Macro footage shot with my HV20 and a Canon 50mm 1:1.8 macro lens, turned into a slow-motion abstract project. View in HD and/or download it here.

Update: Video updated and replaced with a new version.

This is a generic tutorial on audio/video compression. It does not adhere to any specific application, but it teaches you what is what, so by using this knowledge you can export from ANY application in the market. So, here’s what you need to know:

* Many video editors require that you “tell them” in their project properties what kind of source footage you got. For example, if your source footage is NTSC HDV, you need to tell them to accommodate/optimize for 1080/60i. For PAL that would be 1080/50i. Then there is 24p footage, different resolutions, or plain DV or just digicam VGA video. You must always know what kind of source footage your camera outputs, so you can configure your video editor or utility to accommodate it. For example, if your camera shoots HD at 24fps and you let your video editor to use the default DV in a 60i timeline, you will get bad quality, and performance degradation during editing. So, get it right! Not all editors automatically recognize your source footage and auto-configure themselves.

* If you are using Sony Vegas, I would recommend you right click on clips in the timeline that lots of motion, select “Properties” and then “Disable Resample”. This will get rid of the “ghosting” effect during the final export. Not sure how to do that on other video editors, most of them don’t have the ability to turn it on/off.

* When you are finally ready to export, the first thing you need to decide is the medium you want to export to. Is it the web, the iPod, the PS3, a DVD, or simple archival? You see, depending the device you want to export for, different codecs and options apply. Here are some basic codec guidelines for some popular devices:
- iPod: 320×240 resolution, native frame rate as source footage, h.264 video codec 512 kbps, AAC audio codec 64 kbps stereo.
- Sony PS3 and XboX360: 1280×720 or 1920×1080 resolution depending if you have a 1080p or a 720p TV, native frame rate as source footage, h.264 or WMV or XViD video codec at 5 or 9 mbps, AAC audio codec 128 kbps stereo or 5.1 surround.
- DVD: use the DVD templates that video editors usually come with.
- Youtube: Same as iPod.
- iPhone and PSP: Same as iPod, but at 480×270 resolution at 1mbps bitrate.
- Vimeo SD and HD: Tutorials here and here.
- Zune and Creative players: 320×240 resolution, native frame rate as source footage, WMV video codec 512 kbps, WMA audio codec 64 kbps stereo.
- Archos and other PMPs: 320×240 resolution, native frame rate as source footage, XViD video codec 512 kbps, MP3 audio codec 64 kbps stereo.
- Archival: a lossless or near-lossless codec, like DV, HDV, mjpeg, Huffyuv, Lagarith, Cineform, ProRes, AIC etc.

* Video codecs. There are two kinds of codecs. The Delivery grade codecs and the Intermediate grade codecs. The first ones are supposed to be exporting/view-friendly, and the other ones archival/editing friendly. More information here on the subject. I would suggest to hunt for h.264 video, AAC audio, inside the .mp4 container format. That’s the most compatible and widespread format today on devices.

* Frame rate. Keep frame rate the same as the source footage (you should be able to find some “summary” information about your footage somewhere on your application). For NTSC that would be 29.97 (60i), for PAL 25 (50i), and then there’s 24 progressive frames, which actually in reality is 23.976 fps. I would suggest you export with the right frame rate each time and not round these weird numbers to 30.00 or 24.00.

* Resolution. When you export, depending on the camera you got, here are your resolution options and bitrate you should be exporting at with delivery-grade codecs:
NTSC DV 4:3: 656×480 or 640×480 (at 1800 kbps)
PAL DV 4:3: 768×576 (at 2300 kbps)
NTSC 16:9: 874×480 or 880×480 (at 3000 kbps)
PAL 16:9: 1040×576 (at 4000 kbps) or 880×480 (at 3000 kbps)
HVX or DVX 16:9 DV Panasonic cameras, export at 848×480 or 852×480 (at 3000 kbps)
AVCHD/HDV: 1920×1080 (9 mbps) or 1280×720 (5 mbps)
Canon TX1 or Kodak 720p digicams: 1280×720 (5 mbps)
Digital camera VGA 4:3: 640×480 (at 1800 kbps)
Please note that you should never export for web/viewing purposes with aspect ratios that are not 1.0000. I see a lot of people for example exporting DV at 720×480 or 720×576, but this is not the safe thing to do, because most applications don’t take into account the special aspect ratio value that these kinds of exports have, and so you end up with squashed heads. Use the above guide to get it right, and make sure that aspect ratio is set to either 4:3 or 16:9 for the above resolutions, or at aspect ratio 1.0000 if your application uses this way to represent pixel information.

* Bitrate. Bitrate is the amount of bits per seconds that the video uses. That’s the only factor that decides how big a video file will be or not. There are two kinds of bitrates: constant and VBR. In constant bitrate, you just tell the application to use a specific number of kbps or mbps, as shown above. But in VBR, you give two numbers: one for the average number and one for the peak number. For example, if you want an average bitrate of 3mbps, you can also ask the application to have a peak number of 5mpbs. The application will go as high as 5mbps only when there is a lot of motion on the scene and it requires more bits to encode it properly, while it will stay at 3mbps or below if the image is static and doesn’t require lots of bits. VBR is generally preferred for best quality, but it’s more difficult to figure out, if the application’s user interface is not well done.

* De-interlacing. If your camera does not shoot in progressive mode (most camcorders don’t, only digital cameras do), then you must de-interlace during export, otherwise your video will have “jaggies” (horizontal lines). You must hunt for a checkbox or option that’s called “de-interlacing” or “progressive”. Please note that if you are exporting in an intermediate format for editing/archival reasons, or back to the camera’s tape, you should not be exporting as progressive, because this is not the footage’s native format. Only de-interlace for “viewing” purposes, e.g. youtube, ipod etc.

* Audio bitrate. For audio select between 64kbps to 128 kbps (although you can have VBR for audio too). More than that is usually goes unused by many devices/headphones, so there’s no reason to use more. Use 44.100 Khz for sample rate, and Stereo.

* Audio Codec. Like with the video codecs, there are many different audio codecs you can choose from. But the audio codec should always be decided in conjuction to the video codec. For example, if you want to export in h.264 Mpeg4, you must use AAC, because that’s the way most players are optimized to read h.264 with. For DivX/XViD it’s mp3. For Theora video it’s Vorbis audio. For WMV it’s WMA, and so on.

* Containers. Don’t confuse “containers” with “codecs”. MOV and AVI are *not* codecs, so saying that “I exported as AVI” says nothing to others who might try to debug your problem. AVI/MOV are simply file formats that inside them can “host” actual video and audio codecs. In theory, you can have a gazillion different codecs in them, it’s just a container format to keep “glued” the different kinds of audio and video together…

There are three situations that most amateur video users just can’t put their head around. On online forums these are the most common exporting issues.

1. Pixel aspect ratios
I bet that trigonometry might feel simpler to some people. No matter how many tutorials are on the web about pixel aspect ratios, users just don’t understand them. I really hope that camera manufacturers stop using non-square pixels because it confuses the hell out of people, and most of the time they f*ck up their web exports resulting in youtube videos where people in them have heads looking like eggs, or in videos with vertical letterboxing.

2. Interlacing
Ah, this one is funny. Users who have never seen interlacing before are in panic. They come over to the forums and scream that their camera is dying! There are “weird horizontal lines everywhere” and they wonder if warranty is still good to send the camera back! Again, it’s a matter of the camera manufacturers to stop using freaking interlacing. We are not in the 1950s anymore.

3. AVI and MOV
There’s the misconception that all AVI files are DivX/XViD, and all MOV files are “the same”. People don’t understand that avi/mov are simply containers that can hold any kind of media format inside them, and instead they associate them with specific codecs. I got an email the other day: “why my 2 minute footage is 6 GB, I exported as AVI, it should have been smaller than that”. Ah, well, because you saved using the uncompressed codec. Duh.

Amazing work by Jeremy Saville! Very funny, very well executed — Jeremy is a professional in the entertainment business. Not to be missed!

Two more short movies of Jeremy’s here.

A second person in the past few weeks told me that hobbyist artistic videography is “meaningless”, and that has pissed the hell out of me. I will say this only once:

Videography done like on some of the videos below is art. Art is anything that can create or boost an emotion. And many of these videos do that. Not everyone’s videos are art, but there are some very capable amateur enthusiasts that know how to shoot, know how to edit and know what they want to present on screen (e.g. Charlie McCarthy, Remyyy). And this is NOT something that everyone can do. Absolutely not. It’s not as easy as it seems. It’s not as simple as taking a camera and pointing into random things. And editing at the end is an art in itself. I CHALLENGE you, the random reader of this blog, to take your camcorder and try to do something “beautiful & artistic” with it (expensive equipment is not required). Chances are, you will fail. It’s not as easy as it seems.

Beautiful imagery is not meaningless either. Cameras are not created just so we have a plot each and every time with a John Wayne style script. That’s just only one usage of cameras. Video cameras are nothing but “moving pictures”, and as beautiful still pictures are considered art, same for the moving ones.

If I haven’t convinced you yet, we will have to find an equivalent art to compare. And that’s the DEMOSCENE back in the ’90s. In fact, these two scenes are producing similarly-looking products. From wikipedia: “What began as a type of electronic graffiti on cracked software became an art form unto itself” and “the most experimental, unusual and controversial demos are often referred to as art demos or abstract demos.” As the demoscene developer has to be both an artist and a technical person at the same time, same for a videographer, he/she will have to master both. And some of these people out there, have done so. But they are VERY FEW who have done so successfully (no, I am not among them). Which is why videography is interesting, and why it’s true modern art: it’s challenging.

* I watched “El Mariachi” tonight. This is a 1992 Mexican movie that was shot for just $7000 and made it to the big screen with a 7.0 rating at IMDb. Another cheap movie is “Primer“, also shot for $7000, but in USA in 2004. I keep thinking that if the “Primer” creator had used a digital camcorder instead of 16mm (e.g. the Canon Optura Xi which was a pretty good model for the money at the time) he would have been able to make the movie for cheaper. If you are a video enthusiast you owe it to yourself to watch these two movies. They are the epitome of true indie film making and already legendary for that status. And of course, buy the “DV’s Rebel Guide” book. It’s the ultimate “that’s how you do it for dirt cheap” indie film book.

* Last night I watched the lighting tutorial DVD that came with the reflectors I bought two months ago. It was very interesting and simple, it really demystified lighting for me. I feel that I am ready to actually correctly use lights if I want to shoot a music video clip or short movie. Highly recommended to get these reflectors, if not for the included DVD too!

* I spent the day re-encoding some of my favorite Vimeo videos in a format that the PS3 can playback. I have a collection of about 85 videos so far. The PS3 is powerful and doesn’t sweat in HD playback, but its UI is not as good as AppleTV’s regarding media. Hopefully, Jobs will announce the AppleTV 2.0 on Tuesday, but I need it to not only be able to deliver 1080/30p, but also to playback WMV and DivX/XViD in order to be useful to me.

* I published a review of the Nokia N82. Expect next week a review of a 28″ PC 1920×1200 monitor too. I used it for editing my HD footage in 1:1 size, and I must say that for the first time I was able to see clearly how noisy the HV20 is — and the HDV-encoded artifacts too. In all truth told, if you want a “super clean” final cut, the HV20 is able to deliver a good 720p video, but at full 1080p it’s noisy and artifact-prone if you stick your nose to the monitor and you get careful on what you are looking at.

* Havoc Pennington left Red Hat. I don’t know where he’s going next, I wish him good luck, but I also feel sad because the last true leader on Gnome has left the house.

I created a new HD channel over at Vimeo, holding some of my favorite clips in it so far:

In the meantime, I discovered Remyyy, possibly the best video artist enthusiast in the whole Vimeo world:

Update: I just finished IM’ing with him, I think his future shorts will be filmed in widescreen and will have the “HD” treatment on Vimeo. Can’t wait!

Many HD cameras shoot 24p these days, and while some consumer-grade video editors can deal with 24p, there aren’t many cheap DVD authoring applications that support 24p. For example, ‘Sony Vegas Platinum 8′ unofficially supports 24p timelines, but its accompanied ‘DVD Architect Studio’ application doesn’t. If you are shooting in 24p, it’s better to edit and burn a real 24p DVD and let the TV add pulldown during playback, rather than adding pulldown during the DVD authoring process. Besides, there are some new playback devices in the market these days that can output true 24p to 24p Sony TVs, without adding pulldown, so these cases while rare, can benefit from pure 24p DVDs. Here’s how to go around DVD authoring limitations and produce a 24p NTSC DVD:

1. Download and install the latest “nightly by clsid” ffdshow build.

2. Capture your footage the way you usually do. If your camera does not capture as true progressive in the 23.976 frame rate, but instead it uses something like Canon’s PF24 or some of Sony variants, you must remove pulldown (methods for the HV20 here, or for Canon’s AVCHD line here).

3. Then, bring your footage into your video editor, just make sure your editor does support true 24p editing (e.g. VMS Platinum). In the “File”/”Project Properties” select 1440×1080 size, frame rate of 23.976 (type it if there’s no such option), progressive field order, aspect ratio of 1.3333, rendering quality “best” and “none” for de-interlacing method. Then, edit as usual.

4. When your editing is done, you export in an intermediate format. Click “File”, “Render As”, select the “avi” type and its “HDV 720-25p intermediate” template. Then, click “custom”. In the “video” tab of the dialog that pops up select “23.976 (IVTC Film)” for frame rate, and then from the video format menu select the “ffdshow video codec”. Click “configure”. From the newly created dialog select the “encoder” tab, and from the encoder menu select “Lossless JPEG” and “YV12″ for its Colorspace. Click “Ok” to close that dialog. Click “Ok” to close the other dialog too. Then, hit “Save” to start encoding the .avi file.

5. Install the latest “DVD Flick” version (as I write this, the latest beta version can be found here) and then load the application (regularly check for new versions of this app). Click “Project Settings” and go to its “Video” tab. There, select “NTSC-film” as target format. From the “Burning” tab you can instruct the application to burn a disc at the end of your authoring, or just create the DVD/.iso files without burning. Then, load your LJpeg .avi file(s) you exported from your video editor into DVD Flick. Read the DVD Flick manual to learn how to author DVDs with this application. It is a simple application to use, but it doesn’t have enough templates and beautifications. Save often too, as the application is not super-stable either, but it’s the best we got to do this job for free. After you are done authoring, you can click the “Create DVD” icon to burn or just create an .iso file. Enjoy!

Some notes on this method:

* I don’t use 1440×1080 to export via the video editor on step #4 because the kinds of .avi files that VMS produces are not recognized as widescreen by other applications. This results in DVD Flick having vertical letterbox bars, and that’s not what we want. Besides, downscaling first to 720p and then to 480p is not very lossy to make you worry about it.

* If your 24p camera is a DV one instead of HD, export in 874×480 (progressive, 23.976 frame rate, aspect ratio 1.000), instead of the suggested 720p resolution on step #4 (use the same codec as suggested though). If your camera is the DVX-100 export in 848×480.

* I am using LJpeg instead of the Huffyuv codec in this tutorial because the mode of Huffyuv that produces smaller files uses a colorspace that DVD Flick does not support. As for the Lagarith lossless codec, or Cineform, they are not supported by DVD Flick either, so your best bet is LJpeg (the FFv1 ffdshow codec could be a workable idea, but it’s slower to encode that LJpeg).

Samsung and Panasonic have announced their new camcorder line up, and Canon’s already have been leaked too. The new trend is flash storage and full progressive 1080/30p recording. Panasonic goes the extra mile to include true progressive 1080/24p too, while Samsung can capture 300fps for 10 seconds which yields very “clean” slow motion. We know no tech specs about the new Canon cameras yet, but we know that they went flash-storage too. Sony didn’t offer major new ideas/features in their new lineup.

I am very happy to at last see true progressive recording, no more wacky aspect ratios, no more interlaced crap for consumers (50% of them don’t know how to de-interlace and then they export interlaced to sites like vimeo and so their videos end up looking like crap), and of course 24p.

However, Panasonic did not go the extra mile to compete with the HV20, and Samsung didn’t seem to even try either. Just because you can shoot progressively and 24p does not make your camcorder a better device than the HV20. Not only because AVCHD is still not as good as HDV in quality, or because 3CCDs are so small in these consumer cameras that yield essentially no background blur compared to 1/2.7 CMOS of the HV20, but also because there is no manual focus ring, no filter threads, no manual controls in hardware buttons.

This reminds me a lot of the efforts of Samsung and LG to compete with the iPhone last year. Instead of actually replicating all the existing iPhone features plus adding more new features, they just kept their old software feature-set and just gave their device a big ass touchscreen. Well, that’s not enough to compete with the iPhone, and in the same way, Panasonic’s solution is not a better deal for the hobbyist/indie filmmaker than the HV20 is.

Personally, I will still wait for this. If Canon won’t deliver that, no one will.

Update: This Casio digital camera looked very promising for a moment, but then I read that it has no image stabilization, not conversion lens support, and not enough background blur (same as in the HV20). This is why followers are not leaders in the market. Because they make a lot of buzz about a single feature that no one else has, but they forget to take care of the basic stuff. Stupid Casio.

Canon’s consumer HD cameras have 24p support, only that it’s not a… truly true 24p. It’s PF24, which incorporates both progressive and interlaced frames in a 60i stream. To get the pure 24p stream, you need to do a “pulldown removal” (aka “inverse telecine”). PAL users don’t need to do anything special, this burden is only for 24p NTSC users. If your final output is the TV (e.g. via DVD or the camcorder itself), and as long as you never de-interlace that footage, then you don’t need to remove pulldown. But if your output is not a home DVD/TV but professional work, the web, or a PC or other viewing device, then you better get to work and remove pulldown because your footage will have an ugly “ghost” effect whenever there is motion on the screen. Here are pulldown-removed images before & after, showcasing the problem and its fix. Here’s another example showing the problem.

There are several ways to remove PF24 pulldown for the Canon HDV cameras, but their AVCHD cameras (e.g. HG10, HR10) did not enjoy the same fate, until now. If you are using the latest versions of After Effects or Final Cut Pro, then you can use the tutorials linked to remove pulldown (although I have not confirmed that these tutorials will work for AVCHD formats as well). On the PC side, the cheapest, fastest and cleanest way to do it, would be to purchase the Cineform NeoHDV or NeoHD tools (discount here). However, there is another way, a freeware method, as long as you already have a video editor that supports the AVCHD format, DirectShow/VfW and 24p (e.g. Sony Vegas/VMS, AE/Premiere, Avid etc). Please note that Premiere Elements, Pinnacle, Magix Movie Edit Pro and Ulead don’t support true 24p so avoid these for 24p editing. This method is an involved procedure to follow and it requires some very basic usage of the DOS prompt (found in Windows’ Accessories program menu as “MS-DOS”), but if you have no money and you need it done, it should work fine (tested with HG10’s PF24 files).

1. Download the plain version of Mplayer for Windows and unzip it somewhere that you can find back easily. Inside that unzipped mplayer folder, create a folder named “input”, and a folder named “output”. This action only needs to be done once for your system.

2. Download the Huffyuv lossless codec. Unzip the .zip file on your desktop, right click on the huffuyv.inf file and select “Install” (note: Vista might not give you that option, in which case manual registration of the codec’s DLL file must happen via the DOS prompt, google it). After about 15 seconds, you can delete these files from your desktop, as the Huffyuv lossless codec is now installed on your system. Open a DOS prompt and navigate to the mplayer folder (e.g. “cd C:\Mplayer\”, without the quotes). There, run this command: vfw2menc -f huff -d huffyuv.dll -s settings.mcf and on the new dialog that pops up select: “Predict median (best)” from the first drop down menu, and “<-- Convert to YUY2" from the second drop down menu, while leaving unchecked the other options. Then click "ok" to discard the Huffyuv configuration dialog. This action only needs to be done once for your system.

3. Download the Lagarith lossless codec. Unzip the .zip file on your desktop, right click on the lagarith.inf file and select “Install” (note: Vista might not give you that option, in which case manual registration of the codec’s DLL file must happen via the DOS prompt, google it). After about 15 seconds, you can delete these files from your desktop, as the Lagarith lossless codec is now installed on your system. Open a DOS prompt and navigate to the mplayer folder (if it’s not open already at that folder). There, run this command: vfw2menc -f laga -d lagarith.dll -s settings.mcf and on the new dialog that pops up select: “YUY2″ from the “Mode” menu, and “Use Multithreading” if your CPU is hyperthreaded/multi-CPU or it has more than one Core (leave unchecked if you don’t know). Then click “ok” to discard the Lagarith configuration dialog. This action only needs to be done once for your system. Note: Do not change to YUY2 if you are using Adobe’s editors, as they don’t understand that format.

4. I will work with Vegas on this step because that’s what I use, but it should be similar for Premiere/AE/Avid. Open Vegas and load your PF24 .m2ts or .mts AVCHD files in the “Project media” bin. Make sure that these files were indeed recorded as PF24 (don’t mix 60i and PF24-recorded files), and that your hard drives are formatted in NTFS. Then open the “Project Properties” dialog and use the HDV 1080/60i template from the drop down menu. Now you have two options, either pull all your scenes at once in the timeline to export, or export them one by one. The first method does not require your attendance, but it will produce one large file that will have no scene detection, while the second method will produce lots of per-scene files, but it will require that you be in front of your computer quite often to export the scenes one by one. Think hard of what your needs are, and make your choice. I would personally recommend you export the scenes one by one mostly because this way you can avoid the possibility of the audio and video getting out of sync (plus, MEncoder is known to not handle huge files too well either).

5. Once the clip(s) are on the timeline, hit “File” and “Render As”. Select .avi for the “Save as type”, and the “HDV 1080-60i intermediate” template. Then, hit “Custom”. On the first tab select “Best” quality. On the video tab leave everything as is except the video format, select there “Huffyuv v2.1.1″. Then, hit “Configure” and a new dialog pops up. There, make sure that “Predict median (best)” from the first drop down menu, and “<-- Convert to YUY2" from the second drop down menu are selected (leave unchecked the rest of the options there). Then click "ok" to take that new dialog away. Leave unchanged the "Audio" tab, and then hit "ok" in the "custom template" dialog. Then, you must give a name to your .avi file. If you chose the "one big file with no scene detection" method give it any name you want (as long as there are no spaces), but if you chose to export the scenes separately then use the same filenames as in their original AVCHD files, but with the .avi extension this time, and with no spaces in the filename. Export the files in the "input" folder you created inside the Mplayer folder.

6. After all the exporting is done, open a DOS prompt, navigate to the Mplayer folder that has the mencoder.exe file in it, and then run the following command, substituting each time the right input/output filenames with your own video filenames:
mencoder input/INPUT_FILENAME_001.avi -aspect 16:9 -fps 30000/1001 -oac pcm -vf detc,scale=1440:1080 -ofps 24000/1001 -ovc vfw -xvfwopts codec=lagarith.dll -o output/OUTPUT_FILENAME_001.avi
If you are proficient with DOS scripting, you can even write a script that will automatically read all the files on the “input” folder and export them one by one without requiring your attendance. Phoric wrote such a script and sent it to us (save it down and put it in the mplayer folder before you execute it). Thanks Phoric!

7. After MEncoder is done re-encoding all your files, you can choose to delete or move away the .m2t files on the “input” folder (so they don’t mix with video files you will place there in the future), while all the .avi files found on the “output” folder are now ready to edit! Windows Media Player won’t recognize these files as widescreen, but don’t worry, they are widescreen: Vegas will recognize them as such (if not, just set manually the clips’ aspect ratio to 1.3333). Create a new project in Vegas (or any other 24p-enabled video editor), select again the HDV 1080/60i template, but change the following this time: Field Order should be “progressive scan”, frame rate should be “23.976″ (type it if it’s not in the list), rendering quality “Best”, and de-interlacing method “none”. Then, edit as usual and enjoy true 24p!

NOTE: Trevor Marshall suggests an alternative method (not tested by me, might have an audio sync issue).

1. Download, unzip and drop the MEncoder files for Mac OS X on a folder of your choice, e.g. on one that’s called “mplayer”. Create two folders in that folder, one called “input” and one called “output”. This action only needs to be done once for your system.

2. Use a Quicktime-enabled application (any will do, as long as it has access to Quicktime’s exporting dialogs) to export your AVCHD files as MJPEG/PhotoJPEG inside the “input” folder. MJPEG is not exactly a lossless codec, but it’s the only high quality near-lossless codec that comes out of Quicktime that MEncoder will support.

3. By using OSX’s Terminal.app, navigate to the mplayer folder (it requires basic understanding of Terminal usage), and then use the following MEncoder command line to remove pulldown and export in Huffyuv, like this:
./mencoder input/INPUT_FILENAME.mov -aspect 16:9 -fps 30000/1001 -oac pcm -vf detc,scale=1440:1080 -ofps 24000/1001 -ovc lavc -lavcopts vcodec=huffyuv:pred=0:format=422P:vstrict=-1:aspect=16/9 -o output/OUTPUT_FILENAME.avi
You need to manually replace the filenames for INPUT_FILENAME and OUTPUT_FILENAME for each of your files.

4. After all the encodings are done for all your files, you can safely delete the files inside the “input” folder. Your final, true 24p files to be used with editing can be found on the “output” folder.

5. Install the “Perian” utility which will enable Quicktime-enabled video editors to read and edit these created Huffyuv .avi files. This action only needs to be done once for your system.

If you are proficient with Bash scripting or AppleScript, you can write a script that will automatically read all the files on the “input” folder and export them one by one using MEncoder without requiring your attendance. If you write such a script, please send it to me to incorporate it in this tutorial and help others. Please note that iMovie and Final Cut Express don’t support true 24p editing, so if these are the only video editors you got, then either don’t bother removing pulldown, or buy a PC to do the job (it will be cheaper to buy a fast PC with 2 GB of RAM and Vegas Platinum and use that as a dedicated video station, than buying Final Cut Studio or After Effects for the Mac).

There is no way to apply this method under Linux because MEncoder crashes when you try to read an AVCHD file, while ffmpeg has a known bug in the reading of such files that results the video to export twice as long as it should. As of this writing, no Linux library or application can read AVCHD files properly.

1. Admittedly, this is a slow way of doing things. For 1 hour of footage, it will take between 8 and 10 hours to complete the various exports, depending on the speed of your computer. If time is money for you, buy Cineform. If not, let it run overnight. A trick you can do to speed up the various exports is to have the input and output files on different hard drives. For example, when exporting the AVCHD files in the beginning, have the AVCHD files on one drive, and export into the “input” folder of MEncoder on another drive. Then, while using MEncoder, encode the “output” files back to the first hard drive again. This can save up to an estimated 15% of time in the overall process because the hard drive’s head won’t jump like crazy back and forth to service both the input file reader and the exporting encoder, but instead each hard drive will dedicate itself to a single action.

2. You will need a lot of storage. We are talking about 180 GBs per 1 hour of footage, because there are two stages of Huffyuv/Lagarith encodings going on. After the export to the “output” folder is done, you can delete the files found on the “input” folder, in order to save some storage space. Always keep around the original AVCHD files though.

3. Editing Huffyuv or Lagarith files can be pretty slow, because their filesizes are so large. You can choose to edit them directly and be done with it, or export again in Cineform 1080/24p/progressive through Vegas/VMS’s .avi custom codec dialog, and then edit these Cineform files instead. It’s pretty fast to re-encode in Cineform, its format is fast to edit too, and because all these codecs are visually lossless you won’t see much of quality degradation after all these exports. But I leave this decision to you as to in which format you want to edit with, because exporting and re-exporting until you reach a fast codec can be nerve-wrecking.

4. Instead of exporting in Huffyuv out of your Windows video editor, you can choose to install the latest “nightly by clsid” ffdshow build and then export in the “FFv1″ lossless codec instead (you select FFdshow instead of Huffyuv in the .avi dialog and then you “configure” the FFdshow dialog to use the FFv1 422p codec). This codec creates much smaller files than Huffyuv does, but it’s twice as slow to export.

5. Vegas Platinum won’t export in full 1920×1080, so you will need this method to go around the problem and export in full 1080/24p after you edit your footage. If you are ultimately exporting for Vimeo’s glorious 720p HD web service, check my tutorials here.