Monday 8 August 2011

Anatomy of an Audiosurf video

Audiosurf is a great game, if you're a gamer, and music-lover such as myself. It's the only game I've ever recommended to people both in real life, and using Steam's recommendation thing, and I've been recommending it pretty much since I first got it.  Whenever I get a new computer, or do a system recovery on a current one, it is always the first game on Steam that I install. It is also the bread and butter of my videos on youtube.

Other than WoW, it's the first game that I started putting footage from up on the web. I use videos from it to keep my recently-imposed schedule of (at least) one video a day, as it's fairly simple to put an AS video together. So I figured I'd do a post, going through, step by step, how I put together a video for it these days, while I edit some videos of it.

First off, I'll launch the game itself, and I'll sit down and play through an entire album by an artist. I'm editing videos from four albums (one album each) by Celestial Aeon Project, Metatronik, Obsidian Shell, and Oleg Serkov. As I play through an album, I'll write down the names of songs where I got Clean Finish and Stealth bonuses. If you're unfamiliar with Audiosurf, those are just multipliers that affect your final score, that you get by finishing a track without having hit any grey blocks and having any hit blocks on your grid. This gives me a rough guide of which songs from each album I'm going to be editing.

Playing through an entire album, and selecting the songs that I did well on is a lot more efficient than how I used to select tracks for editing. I used to select a specific song, and record footage of that specific song. Sometimes it would be an easy ten minutes of recording, but a lot of the time, one single track could take upwards of an hour to record a good run, with minimal framerate lag from the recording software. I still do a song or two, here and there this way, (most recently, my nyan cat video) but I mostly do it the album play-through way now.

After I've recorded the footage, and made my notes about which songs might be edited, I go to the artist's Jamendo page, and the pages for the songs that I've selected. For what I'm editing today, the first video I'll be working on will be a song from Celestial Aeon Project's album "The Fall of Ragnaros". Looking at my notes, I see that I did well on the songs "Stormfront" and "To the Depths". I open these pages up for a couple reasons. The first, and most important reason is to check which Creative Commons licenses are being used for the songs. Some of the CC licenses prohibit the use of music in the way that I'll be using them. Usually, the same license will be used for the entirety of an album - like with this album - but sometimes, there is a different license being used for a song, and the song's page will reflect that. In the case of this album, the CC license that's been used means I can use these songs for my Audiosurf videos, as long as I credit to the artist (accomplished by a link in the video description), and don't make any money off of the use (accomplished by disabling monetization of the video on the Youtube (if I had a Partner account on there and was making money off my videos by default, which I'm not, so it's something I don't have to worry about right now). I leave those pages open until I'm done my editing process, to speed up the writing of the video description (as I link to the artist's page as well as the page for the track so people can download the song, and support the artist).

Moving on, now that I've looked at the licensing info for the songs that I'll be looking to edit, I open up the audio files for those songs in my media player of choice. I use VLC player - despite owning a pro copy of Winamp - for the sake of convenience, but which you use really has no impact on this process. Since Audiosurf videos - at least, in my case - are more about the music, than the skill of my playing (though that's still a concern, otherwise the songs I did poorly on would still be in the running) I want to choose the most interesting track to listen to. For this album, it's a relatively easy decision, as I only did well on two songs. Both songs are good songs, but to me, "Stormfront" is the more interesting track, so that's the one I'll be editing.

It is only at this point that I finally launch my editing software. Once it's launched, I import the folder that FRAPS outputs its footage to. I did a lot of recording lately, so I have 641 GB of Audiosurf footage sitting in there, as well as another 295 of Torchlight footage that I'll be editing in the future. Needless to say, the importing process could take a bit of time to do. I usually put the kettle on while I wait for it to import. I then select all of my Audiosurf footage and place it on my timeline. Yes, all of it. Looking at it on the timeline, I skim through until I get to the footage of the specific song I'm looking to edit, which in today's case is about 18 minutes in. Then, I select the video tracks that involve that song, and put them on the second track of my video editor. If I'm editing more than one song I'll go ahead and do the same thing for the other songs as well. After I've moved the clips that concern what I'm going to edit onto the second track, I remove all the other clips from the first track.

Once I've done that, I go back to the media room section of my editor, and select all the (Audiosurf) video clips that don't have a little green checkmark on their preview image, I right-click them, and then select "Delete from disk". Audiosurf videos are the only videos where I do this, as there is a lot of footage of the game, that I straight up, just won't be using, and the raw footage that FRAPS puts out takes up a lot of hard drive space. When editing footage from other games, there's a lot of back and forth in my file usage, and transitions, and sometimes even if I'm not going to use the video footage from a clip itself, I may use a freeze-frame from it, or the audio track. But how I handle my Audiosurf videos is a lot more cavalier, and I use an entirely different method for them.

Once I've removed the unused video clips from my hard drive, I go about with the editing process. This normally doesn't take too long, but it depends. For my "album in one go" clips, it's just a matter of splitting the clips at the start and end points of the song, and removing the unnecessary footage from the timeline. For the songs where I specifically chose that song, it's usually the same process, there's just more footage being removed from the timeline, as well as a bit more video splitting. For those videos I'll do the necessary editing at the beginning and the end, but I'll also split the clip right as the video fades from the summary pre-loading screen to the track. When you hit restart while playing a song, that doesn't get shown again, and I like that information to be shown at the start of the video. I then go on to the clip where I finally succeed at completing a track, and go frame-by-frame to the very first frame after the Esc menu is off-screen, and split the clip there. I remove any footage that happens to still be between those important splits, and I close the gap on the track. None of this extra work is necessary for today's footage, though, so I just have to do my basic splits and timeline removal.

At this point, I like to watch the whole clip in the preview pane to make sure that I don't have any small gaps between my video clips that will show up in the final project as a quick flash of black. It's important to note that when dealing with raw footage recorded via FRAPS, that most video-viewing software isn't properly equipped to deal with the raw data available in the unprocessed footage, so there may be the appearance of extreme framerate lag. Video editing software's preview pane is better equipped for viewing this type of footage, but it is still going to look like there are dropped frames. I'm just looking for gaps in my playback, at this point though, so I try not to let those "dropped frames" bother me. Assuming, everything is okay, I move on to the end stages of the editing process.

What I do next depends on a couple things. If I intend to upload footage from another game today, or I've more videos that I intend to edit more videos, I'll save the project, and close the program. When I want to produce the video, all I have to do is open up my editing software, open the project file, and I have my fully edited video waiting to be rendered. If the video I'm editing is not an Audiosurf one, or it's a video that I want to use as the picture, for a picture-in-picture of a different video, I'll go to the produce tab, and render it as an mpeg-2 NTSC 1080p video. Your video editor's codecs and desired picture quality may differ, but I render all my videos in 1080p. For those videos, after they're rendered, I'll often go look at the finished file, and see if there's any issues with the final product that I didn't catch during the editing process, and if so, I'll fix them, re-render and watch it again. I don't do any of these pre-upload renders for my Audiosurf videos any more, as the editing process for them is fairly straight-forward, but I do this for all my other videos still.

Assuming, everything is ready to go, whether the video is an Audiosurf one or not, I'll go to the Produce tab again, and select the Youtube tab within that tab.This tab takes a lot of the work out of uploading my videos - though sometimes it doesn't work properly and I still have to manually upload. I fill in my youtube account information, and I select 1080p from the quality drop-down menu. This information is stored in between sessions, so I only really have to do this once. I then fill out the title, description and tags for the video as well as selecting the category for the video. The title and tag fields are a bit more limited than uploading directly from Youtube, as you can't put as much text in, but this is still handy if what I'm uploading is more than 15 minutes long. If I'm uploading a video that is more than 15 minutes long, PowerDirector automatically splits the video at the 15 minute mark, makes a playlist on Youtube for the videos, and fills in the tags, description and title for each of the videos, appending a"(x/y)" (where x is the current part number, and y is the total part number) to the video title. At this point, I also decide whether I'm going to upload the video as public or private. If you upload a video as private, you can then make it public later, by going to the video details on Youtube and changing it in the privacy settings. It will still show up in your subscriber's subscription boxes once it goes public, but depending on how long ago it was uploaded, it may end up being buried under more recent stuff. I generally select public, but sometimes I'll upload it as private, and then make it public the next day. It's up to you.

I then hit the Start button (Not to be confused with Windows' Start button), and it begins the rendering and uploading process. I generally do this even when I know that the uploading portion of it isn't going to work, as the video still gets rendered into files that you can manually upload, and are conveniently split at the 15 minute mark, if applicable, and are in your desired Youtube video quality. Also, in the case of my videos, I still end up with a 1080p quality video in a smaller filesize than I can easily get by rendering using a different codec. Depending what time it is, while the program is doing the rendering/uploading process I'll either go to bed, read a book, or play a casual game.

And that, is how I go about, from start to finish, making an Audiosurf video. And remember, for all that work, these are videos that take the least amount of work for me to publish. Hope this was informative, and thanks for reading. There will be four new Audiosurf videos today, so be sure to subscribe to my Youtube channel to catch those, or follow me on Twitter so you can see my updates on their release.

No comments:

Post a Comment