Adventures with DL disks

[Spoffo]

A month ot two ago, I started backing-up some of my DVDs on DL disks. Yeah, they're pricey, but for some of the ones with terrific extras, it was worth it to have a complete clone of the original.

I started with high-quality (Verbatim) DVD+DLs, and the process was simplicity itself. Rip the whole original with MTR, and then burn the complete VIDEO_TS file with Toast. Perfect results every time.

However, the Verbatim blanks are around $6.00 each, and I could get Memorex blanks for close to $3.00. So I decided to try Memorex, despite their "mixed" reputation for quality.

The initial results were just that - Mixed. Some copies played perfectly, others would start off fine, but then have playback problems - ranging from just a few "sparkles" to bursts of serious macroblocking to spasms of freezing and breakup that made them unwatchable. As I was about to write-off the whole experiment, I realized that there was a pattern: If a disk had a problem, it was always somewhere around the middle of the feature; never near the beginning and never in the menus or extras. And, on further checking, the problem section never lasted more than 5-6 minutes, after which the rest of the disk was just fine.

This led to some detective work with DVD2oneX and MyDVDEdit which revealed that the bad section on the Memorex disks was ALWAYS in the first 5-6 minutes after the layer change. In other words, the quality problem with Memorex DL blanks is inconsistent material at the very beginning (i.e. the outer edge) of layer #2.

[I would love it if someone here with more understanding could explain how the layer break is programmed and recorded in the cell tags. On some originals DTOX will cleanly report which cells of the feature are on each layer, but on others it will say that everything is on layer 1, no matter what. Likewise, MyDVDEdit will sometimes show a Command 1 (which I understand means nothing) at the layer break, or show a change in the cell numbering scheme. Other times it doesn't. And, extras seem to always be reported as being on layer 1, even when the files sizes mean this is impossible.]

But I digress.

Following this lead, I went back and checked the Memeorex burns that had come out perfectly. With one exception, they were all the smaller disks, with total size between 6189mb abnd 7446mb. The problem disks were all larger, ranging from 7593mb to 7873mb. That makes sense. A smaller total size would usually mean an earlier layer break - i.e. in from the edge, avoiding the bad material.

So, it would seem that the strategy is clear: 1) rip the whole original, 2) if necessary, compress with DTOX to get the total size to around 7.4GB 3) burn the (slightly) compressed file. I did that a few times, and it worked great. I was able to get clean back-ups of disks that had given me problems the first time around. I didn't have to throw out the rest of that $80 Memorex cakebox.

However . . . . there are a couple of items that have me thinking that it's more complicated than just a matter of file size.

1) One of the disks I made early-on that plays perfectly is Shindler's List, and it's the biggest file size of all - 8015mb. However, because the original is a 2-sided flip-to-play, I had made an uncompressed "join" of the feature only, using DTOX and burned the DL from that.

2) In one of my early experiments, I took a file that was 7757mb in the original and gave it a teeeny compression in DTOX to 7736mb. The burn from the original had terrible breakups after the layer change, but the slight compression played perfectly, even though it was bigger than several of my problem disks.

In both of these, it would appear that the fix wasn't the reduced file size, but the fact that they had been altered, maybe because this gave Toast more power over how the layer change was done? (This is now way beyond my understanding of how these things work.)

But size must matter, too. Otherwise, how would you explain all those smaller disks - which were never touched by DTOX - that played perfectly, as well as the fact that even the biggest files burn just fine without DTOX when you use the more expensive DL blanks.

Anyone with a deeper understanding of this stuff care to venture an explanation?

[pants]

i wonder if it matters to maybe burn from disc images instead of the video_ts folder? it sounds like now you're burning from the ts folder, right? can mtr make disc images instead of just the ts folders? i've tried to rip like this from mtr before so i don't know.

[rumplestiltskin]

The beginnings of a great thread! I'm getting ready to buy some DL's and this will be most educational.

[dcsos]

But shouldn't we preseve the original layer break, not generate our own where the "halfway" point is in the file weight, which is what you seem to be getting?

Off topic: but perhaps illustrative as per PANTS suggestion to use an image (TOAST IMAGE, DISC COPY IMAGE?)
on the PC side the layer break is maintained by RECORD NOW by SONIC SOLUTIONS
but re-done in the middle by NERO

https://www.videohelp.com/forum/viewtopic.php?t=225001&postdays=0&postorder=asc&highlig...0break&start=0

[Spoffo]

On most of my disks, the position of the layer change with regard to the program material seems to be locked-in by the original authoring, even if you compress it. In other words, if you shrink a file, the layer break will still be in the exact same place - - say, 04:15 into chapter 12 - - as the orginal disk, even though the file sizes for all the cells and chapters have changed.

The one disk I've burned where this isn't true is Shindlers's list, where the original is 2-sided (single layers). On the joined back-up, Toast put the layer change at the beginning of chapter 20, creating an almost perfect 50/50 split between the two layers 4019mb and 3995mb.

As I think about it, this may be why the back-up plays perfectly despite its large size: This even split keeps the change-over point somewhat father in from the edge than other large disks with a less even spilt.

I'll continue to do some systematic experimentation in the days ahead and report back. Unfortunately, since this involves playing around right near the area on the Memorex disks where we know they have a product problem, it may be tough to sort out the "science" from disk-to-disk quality variations. (And, even at only $3 a disk, I'm not interested in trying for statistical validity!)

More later.

[Spoffo]

One more thought. looking at those layer sizes, I realized that to get a smooth layer change (or maybe any change at all) layer 2 MUST be smaller than layer 1, so that the read laser can stay in the same position on the disk and just start reading outside-in. If layer 2 were larger, the read head would have to physically jump outward from the end of layer 1 to pick up the start of layer 2 (and unless you had it reading empty tracks, would this even be possible?)

Does anybody know for sure that this is true?

[AntnyMD]

The buffer should handle any large gap between layers when layer one is smaller than layer two. I surmise the layer change glitch sometimes observed is a combination of a too small buffer on the player's end, and funky layer change placement in authoring. I havent noticed layer change glitches in years -- not since the early days of DVD.

Keep in mind though most folks who are burning DL are trying to maximize the size of layer one so are trying to fill up as much of that first 4.4 gigs of space (making layer two nearly always smaller than layer one -- layer two can hold only 4 gigs).

Are you getting these $3/disc DL media online or in a store? Around my place, the Memorex are only about $1/disc cheaper than Verbatim, so I go with the big V.

[Spoffo]

Out here, CompUSA has had an advertised special at least twice on the Memorex DVD+DLs (they're actually Ritek D01) at $79.95 for a cakebox of 25. They seem to come up about every 2 months.

Also, several of the online retailers offer the same Memorex and/or Ritek pack for $100 plus a $20 rebate.

[pants]

a few weeks ago compusa also had 3 packs of the memorex dvd+r dl's for $9/each.