A Newbie's Research: A&E's Technical Guide for Dummies (Like Me)

I'm a relative newbie to the world of vidding. I've made 4 vids in Windows Movie Death (as we call it at home), but nothing I'd try to pass off to a blind person as high-quality (they all have the DivX logo, for crying out loud). I knew next to nothing when I made these, and I vowed to take the time to do some research before I made any more. I owed it to myself to do this right if I was going to do it at all. But you know how there's a vast universe between WMM and Premiere? I was lost in that space. I get the feeling that a lot of vidders were stuck in that in-between, but there is little being said about it. Most of what I can find in the way of online vid discussion is either for WMM or Premiere. It's like in Musical Chairs when there are only two chairs left on opposite ends, and if you're going to leave the WMM chair, you have to run like hell to get to the Premiere chair before the music stops. That's how I felt sometimes, anyway.

Since I'm not ready to run to that Premiere chair yet, I decided to take control and find a way to be happy in that space between, not lost in it. WMM is such a hand-holder that it can be scary letting go. To start, I decided to upgrade to one of what I consider "intermediate" editing programs (Ulead, Pinnacle, etc, which are linear like WMM, but with far more features, options, and required knowledge to maximize them.) I also started to read everything on vidding I could get my hands on. Fortunately for me, I managed to stumble upon A&E's Technical Guides to All Things Audio and Video by ErMac and absolutedestiny. I know I don't speak only for myself when I say that it rocks my little vidding world. It gave me both the tools and know-how to make some quality vids if I was willing to put some work and thought into it. I dove in headfirst. Thanks to their help, I've gone from having the video savvy of a green bean to having the savvy of...well, a green bean who at least understands the difference between interlaced and progressive video. I can not stress enough how much gratitude, appreciation, and respect I have for these guys for providing the vidding community with such an amazingly helpful resource.

Because my day job could actually be performed by a green bean, I have lots of free time on my hands to research vidding online here at my desk. I've spent the better part of the last three weeks reading A&E's guide from top to bottom, reading vid meta posts in this community, watching rec'd vids, going back to A&E's, playing with vid program trials at home, starting my own vidding sources masterlist, going back to A&E's, repeat as desired, yada yada. I also started extracting quotes and sections from A&E's for my quick reference (Cliff's Notes to A&E, if you will), and upon doing so, things started making even more sense. I highly recommend doing this. By arranging things into the order my own mental framework best understands, all kinds of lightbulbs were switching on all over the place. I chalk it up to the different learning styles people have, not some failure on A&E's part to successfully communicate something. After all, they just put the information out there; how much we already know or at what speed/order we read it is completely out of their hands. They've done their part, (much more than their fair share, if you ask me), and now it's up to us to do something with it.

At the same time, the site can be a bit intimidating for someone like me who was just getting started. On the offchance it'll help some other newbie out there, I figured it probably wouldn't hurt to type out what I've managed to piece together and put it where someone other than myself could see it. Don't see this as any kind of replacement, disrespect, or insult to A&E--that's the last thing I'm trying to do. Their site is my personal vidding bible, and I'm an aspiring missionary to help those who were lost like me. I'm not going to say anything here that they haven't already said in one way or another. I was a TA in college, so it's in my blood to stick around after school to present the class material in a different way for the other folks who, like myself, didn't quite get it the first time for whatever reason. It's also entirely possible that I'm the only one who had trouble wrapping their head around this stuff at all, and this post is just for me. I'm fine with that. After all, this post just covers the sections I had a rough time with at first.

DISCLAIMER: In no way, shape, or form am I trying to pass myself off as any kind of expert. Green bean here, remember? Very green. It is entirely possible that I've got some things wrong. Please, please correct me if I do. I'm surprisingly humble about these things. Otherwise, what would be the point?

First of all, everyone should open up A&E's guide. If you're new to the site, start at the top. It's extremely well-organized (especially considering how much information is in there), and it's quite linearly structured from top to bottom. It's thorough and well-explained, and it's worth digesting in full. If you don't read the Theory Guide in the Video section, you're going to be lost. (Yes, I'm speaking from experience here. I'm a honest green bean.) They do a great job of giving you the quick-and-dirty one paragraph versions of the highlights and buzzwords before digging a little deeper. Read it, then go back and read it again, then read it again. And when you're done with that, go grab a glass of juice and a bagel and read it again. Do not venture below the "Producing the AMV" line, do not pass Go, do not collect $200. At least finish your juice first.

Once you've got a decent grip on the basics, go ahead and soak up the next section, The AMVapp, AVIsynth and VirtualDubMod. Install the AMVapp, read about AVIsynth, read about VirtualDubMod. They really laid this out well. Play around with the programs and make sure you have a basic understanding of what they do. If you're going to be doing any video editing beyond throwing downloaded torrents into Windows Movie Death, you're going to need VDM. You'll get the most out of VDM by using Avisynth. Those, and all the other pieces you'll need to use the guide, are in the AMVapp. It's as simple as that.

Up to this point, I probably haven't told you anything you don't already know. And up to this point, we're all more or less walking the same vid path. It's from here on out where things start to get confusing, which is why there are so many posts asking for help getting started. But sometimes, being pointed to a certain tutorial or tech guide isn't enough. Maybe it's too specific (only for software you're not using), maybe it's too general (doesn't even mention software), or maybe you just need to bridge a few gaps between what you know and what you're reading.

A&E's does a great job of describing different places you can get your source video from, how to rip from dvds, and lists the many steps involved in getting it from raw footage to a finished vid. But what I was having trouble with as a newbie was sorting out the various combinations of options that get us from A to B. To me, this site is a bit like those Choose Your Own Adventure books I loved as a kid. You constantly have three options to choose from, each one leading you to three more. Since I'm a "big picture" person, I very much appreciate having that list of all the steps involved and seeing a breakdown of the basics. I love that page and refer to it often. But the piece my brain was missing to put it all together was seeing how these basic components connect to each other across category lines as well as within them. A&E's guide does explain the many variables that go into making a quality vid, (far more than WMM would lead you to believe are worth paying attention to!) For example...

A Few of Many Variables/Buzzwords to Consider:
+ Video Type
+ Framerate
+ Interlaced vs Progressive
+ Editing program
+ Codec for editing
+ Aspect Ratio
+ Window Size
+ Colorspace
+ Codec for distribution

And the list goes on! All of these are taken from the list of steps A&E take you through, but I think it's helpful to look at these from afar before diving in headfirst. You see, the catch is: more often than not, what's "best" in one category of variables doesn't work with what's "best" in the next category. For instance, it's entirely possible that you can have clean, beautifully ripped dvd footage and a functioning editing program, yet the two aren't compatible (without some tweaking in Avisynth first). Why would that happen? Quite simply, a mismatch of variables in our list. They're not all friends, you see. Let's break down the list a bit. In some cases, such as codecs, I'm only including what A&E say are worth using.

A Few of Many Variables/Buzzwords to Consider:
+ Video Type (NTSC, PAL, FILM)
+ Framerate (frames per second, fps) (23.976, 24, 25, 29.97)
+ Interlaced vs Progressive (edit interlaced, distribute progressive? edit and distribute progressive? edit interlaced, distribute both? etc)
+ Editing program (Adobe Premiere, Vegas, Final Cut, WMM, Ulead, Pinnacle, Premiere Elements, PowerDirector, the list goes on)
+ Codec for editing (DV, HuffYUV, Lagarith)
+ Aspect Ratio (4:3 or 16:9)
+ Window Size (width of 720, 640, 512, 320, etc)
+ Colorspace (RGB, YUV--YV12 or YUY2)
+ Codec for distribution (Xvid, other lesser options)

And that's just the tip of the iceberg; I'm not even touching anything regarding editing techniques, clip order, or anything else within your editing program. This is just for preparing the source! It's no wonder we newbies run away screaming when we see what really goes on behind the scenes :-) Those vidders make it look so damn easy.

The trick is learning what variables go together and what don't. Some of these are more constricted than others, some will go with anything.

Before we jump into the juicy stuff, let's look at colorspace.

The guide includes a fantastic breakdown of RGB and YUV color. There are two kinds of YUV that we run into in vidding: YUY2 and YV12, the latter being more compressed and more prevalent. The main point to glean from A&E's colorspace page is knowing when we'll be in particular colorspaces during each step of the way. Because hey, guess what. Different programs like different colorspaces, and they'll switch on us when we're not looking if we don't tell them what we want. Just to make our lives a little more difficult.

We all start in YV12 if we're ripping from dvds. We'll (probably) be editing in RGB because that's what editing programs like. We'll (probably) be doing post-production with Avisynth, which means back to YV12 because that's what it likes.

As you can see, we'll have to do some converting along the way. Switching from one to the other doesn't make any noticeable difference that I can see, and it's not inherently bad, but multiple conversions back and forth and back again can lead to fading. So. It's good to be aware of when your files are in a particular colorspace so you can minimize conversions. You're not going to be able to avoid them altogether, but you can control how often you're losing color.

According to A&E, the optimal scenario involves only these 2 colorspace conversions:
1) DVD in YV12, processed with Avisynth in YV12 and then converted to RGB32 ready for editing. (YUV to RGB)
2) RGB export from editing program using HuffYUV codec (in RGB mode), then converted to YV12 for the final video compression. (RGB to YUV)

Sounds easy enough, it's just another thing we have to think about in this craziness. :-)

One other thing to keep in mind: If you're going to be ripping AVI clips to edit with, you have the option of selecting Lagarith or HuffYUV as your codec. Lagarith can store the original YV12 color from the dvd rip, but HuffYUV needs to be in RGB or YUY2. Since your editing program is going to kick out an RGB product anyway, it's up to you if you want to put YV12 or RGB into it. You'll be doing a YUV-->RGB conversion either way, it's just a matter of when (and who) you want to do it. Choose Lagarith and you'll work with smaller (but slower) files, leaving it up to your edit program to convert the vid to RGB on its way out the door. Choose HuffYUV and you'll convert to RGB before you reach the editing room, working with larger (but faster) files, and what you see onscreen is what you'll get when you leave. It's up to you. To be clear, Lagarith can store RGB, too, but since HuffYUV is generally a better codec, there isn't anything to gain by doing RGB in Lagarith.

Now. Take a deep breath. It's time to get into the messy part. As a newbie, this was the hardest thing for me to piece out: Interlaced vs Progressive Footage

Here's somewhere to start. I'm going to color-code this from here on out to help reduce confusion (those numbers look awfully similar at first!) Your dvd source is (generally) going to be one of the following:

A) 23.976 fps Progressive FILM
B) 25 fps PAL (note: I do believe PAL can be either interlaced or progressive, but I'm not positive.)
C) 29.97 fps Interlaced NTSC

You should go to the tech guide to read about the difference if you haven't yet. According to A&E, choice A is your higher quality option, choice B isn't found much in the USA, and choice C can be problematic in editing and internet distribution. "Fine," you say, "I'll go with A. I want my vid to be all that it can be."

Sadly, it might not be that easy. If your source does turn out to be encoded as progressive, then awesome, you might be well on your way. (You can find out if it's progressive or interlaced by looking at the "preview" window in DG2Index when you load your vob files.) If your footage is interlaced and you want it to be progressive instead, you'll have to use IVTC (Inverse Telecine) or similar to make it so. Either way, you'll eventually end up working with 23.976 fps progressive footage in the editing room.

The kicker is that Adobe Premiere doesn't like working at 23.976 fps, and other editing programs might not, either! All editing programs are built to work with 29.97 fps interlaced NTSC footage, though the editing can be messier than with progressive. A&E say the best editing codec is DV (it's lossy without sacrificing quality), but it also only works with 29.97 fps. Seems a bit silly, doesn't it? To put it plainly: the "best" footage type to edit with might not work with the "best" editing programs or the "best" editing codec. And remember that dvd footage encodes in YUV colorspace, and you have to be in YUV to run Avisynth scripts, but editing programs work in RGB! We have our work cut out for us. All the same, I found myself less concerned with the technical details about what makes up each type and more concerned with the general whats and whys about using one type over another.

So what are the perks of each footage type? Let's break it down like this:

+ 23.976 fps Progressive FILM: creates higher quality source footage, higher editing ease/quality, used for computer playback
+ 29.97 fps Interlaced NTSC: works with the best editing codecs, better compatability with editing programs, used for tv playback

Put another way, here's what's preferable/easier/better at each step of the vidding process:

1) Ripping high quality source footage
2) Save with codec for editing purposes
3) Create a new project in editing program
4) Worry-free editing re: effects, moving frames around
5) Save with codec for internet/computer distribution purposes
(optional) 6) Save with codec for tv/dvd/convention distribution purposes

It looks like a damn Christmas card. And just for kicks, let's throw the flow of colorspace conversions on top of that.

1) Ripping high quality source footage (YV12)
2) Save with codec for editing purposes (YV12 --> RGB)
3) Create a new project in editing program (RGB)
4) Edit much more worry-free re: effects, moving frames around (RGB)
5) Save with codec for internet/computer distribution purposes (RGB-->YV12)
(optional) 6) Save with codec for tv/dvd/convention distribution purposes (RGB-->YV12)

What gives? Is a little consistency too much to ask for?!

But this isn't a lost cause, not by a long shot. Remember, that's not a breakdown of what we're actually going to use at each stage, just a list of which type is "easier to work with" (or "has less compatability restrictions" or however else you want to spin it) at each stage. The actual vid path we walk is much more consistent. Whether you decide to edit interlaced or progressive, you can get from start to finish with minimal yo-yo-ing back and forth. There isn't a "good guy" or "bad guy" here; both roads are valid, and both have their pros and cons. You have powerful tools like VirtualDubMod and Avisynth at your side, and they're going to do everything they can to help you.

To tie it all together, let's go back and look at that list of variables. Based on A&E's recommendations for each step of the way, how do they all piece together?

Your base combinations were:
A) 23.976 fps Progressive FILM
B) 25 fps PAL
C) 29.97 fps Interlaced NTSC

For A) 23.976 fps Progressive FILM:
+ Video Type: FILM
+ Framerate: 23.976 fps or adjustable to 24 if necessary
+ Interlaced vs Progressive: source is progressive, edited as progressive
+ Editing program: almost all (note: Does anyone know of software that doesn't work with 23.976? There must be some.) Remember, in Premiere you have to change to 24 fps first.
+ Codec for editing: HuffYUV, Lagarith
+ Aspect Ratio: Can be either 4:3 or 16:9
+ Window Size: Any size you want
+ Colorspace: either YUV or RGB so long as your codecs are compatible
+ Codec for distribution: Xvid or other lesser options (as laid out by A&E)

I'm not going to address B) 25 fps PAL specifically here because honestly, I don't know squat about it. But you can find the bits and pieces in each relevant section of A&E's guide if you need.

For C) 29.97 fps Interlaced NTSC:
+ Video Type: NTSC
+ Framerate: 29.97 fps
+ Interlaced vs Progressive: source starts as interlaced, but can be edited as either
+ Editing program: All available programs. Interlaced is the universally compatible editing method.
+ Codec for editing: DV, HuffYUV, Lagarith
+ Aspect Ratio: Can be either 4:3 or 16:9
+ Window Size: Any width you want, but never change the vertical size of interlaced footage. If you need to resize, go progressive.
+ Colorspace: either YUV or RGB so long as your codecs are compatible
+ Codec for distribution: Xvid or other lesser options (as laid out by A&E)

The differences are subtle but important. You can (and should) read about the pros and cons of progressive vs interlaced editing at the guide.

The best way to decide what path you're going to take through these woods is to look at your bottlenecks:

+ how the footage is encoded
+ what editing program you're using
+ how much effort you're willing to put into making this work

Fortunately, most editing programs can work with either 23.976 fps progressive or 29.97 fps interlaced, but you need to investigate this on your own. Open a blank, new project and see what the options are. If you can work in either framerate, then you should be fine. And for those of you working with Adobe Premiere, remember A&E said don't trust their 23.976 fps setting. You can convert it to an even 24 and go from there. If you're not willing to do even those steps, then you've found your next bottleneck ;-)

But really, the ultimate dealmaker is whether your footage is progressive or interlaced. Without knowing that, you're not going to get anywhere positive in your editing program no matter which product you're using. If you ignore this distinction from the get go, you may very well regret it later. Undoing interlacing is much more difficult on a completed vid than fixing the source beforehand. It's best to take sixty seconds and use DG2Index or VirtualDubMod to determine what kind of source you're dealing with, then go from there. If it's progressive, you're free of a fair deal of headache, but your options have narrowed. If it's interlaced, the compatability field is wide open, but you have a lot of thinking, decisions, and extra steps ahead of you. Now we can start piecing together combinations of options.

1) 23.976 fps progressive FILM source, edited as progressive
2) 29.97 fps interlaced NTSC source, converted to 23.976fps progressive FILM, edited as progressive
3) 29.97 fps interlaced NTSC source, edited as interlaced

Note that we don't have a fourth option of converting progressive footage source to interlaced (reverse of #2). It is possible to do that, but there isn't a practical reason why you'd want to for vidding unless (for whatever reason) you prefer to edit and distribute interlaced. As far as the framerate is concerned, it is possible to speed up your progressive footage from 23.976 fps to 29.97 fps to accomodate your editing program (if it won't do 23.976 fps), but it's not recommended. Speeding up footage is more noticeable in live-action than slowing it down. If you do have to speed up a video to 29.97 fps for convention distribution, that should be done at the end. (Messing with the finished vid, not the source.)

One quick note: if you have a hybrid source (part interlaced, part progressive), there are plenty of options for handling it in the guide.

Now that you've determined if your source is interlaced or progressive, and decided if you're going to edit it interlaced or progressive, what does the road ahead of you look like?

1) 23.976 fps Progressive FILM, edited as progressive
-Rip progressive footage from DVD, using "Force FILM" option if preview reads between 96 and 100% FILM (disable audio stream to save space)
-Deal with any aspect ratio problems
-Decide if you're going to work directly with AVI files in your edit program, or instead work with Avisynth files in your edit program (note: for linear editors like WMM, Ulead, Pinnacle, etc, working directly with Avisynth files is not an option)
-If making AVI clips, convert from YUV to RGB using Avisynth in VirtualDubMod, and export using HuffYUV or Lagarith
-Edit vid in whatever program you have
-Export using HuffYUV or uncompressed without audio stream
-Reattach audio stream and convert to XviD AVI in VirtualDubMod

Things to note:
+If at all possible, this is the all-around "preferred" method since it avoids the potential messiness of interlacing altogether.
+If you're going to distribute for tv/conventions/dvd, you may need to remaster a 29.97 fps version using Avisynth


2) 29.97 fps Interlaced NTSC, converted to 23.976fps Progressive FILM, edited as progressive
-rip interlaced footage from DVD
-Convert your footage to progressive using an Avisynth script found in the guide. Almost all the scripts will give you 23.976fps footage in the end.
-Check the field order and tackle aspect ratio issues
-Decide if you're going to work directly with AVI files in your edit program, or instead work with Avisynth files in your edit program (note: for linear editors like WMM, Ulead, Pinnacle, etc, working directly with Avisynth files is not an option)
-If making AVI clips, convert from YUV to RGB using Avisynth in VirtualDubMod, and export using HuffYUV or Lagarith
-Edit vid in whatever program you have
-Export using HuffYUV or uncompressed without audio stream
-Reattach audio stream and convert to XviD AVI in VirtualDubMod

Things to note:
+This is the "preferred" method for dealing with interlaced source footage.
+Other than converting to progressive in the beginning, this is identical to method 1.


3) 29.97 fps Interlaced NTSC, edited as interlaced
-rip interlaced footage from DVD
-Check the field order and tackle aspect ratio issues, keeping in mind never to change the vertical size of interlaced footage (unless you use the provided script)
-Decide if you're going to work directly with AVI files in your edit program, or instead work with Avisynth files in your edit program (note: for linear editors like WMM, Ulead, Pinnacle, etc, working directly with Avisynth files is not an option)
-If making AVI clips, convert from YUV to RGB using Avisynth in VirtualDubMod, and export using DV, HuffYUV or Lagarith
-Edit vid in whatever program you have
-Export using HuffYUV or uncompressed without audio stream
-Reattach audio stream and convert to XviD AVI in VirtualDubMod for tv/convention distribution
(OPTIONAL:) -Convert back to YUV and remove any interlacing in post-production to export for internet distribution with XviD

Things to note:
+This method is only necessary if your editing program can't process 23.976 fps footage OR you don't require internet distribution format.
+If you're going to distribute online, extra steps must be taken to make your vid progressive.
+DV is an available export option now that we're dealing with 29.97 fps footage.
+If you're going to use any Avisynth scripts in post-production, you'll have to convert the color again.


Now that I've pieced out these three roads and what my options were at each step of the way, I feel much more confident moving forward in my new editing software. If I've taken all these factors into consideration, I should get a great looking vid. I'm going to hand it back over to the real teachers now, but please be sure you check out their sections on using Avisynth scripts to improve visual quality and compressing your audio and attaching it in VirtualDubMod, two important steps I've completely skipped over here. I think I've rambled on long enough for the time being :-)

Thanks for reading. Class dismissed.