Randon Pixelation

[kougei]

Hi, I am new to mkv. Many years ago I embarked on a project to rip lossless episodic DVD and BD to Emby Server for Netflix-Style on-demand playback. I ripped DVD using CloneDVD and BD using AnyDVD and BDRebuilder, but project faltered due to time.

I have a Ryzen7 2700X/32Gb Workstation 64-Bit Windows10, Asus BW-16D1HT (3.02fw) and end-goal will be direct NFS share playback via DuneHD 4K. I was not quite decided on mkv episodes or ISO Rips, but have tried both methods with the latest MKV build. I am favouring 1:1 full ISO as it seems a much quicker process, and Emby Server now has limited support for indexing full ISOs, and storage is so cheap compared to when I last did this.

Using Makemkv in default options, on both BD Backup and MKV episode rips, I see random pixelations on file playback. When I say random, I mean the pixilation can occur in different places when ripping the same disc with same settings over again. But on an output file, the pixilation on playback is always in the same place, and can be replicated on network playback via Emby App on Xbox/AppleTV, Panasonic UHD Player, as well as the preferred DuneHD device. So the issue seems to be inherent pixilation/artifacts within the created video file. I have tried cleaning/wiping the discs - 3 separate discs tried thus far, and disks play flawlessly in the Panasonic UHD Player.

So I reverted to AnyDVD ripping to drive instead of Makemkv, and also BDRebuilder - I have the same pixilation issues on those methods as well. So it's not a specifc Makemkv problem, but as it's the most popularly used tool, I'm still asking for advice here.

Looking though posts on the forums, I have already troubleshooted as follows with no change :-

1. Tried a different BD Drive - Pioneer BDR-206D - same issues
2. Tried changing target Output Folder from underlying Hard Drive to SSD Drive - same issues
3. Tried Backup to Drive before running MKV or BDRebuilder to output mkv files - same issues
4. Used Single Drive Mode and BD+ dumps in Expert Mode
5. Ran Windows Memory Diagnostic Test - No Issues at all

I have seen other references to limiting CPUs in some way, and running MD5 checksum on original versus output files, but it wasn't clear how to do this. Any suggestions or other ideas herein?

[Woodstock]

If it moves around, it's not the file, it's the playback environment. And usually that comes down to transport between storage and player. If the bitrate of a file exceeds what the transport can deliver RIGHT NOW, the video decoder is the first place that falter.

[kougei]

It's a Gigabit Network between server and player, hard-wired, Also MKV Rips from many years ago still playback perfectly. I can try playing it back from a USB Stick though in the Dune, will try that now.

[kougei]

To be clear as well, it doesn't move around, the pixilation occurs at the same time, on multiple playback devices, on the same output file, so is inherently repeatable on that file. But if you then re-rip the same file from the original disc, the pixilation will occur on other random scenes, and not the same ones as the previous file.

[SamuriHL]

Then you're getting read errors that aren't getting caught. That seems bad.

Sent from my Pixel XL using Tapatalk

[kougei]

Now that is interesting, it plays fine from the USB Stick. I will try a fresh rip from a different disk just to double-prove the case. You have certainly given me a big clue as to where the problem could lie, so thankyou. I would still have thought it odd that the pixilation artifacts are in the same place, on repeat viewings via the network, and repeatable also on different playback devices, which is why I never considered it could be related to networking. So I guess the area of investigation is solely on the networking setup on the serverside. Maybe I will try from Samba shares instead of NFS as the first investigation.

[SamuriHL]

So if you put it on a USB stick it plays without error? But playing it back over your network shows errors in the same spot? That's not a rip issue if that's the case. That's rather fascinating.

Sent from my Pixel XL using Tapatalk

[kougei]

Hi, unfortunately, having gone full pelt into ripping BDs, it now seems that the USB versus Network diagnosis was faulty. In further testing I have examples where what manifests as pixelation/blocking at points in the playback, if I play the same file using Videolan on a Windows desktop, the same place shows as a massive skip/jerk, so I am back to being inherent flaws in the rips. I have 2xBD drives in the PC and a rerip with the older drive to a different hard drive, doesn't have issues. Is there anyway to do a RIP/Backup to a drive, then compare the ripped BD folder/files to another copy, and see a difference? Something like an MD5 checksum, but this can only be done on individual files rather than the ripped directory? Then I can experiment with different disc-drive/hard-drive ripping combinationsm

[Kettykrueger]

I've had this problem for a while now, and like you, couldn't find the root cause.

Did you get any where with your diagnostics?