I'm writing this concurrently as I read through the Afterdawn article that you've linked to...
Right away, I have to disagree with what's being written in the Afterdawn article. TMPGEnc was written by a hobbyist for a hobbyist, it did not start out as a piece of software that was geared towards the lowest common denominator, i.e. the lay consumer who is unlikely to know anything at all about video. TMPGEnc 2.5, the "original" encoder, had many settings that were selectable, but no explanation was given -- you had to know about video to know how to properly set them up. Missing such an obvious historical aspect makes me cringe, and I wonder if it will affect their article as I read through it.
Ah-ha, the guide both contradicts and corrects itself a few paragraphs later. I wonder if this was written by one person?
Sentences like "depending on what your purpose in encoding an MPEG file is, the range of options supported for playback may be more limited than the MPEG-2 specifications allow" don't make sense. I can only assume this means that the templates are sometimes too limiting for a project, meaning you need to manually set up the encoding parameters. I would agree with that. In fact, I would suggest NEVER using the templates, as the "one size fits all" approach tends to fit none.
They're making some mistakes again. Very often an .MP4 file can be renamed to a .AVI file, and TMPGEnc will see the data. The correct system-wide (DirectShow) codec simply needs to be installed.
I honestly have no idea why somebody would want to re-encode a DVD-Video file set. I can understand if it is being done for restoration purposes, but this guide makes it sounds as if this is a routine choice in source. No wonder so many people create atrocious quality videos, they are butchering it with needless re-encodes.
I'm having to cringe again --- "Although you can change the resolution using TMPGEnc XPress, you may want to do it ahead of time, either by encoding an intermediary file with another tool or using AviSynth to resize on the fly" --- That is awful advice. Did they honestly suggest yet ANOTHER re-encode of the video, simply to change the resolution? Does this person writing this actually know what he/she is doing? (Yeah, I'm being a bit mean now -- hopefully you're not the author!) Is this video supposed to be viewed on an 8" portable DVD player where quality is not important?
The Aspect Ratio information is a bit butchered. For example, you can only use 16:9 when the video is 720x480, at least if you want to be 100% DVD-Video compliant. My DVD recorder can make 352x480 DVD's, but only because it uses the DVD-VR specification for decisions, and not the DVD-Video one. Another example is 4:3 is the standard "full screen" traditional tv size, while 16:9 is the widescreen tv size. All that other stuff they mention is confusing and needless.
Framerate info is accurate.
The rate control information is way off-base. CBR is simply constant bitrate. The bitrate is continuous at the settings selected. At a high bitrate, the video gets bitrate whether it needs it or not, not maintain image quality. At a low bitrate, the video has a ceiling and quality may suffer. VBR is a "smart" method of encoding, putting bitrate only where it is needed. The description of "allows higher peak bitrates for more complex scenes by limiting bitrate in other places" is very misleading, because the bitrate only goes to the maximum or minimums you set. It will not magically go "higher" just because it is needed. If you set the maximum too low, then it still will not get the bits adequate enough to maintain proper image quality. VBR was simply created as yet another way to compress the video data. For example, you can put 2 hours of good-quality 720x480 DVD-Video on a single-layer disc with VBR, but you could not do so with CBR. The specs for a VBR disc would be 8000k max, 5500k avg, 0-2000k min, but CBR would be 5500 constant. The CBR would suffer when it needed a higher bitrate, and be throwing away bits when not. The VBR would give and take, as needed, to allow for image quality to stay as close to perfect as possible, within the limits of an 8000k ceiling.
Setting up an encode size based upon the blank disc to be used can be a bad idea. It does not take menus into consideration, which can add 100-200MB or more, depending on how complex the disc is. It would also preclude "bonus material" of any kind. It is best to calculate all of the values in advance, and then select an encode size in MB based from that. Your bitrate largely controls file size anyway, so I'm not sure why this is in their guide. It's usually an "either/or" type of selection in TMPGEnc, where you either pick the bitrate, or you pick the disc size, but not both.
Always pick your video format -- PAL or NTSC -- never set it as "unspecified".
The DC information is not necessarily accurate, but it's been proved many times that such a setting makes little difference with the TMPGEnc encoding engine (Plus 2.5 version especially). Most encoders set it for 9 or 10.
The field order information should specify the "jerky" description as being something observed on an interlaced display device, such as a television set. It may go unnoticed on a progressive computer display, when de-interlace filters are in effect in the playback software.
Motion Search set to high.
They seem to be confused about what a GOP is. It is not a "display format" (how ridiculous), but rather another compression method. I-frame only video is temporally uncompressed, there is not intra-frame compression because an I frame is independently (hence the "I") encoded from other frames. It's the reference frame for the other frames, when you start to compress and include P and B frames. The I-P-B standard is what DVD-Video largely uses, although many commercially produced DVDs will do custom GOP settings based upon scene content. Studios do not open a video file like us consumers/hobbyists do, and cram it through an encoder in one fell swoop. A chapter mark on a DVD can only be inserted at an I-frame.
Open GOP used to be a problem for early DVD-Video authoring software, but not anymore. You had to close the GOP some years ago, for some applications. Having a closed GOP has both its benefits and its drawbacks, and it really does not matter in TMPGEnc, which is not the most robust encoder in terms of highest possible image quality. Sort of a waste of debate, given the circumstances.
DVD-Video requires a sequence header, end of story. It should always be checked, not "normally" be checked as the guide has stated. Only in non-DVD application would it be allowable.
Some honesty here: "details of exactly what the numbers mean are beyond the scope of this guide or the knowledge of the author". Matrices are the math behind the encoding setup. In general, leave these numbers alone. Some folks, like the xVCD/xDVD compressionist Karl Wagner (kwag), like to change these numbers so much that it starts to create MPEG files that are considered out-of-spec by DVD-Video authorware and DVD-Video hardware players. Much like GOP length, these are criteria that a studio would tweak to the content of the scene being encoded. It's not something a consumer or hobbyist should worry about -- and I honestly don't think they should even have the ability to touch it, as more harm than good will come of it.
The whole keyframes thing is neat, but I'd never use it. Most people would not need it.
Amusingly enough, the "Format Specific Templates" page appears to be correct, and in fact some of the information written here contradicts the incorrect information found earlier in the guide.
Ideally, you want your work folder and save folder to be on separate drives -- and neither of them would be C: (assuming C is the drive where you have Windows installed, and where Windows creates its page and temp files). These must all be internal drives, not USB/Fireware drives. The less drives you have internally, the less choice you have. The C: temp and D: save folder are okay if you only have the two drives.
This is not correct: "using MPEG for MPEG-2 is a personal preference, and either extension is accepted". A lot of Windows-based authoring software will not see an MPEG-2 file if it is not .MPG. This may work when TDA is the authorware being used, but other software will balk at it.
That's my analysis of what I read. Hope it helps.