www.makemkv.com

Hello, I'm considering ripping my entire DVD and BD collection to MKV. The first (and obvious choice for ripping to MKV, given its reputation in the community) was MakeMKV. The only thing I don't like about MakeMKV is the fact that there is no preview available, which would be extremely helpful in identifying individual episodes in series disks. This led me to search for additional/alternative tools that do the same job, which brings me to my question.

Shouldn't two applications that merely rip in lossless format the same content from the same original source (disk), produce the same MKV file size? At least, that's what I expected. I ripped the same BD twice, (obviously keeping the same audio streams and subtitles in both applications), but the resulting file was several MBs different between MakeMKV and the other application. I am deliberately not mentioning the name of the other application, being this the MakeMKV forum, I will gladly do so in a follow-up post, provided the moderator is in agreement.

I would expect a digital rip (byte-per-byte) copy to return the same size regardless of the tools used. Obviously, at least one (if not both) of the two applications does something more than just rip. Am I wrong to think that?

Thank you for your time.

Hi!

Shouldn't two applications that merely rip in lossless format the same content from the same original source (disk), produce the same MKV file size?

Yes - except for the fact that programs such as DVDFab and AnyDVD have additional optional settings not found in the purer MakeMKV, such as elimination of "BD+" and "PUOs" (Prohibited User Operations)...

Thanks for your reply setarip. This could very well be the answer provided the output file were a few kbytes different. However with the difference being MBs instead of KBs, I'm having difficulty believing the these can be attributed to the omission or inclusion of metadata.

well extract the raw streams from the mkv and compare them

skittle wrote:well extract the raw streams from the mkv and compare them

I did just that on two DVDs. I extracted the video stream from the mkvs and on both occasions I had a difference in size ranging from a few KBytes for the first DVD, to 3.5 MB on the second. One thing that remains constant is that MakeMKV creates consistently a larger file that the one generated by the other application. The tools used to extract the raw streams from the mkv files are mkvcleaner 0.5.0.3 and mkvtoolnix 4.9.1

I know that in reality this probably doesn't make any difference whatsoever; I just wanted to point out the fact that ripping even to a lossless format does not necessarily mean that we get a 1:1 copy of the original content.

setarip_old wrote:

I extracted the video stream from the mkvs and on both occasions I had a difference in size ranging from a few KBytes for the first DVD, to 3.5 MB on the second.

And now perform the next step - do an individual file-by-file size comparison of ALL files (.M2TS, .MPLS, CLPI, etc.), including ALL Java files...

Java files? I've lost you. All redundant content has already been removed during the mkv creation and the only thing left is the actual movie, 2 audio streams and 1 subtitle. I'm not ripping DVD ISOs or BDMV structures to have Java files. I am comparing the extracted raw video streams from the mkvs and they are different. What does Java or anything else have to do with this? Isn't this proof that at least one of the two programs is NOT doing a 1:1 rip?

The other program is Pavtube ByteCopy. I still don't know whether I should have mentioned a competitor program in this forum; my apologies to the moderators if I have violated any forum rules.

I have confirmed "cutter's" findings by randomly choosing 1 DVD + 1 BD from my collection. Upon converting them to .mkv by using both makemkv and bytecopy, the comparison revealed that the .mkv file is ~1% larger compared to the one created with bytecopy!

Furthermore, not only the output .mkv files are different in size but the same applies to the actual video and audio streams when extracted with MKVToolnix.

Frankly I am little worried from this rather strange behavior, especially due to the fact that I am intending to buy one of these tools for converting ~500 of original BD discs to .mkv - Is it possible that makemkv injects some kind of copyright watermark to the converted files for future usage or bytecopy partially "drops" chunks of data when converting to .mkv?

MKV rip sizes will be different between all tools on the market. They all have their own header structures, which differ, thus a different number of bytes. Most of them attempt to detect and remove the zero padding found on discs, and do so with varying degrees of success. Some programs include the random number that is supposed to be generated per the MKV standard, others do not. Different programs handle disc authoring errors differently. Etc. etc. etc.

No two programs will give you the same file size. And no program is necessarily the "correct" program to use.

For example, I mostly use MakeMKV for my rips. But I recently ripped Burn Notice Season 1. It has a disc authoring bug on the DVDs. MakeMKV doesn't appear to detect the bug, and rips the disc as it normally would (at 29.97 fps, as flagged on the disc). The result is that the audio on every rip will get out of sync after about 30 minutes. However, if I extract the elementary streams from the disc, and then make a MKV using mmg, it will detect the error and encode the MKV at 59.94 fps (vs 29.97 fps), which allows the audio to stay in sync. If I force mmg to use 29.97 fps, it produces the same sync problem as MakeMKV.

If you intend to just use one program to make mkv files, then be prepared for a world of frustration. I currently have ripped about 500 movies and 2000 TV episodes and have had to use no fewer than four methods to successfully rip everything I've came across.