How to make a good drifting video.. comments and opinions?

  • Thread starter Ske
  • 18 comments
  • 3,701 views

Ske

Gimp
Premium
939
Norway
Home
The_Ske
PzR Ske
I've seen a number of vague threads on this topic come and go, and since I'm now embarking on this bit myself I thought I'd try and open up this discussion here in the vid forum. This is a long post, but I hope some of you will take the time to read it and comment, voice opinions or contribute with knowledge.

I've seen quite a few very good videos around here, so no doubt a few of you can clear up a few of the points here. I just hope you're willing to share with us noobs how to make the most out of our saved replays ;)

The big question is naturally:
How do you make the best possible GT4 drift video?


Step 1: Get set up. Obviously, you need:

  • A video capture device (TV card, video-in, etc)
  • Capture/recording software
  • Editing software
  • Replays of your favorite cars going sideways
Step 2: Capture. From here, it gets more difficult. Get ready to deal with:
  • Which software and codecs to choose
  • Interlacing and other problems with capturing
  • Capture resolution and compression
  • Filters for color and contrast preservation
Step 3: Plan. What kind of a video are you going to make?
  • Drifting. Do you keep the subtle but clean drift, or should you go with the sloppy huge angle one, or both? What's more important, angle or line? Are including entries and exits necessities?
  • Concept. Do you need one?
  • Ambition and project size. How long should the video be?
  • Choice of music, if any. How long should one track play before you switch to another tune?
  • Parts. If you're planning a long video, you might wanna divide it into parts. What can be the "concepts" behind each part?
Step 4: Edit. This one's as tricky as it is personal, and it can definitely make or break your video.
  • Structure and chronology. Should you group clips with the same cars, on the same tracks, or similar types of drifting? Or simply try to spread everything out randomly?
  • Camera angles. Should the focus be to follow the car in linking drifts to show everything, or do you go for more artistic or static angles that switch mid drift? Managing the synced replay mode for snappy visuals during capture can be a major pain, so is normal replay good?
  • Overlay text and transitions. Less is more?
Step 5: Distribute. How to get the most value for your MB?
  • Which file formats and codecs are acceptable or desireable?
  • What's an acceptable target file size, total or per minute?
  • How do you optimize quality while minimizing file sizes?
These are some of the points I'm more or less unsure of at the moment, and I'm guessing more people than me could use a helping hand.

:)
 
Boundary Layer's Recording/Rendering Setup


Just going to post my thoughts on the first few steps in detail. My approach to editting changes from vid to vid and really depends heavily upon the music I choose.

1st Step:
  • I have a ATI TV Wonder Pro for my capture card. I capture with composite cables coming through the Record Outputs of my stereo reciever to which my PS2 is also connected.
  • I use VirtualDub (or VirtualDubMod) as my capture program - ATI VCR is a useless pile of poo.
  • I do my finals cuts and renders with Vegas Video 6.

2nd Step:
  • When capturing in Virtual Dub, I use the huffyuv compression codec. It is a lossless codec, does an extremely good job. I only ever lose 2 frames MAX at the beginning of every capture. ...just make sure there is a lot of HD space available for your captures before doing anything (1min in 640x480 ~ 1.5GB).
  • At this point I do not do any deinterlacing or filtering. It is an additional load on the machine, and I want everything as freed up as possible so I don't drop any frames during the capture.
  • I use capture resolutions of 640x480, or 720x480 depending on the format of the media I'm recording.
  • After doing the initial captures, VD lets you re-render your clips and save as an avi. I use huffyuv for the compression during this step as well. This is also where I do my deinterlacing and filtering. For my capture card I usually find I must increase brightness by 7% and contrast by 12% when capturing from my PS2. I deinterlace by blending fields together.
  • I'm kind of skipping steps here, but whatever. (Ske, I'd suggest putting in a step between Editting and distribution for final render settings)
    For final renders I take the AVI's generated with VD and plop them into Vegas 6.0. The filtering should have already been done by now, but I still have to add text effects and do all that editting stuff. Eventually I get something that I'm ready to show off to the world.
    I render my project as a wmv with the following settings:
    • Video Rendering Quality: Best
    • Audio: CBR - WMA 9.1 - 32kbps, 44khz, stereo (A/V) CBR
    • Video: bit rate VBR - WMV V7 - 23.976 (IVTC) FPS - 700 to 1000 kbps

3rd and 4th step:
  • depends how many pills I'm on and what color the sky is today.
  • NO EUROBEAT!

5th Step:
  • As mentioned, I like wmv's. With my setup I have found it possible to attain excellent video quality and acceptable file sizes with this format.
  • I would consider 10mb/minute to be as large as I will go with a file. If it is something that is going to be placed on a metered server, I aim for 5mb/minute. If I'm not concerned about bandwidth, then I may goto 8mb/min or higher.
  • To save size I drop my frame rate from 30FPS to the IVTC standard of 23.976. It's hardly noticeable. Also, I render in VBR not CBR.
    You may have noticed I use WMV V7 instead of a newer version. That is also to save on file size. For whatever reasons, I have found that version to offer better video quality at lower file sizes than newer versions are capable of.
  • this wasnt part of the survey, but I want to spread the word. File Front offers a free service with unlimted server space and unlimted bandwidth for uploading your vids. If you do not have webspace of your own, or if you have very limited bandwidth with your service I would strongly recommend you place your files here. The links will not expire as they do at yousendit.



that's the gist of it for me.

Don't view any of the vids I currently have posted as examples of this setup. I fumbled around with settings all summer until I got this right. IDM members have seen what it can do... as well as anyone who managed to snag my HK2ex demo vid for the current GTPDC.
...I'll reupload that right now to show it off again. BRB


edit: Ok, I'm back.
Example of my capture method:
HK2ex Vid - created 06/09/05
 
I'll just add my own personal preferences as of now (the noob stage ;) ) here to separate them from the OP.

Step 1-2: I use a cheap Phillips card with VirtualDub. I'm currently unsure of which compression codec to use, I'm not 100% happy with HuffYUV.. or I might just not have it set up properly. I edit in Vegas 6 (trial) and it's an incredibly easy and versatile program. My captures have horrible motion interlacing that I haven't been able to get rid of without seriously compromising sharpness :(

Step 3: Personally, i'm for clean drifts and keeping a good racing line rather than huge stuff done in a sloppy way.. IMO it should look like you're in control, unless you're aiming for educational comments to your drifting skills ;) IMO a video longer than 5-6 or so minutes should have either a frantic pace, or some sections/parts for easier digestion. Obviously, splitting it according to drivetrain like DR has done is a good idea, but I guess most of us are still learning FR/MR. A concept is a good idea if you're very productive.. to give each of your videos "something else".
And I agree.. Eurobeat/trance is right out ;)

Step 4: I still need to find out some guidelines or pitfalls here.. but from the comments my first video got (and I agree), try to alternate between chase and static cams.. and synced vs normal replay mode. I personally like chasing long links, as it captures the flow a lot better. Big single corner attacks look best from a static cam IMO.
As for text and overlays, I like to keep it simple. Make sure what's there is readable at your final resolution, and don't overdo font and color use. Scrolling text can be unreadable at 320x240.

Step 5: I'm still unsure what is best when it comes to file size vs quality, rm or wmv. Wmv seems to have smoother colors and RM has sharper edges. IMO, WMV is generally more desireable since you don't need a separate player to view them.
It seems like 8MB/minute (1 mbps) is what you need to aim for, though 500kbps RM files are very good for their size (a lot better than WMV, I think) and adequate for previews/demonstrations. The downside with RM is that you don't really have a lot of adjustable options for encoding, unlike WMV.
 
Im new to makin vids, and Im useing my x800xt AIW to capture. I used Pinnacle Studio 9 (came with the card) to capture it, and i made it into a .avi, but its 280mb.......and thats at 50% quality, and no sound. the vid is just under 7min

Now, I have no idea how to compress it and stuff to make it alot smaller. Any suggestions?
 
download the huffyuv compression codec (I assume you can use that with Pinnacle Studio) and use that during captures. This will give you a lossless capture as your starting point.

Then edit and re-render your captured clips with whatever video editting software you prefer into a .rm or .wmv, or whatever. You should be able to get 7mins of decent looking video with sound at 640x480 down to 60-70mb or so.
 
I installed the HuffYUV codec, but i cant capture useing that with Pinnacle studio. But when I go to Make Movie, i can use HuffYUV, but it brings it to ~400mb. Is there any free software I can use? Dont have the cash atm to buy any :(
 
ya, you definitely don't want to use huffyuv for your final render, thats not it's intended purpose. It's meant only to offer lossless captures.

virtual dub is a great alternative capturing program. It's free.
You can use the free Windows Movie Maker to do your final renders in wmv format.
 
Ohhhhhh, I thought I was supposed to use HuffYUV for all of it. Well I didnt have any lost frames when i captured useing Pinnacle Studio. I was able to import it into windows movie maker just fine. I made a short lil 40sec vid, only 4mb :). Thanks for the help boundry, i just gotta do some editing and stuff now :)
 
Does anybody have any experience with DivX / 3ivX / XviD encoding? I've seen some excellent 640x480 tv episodes and movies which have incredible quality at 5MB/minute or less (VUK's 350mb/hour TopGear eps for instance, if anyone's seen those).

I'm having a hard time finding good settings for those codecs when exporting directly from Vegas, but it seems they have better potential than WMV.
 
check out some guides at doom9

i've tried fiddling around with xvid and x.264 codecs but haven't had great results yet.
as soon as my midterms are done I want to learn how to use gordianknot - that piece of software seems to be vital in getting decent quality and great compression out of an avi
 
ok, I have a new Step 2:

  • I am still capturing with virtual dub at 640 x 480 with the lossless huffyuv codec. big file sizes, but huffyuv writes extremely fast, so no lost frames.

  • Rather than applying filters at this stage, I have configured my video proc amp to increase brightness and contrast slightly

  • I suppose I could apply deinterlacing during the capture, but I have avoided doing so to conserve CPU (I don't want to bog down my machine with extra tasks and risk dropping frames)

  • I take my raw captures from VDub and edit them in Vegas 6. I chop them, add text, put it all to music, and do whatever else I wanna do here. When I'm satisfied, I render to a lossless avi with Lagarith lossless codec.
    I use lagarith here because it's a little more configurable than huffyuv and has much smaller file sizes. The write time is slightly longer, but decompression at a later stage is significantly faster, I've found.

so, now I have a lossless editted file.
Time to plug it into 1 of 2 things.
  • option 1: I can go back to vdub, open it, apply whatever filters I deem necessary. Deinterlacing is usually all I need. Sometimes I'll tune up the saturation slightly.
    Then I render to xvid with an avi handler. I use the L5 profile on a twopass-first pass encoding type. If i'm worried about file size I'll drop my frame rate from 29.97 to 23.976fps. I did a little work with quantizers to help reduce file size and keep quality - but I've reinstalled all my codecs recently trying to get someone elses vid to play, and all my xvid settings got reset.

  • option 2: I open up Auto Gordian Knot, open up my lagartih lossless avi, set my bitrate, etc, and start rendering to xvid format avi.
    This looks after all the confusing xvid settings for you, but you are unable to apply filters here. Get all your filtering done ahead of time if you go this route.

  • I suppose I could also render from vegas to xvid without doing the lagarith thing. I still prefer to go with either autogk or vdub for my final render. autogk is dirt simple and does everything for you (who cares what a quantizer is, right?). Or Vdub b/c i think their filters are better than those in vegas.

the whole process is kinda time consuming, but it doesnt bother me. The resulting video quality is justification enough for me to go through the lengthy process.

The wmv settings I posted awhile back still aren't bad. But at a similar bitrate wmv's look just a little bit blurry compared to xvid. On the other hand, wmv's are much quicker to render.

here's a rather large xvid file that I got with my first option. It's at the standard NTSC frame rate of 29.97 fps rather than being reduced to ivtc 23.976. The bitrate (including audio) it approx 950 kbps - I think the audio stream is 128kbps MP3 format, so the video is somewheres around 825kbps. Kinda high... I've gone as low as 650 and it's still bearable. But serious blocking begins to occur in places, and clarity suffers as a result. Good file sizes though.

anyways, I think this looks pretty good. nice and crisp.
DSJ3 Highlights http://www.bgj.ca/GT4/BLvids/BL-DSJ3.xvid.avi
(Evo VIII MR GSR, you may have seen it already)
 
http://www.videohelp.com/ and http://dvinfo.net/ are good friends.

A couple of quick tips for editing to entertain:
Variety is a crucial part in keeping the audience's attention
Unless you really know what you're doing, keep editing simplistic
DON'T USE CHEESY TRANSITIONS-The cut and the crossfade are a beautiful thing
Syncing to music can either make or break a video
We want to see HAMMERS, delete the weak stuff
Every video should have a name

On a side note, if you want to experiment with editing, by all means do it. An experiment is just that, it will either pass or fail. Either way you'll learn something. But if you release something that is "outside of the box" don't always expect it to go over with the audience.
 
I feel pretty stupid here, but I couldn't find a place to put this question.

Is it possible, to capture video from my PS2 to my PC using my Redant PS2 VGA Box?

I'm pretty sure it's no. But I'm hoping not...
 
i'm pretty sure it's no as well.
I don't know a lot about those units, but I thought they were only for converting your video signal into something that can be displayed properly on a computer monitor - no recording capability.
 
Step 1:

Grab a digital camera

Step 2:

Play your lap replay and record it with the camera

Step 3:

Visit fileupyours.com and upload



This is how I do it, simple or what?
 
I got a question about this, I use DVD Express, a crappy little usb box You hook up AV Jacks too, has decent quality, but it only encodes as MPEG2, and the files are so big I can't get smooth playback in Primeire, I was wondering If there was a way I could fix this? Virtual Dub doesn't seem to want to open these files either, so I can't compress them further.. I'm not very good with codecs. lol
 
Back