Video: [R2J Dragon Box](https://nyaa.iss.one/view/1133179) (w/ corrected B part) Audio: [Sync'd Broadcast Audio](https://nyaa.iss.one/view/1240974) (re-encoded with AAC for space) Subtitles: Machine-timed & re-styled DVD subs (unclear who OCR'd the copy I started with) Chapters: [Courtesy of Utsukushii-Raws](https://nyaa.iss.one/view/489801) Now that the Dragon Ball broadcast audio has been shared and sync'd I've been slowly watching the series. And I've been slowly working on making a release that I find easier to watch. Mostly by machine-timing the subtitles to have fewer egregious bleeds across scene changes, as well as correcting obvious OCR errors or typos as I watch. I'll share the release in its entirety eventually, once I've finished watching all the episodes. But I wanted to share this episode before I finish because it has unique issues that make it less watchable than most of the other episodes. Overall, the Utsukushii-Raws release of Dragon Ball does a great job of capturing what's on the JP DVDs. Unfortunately episodes 26 and 41 have errors on the DVD, and those are preserved in the raws. The B part of both episodes is telecined and some of the fields have blended chroma. Thanks to limited animation, many frames can be almost perfectly restored by simply borrowing from an adjacent duplicate (e.g. lip flaps), but action frames cannot be recovered so easily. For frames which simply pan without foreground motion interpolating the chroma works quite well. For frames with both background and foreground movement, or quick action, the interpolation doesn't work as well, and I instead opted to color match the chroma from the [Funimation DVDs](https://nyaa.iss.one/view/1139811) for these frames (as the R1 DVDs do not have these chroma errors present). Unfortunately the R1 DVDs have less visible frame than the R2J counterparts, so there remain some, filtered, chroma errors at the edges (and on dark surfaces where the R1 DVD was crushed), but they are _much_ less intrusive and distracting, especially when watching in real time. And of course colors may not match perfectly, and chroma alignment certainly isn't perfect, but I'm not an artist and I didn't want to paint the individual frames, so this is the best I can realistically do. I don't think still images do a perfect job of capturing the difference, but here is [a comparison to illustrate](http://www.framecompare.com/image-compare/screenshotcomparison/7YKZWNNX) generally what all of this means for impacted frames.