zooming is really going to reprisent the larger screen. what type of zooming(quality) is your player doing compared to what the tv will be doing for example ?
2. nobody i know watches movies by looking at a single frame for a second or two and then advancing to the next. video is watched in motion and some artifacts or imperfections etc you may see in a still frame will not be seen at all when really watching it that way.

Biochemics use microscopes, even electron microscopes, and a collector uses a loop to examine his stamps for imperfections.

So what is wrong with it ????
we are talking about video. how you equate that with looking at a stamp i will never know...unless maybe you stacked a few thousand stamps on each other and had one fly from the top of the pile every 1/25 of a second(PAL of course). :D