For those not familiar with Sony Vegas, it’s thought to be the geek choice for video editing on Windows. It’s much cheaper than the heavyweight solutions in the industry, but at the same time very powerful and robust. Let’s have a look as to what’s new in its 9th version.
The first thing you will notice on Vegas Pro 9, after installing or upgrading it, is the darker interface. The darker palette allows the human eye to have less contrast between the video and the interface, and therefore make more informed decisions when color correcting or color grading.
In terms of new format support, Vegas Pro 9 has a few interesting surprises for us: it can now capture XDCAM EX footage in MXF format (e.g. from the Sony EX1/EX3 cameras), while it now offers supports for RED’s .r3d format! Sony implemented a pretty in-depth dialog/UI to control the RED RAW format as you add it to the timeline. Of course, in order to open these files, Vegas now had to support video resolutions up to 4096×4096.
On top of all that, there’s even better support for AVCHD footage. DV, mpeg2 HDV and MJPEG also worked as admirably as in the past. The only problem we had with format testing was the slowness, and often crashiness, of the application when editing .mp4/.mov files from HD digirecorders, like the Flip, Kodak, Aiptek, and Sanyo Xacti series. This problem has been plaguing Vegas for years, and I was surprised to see nothing getting optimized in that regard — at least for stability, if not for speed.
On a high note, the 32bit floating point processing feature is now better than ever. Now you can switch back and forth between the faster 8bit processing and the (slower, but better) 32bit processing with a flip of a switch without undermining the quality of your video. It is now suggested that editing can happen in 8bit, but export in 32bit at the end — which is what matters.
And if you like these… big numbers, here’s a good one. Among new image formats now supported by Vegas (e.g. OpenEXR, MS HD Photo, DPX), the app can now load images much larger than a gigapixel! This allows to zoom in to a picture, and crop, without introducing pixelation.
On the more user-visible features side, there are now new plugins and transitions added to the existing collection. Users can now use visual effects like the ones found on After Effects or FCP: glint, rays, defocusing etc. In fact, I smiled when I saw the “rays” plugin announcement, because just a few weeks before that, I was watching a music video (directed by online friend Mark Pontius, drummer of the band) using the same plugin via FCP, and I was thinking to myself “hey now, how could I compete with this? I can’t!”. But Sony had this surprise under its sleeve for me. Also, Vegas comes with a powerful SDK that allows third parties to develop their own plugins.
Vegas Pro 9 comes with DVD Architect 5, which supports full Blu-Ray authoring, in mpeg2, AVC, and at frame rates of 50i, 60i and 24p. The Vegas Pro application itself can also burn the current timeline in an “AVCHD disc”, meaning that you can burn HD video on a plain DVD disc using a plain DVD writer. AVCHD discs are supported by some Blu-Ray players (but not all).
Finally, the 64bit version of Vegas Pro 9 has seen a refresh with some fixes, but the fundamental problems it had with version 8.1 remain: no access to [most] 32bit plugins and codecs. I personally run Vista 64bit on my beefed up video editing station, but I prefer to still use the 32bit of Vegas.
Vegas Pro is a very powerful app for the money, as I stated above, but there’s still room for improvement. For example, with Vegas 9 there is no good intermediate codec shipping with it (older versions were licensing Cineform), so I wouldn’t mind seeing some (optimized) Avid DNxHD support in the future. The current implementation for Avid’s free codec doesn’t preview faster than 7 fps, which is too slow to edit properly!
There is no proxy editing support, and one has to resort to scripts or complicated workflows to go around .mp4/.mov editing. Also, doing frame adjustments is a pain: retiming footage, removing/adding pulldown, or slowing down/speeding up footage more than 4x is a matter of long tutorials that the user has to hunt online and carefully follow.
Maybe my biggest gripe with Sony in general is the fact that they haven’t released an update for their consumer version of Vegas, Vegas Platinum 9, to fix their 100% reproducible HD SonyAVC/Blu-Ray crashing during exports. Hopefully in the near future.
Overall though, both Vegas Pro and Platinum 9 remain the best solutions for people with a strict budget but who require immense power and flexibility during editing. Personally, I always suggest Vegas to people who ask me which editor to buy. Sure, Vegas has its problems, but most other PC editors in its price range have an order of magnitude bigger problems than Vegas, so the choice quickly becomes clear (especially if you are using standard camcorders and not digirecorders).
Rating: 8/10
What about Reds? Do they still start at $20000 or was there a cheaper model? I kinda lost track.
The cheaper RED Scarlet models are expected sometime this or next year. The current basic RED can be obtained for $16,000, but by the time you add all the extras you would need, and lenses, a more realistic price point is $25,000. Still, way cheaper than its actual competition.
It is now suggested that editing can happen in 8bit, but export in 32bit at the end — which is what matters.
That doesn’t make much sense to me. Wouldn’t it be better to edit in 32 bit and then downsample the final result to 8 bit for smaller file sizes?
I don’t work in video, but when retouching photos I tend to work in 16 bit and then downsample to 8 bit JPEGs when I export.
I believe that Vegas can re-apply all filters in 32bit when you export, even if you did your editing in 8bit. So, for speed, you can edit in 8 bit, and just before you render out, you change its project properties to 32bit. Then, you export, and Vegas will re-apply all filters in 32bit resulting in less quality loss. However, you can choose to export in an 8bit format (e.g. mp4), or a 10bit one (e.g. Cineform).
From what I know, unlike Photoshop, Vegas doesn’t keep a stack of cached data in 8bit, but it re-calculates everything just before exporting. It’s not real time like Photoshop is when editing. AFAIK, it wasn’t like this in version 8 (it was like in Photoshop).
Here’s the part that’s confusing me. If you’re editing in 8 bit, having you already lost 24 bits of data? Then if you export your 8 bit project as 32 bit, aren’t you just upsampling 8 bit data into 32 bits? In this case you won’t be gaining any advantage at all since the end result is effectively an 8 bit movie pretending to be 32 bit?
You are thinking in terms of a single frame and all the processing that has been gone into it. What I am telling you here is that video is not as real time as a still image, and so when you export, you can tell the editor to re-do the processing from the very beginning. So it will treat it as 32bit from the get go, without upsampling (provided your source footage is 10bit, of course). Vegas re-processes each frame when you export, so there’s no reason why not treat it as 32bit and redo all the plugins processing at that very moment. Although I could be wrong, this is what I get from it.
Edited 2009-06-03 07:00 UTC
Thanks Eugenia, that makes sense now.
I am no movie director by any stretch of the imagination, just doing editing of AVCHD video files of my family on Platinum 8. I’m probably not stretching it. I have a decent dual core AMD with 3Gb on Vista 32bit, but rendering can take some time (obviously)
Is there much of an overall gain moving to 64 bit with Vegas? In overall usability terms and speed of operation?
My thinking is the move to Windows 7 in a few months time might be 64 bit and a good time to upgrade to Platinum 9 64 bit?
I explained it in the article. If you are using *third party* plugins and codecs, these won’t work with the 64bit version of Vegas. Plus, there is no 64bit version of the Platinum version, only 32bit. And, the speed bump is extremely small. In fact, on my Vista 64, it was a bit faster exporting with 32bit rather than the 64bit version of Vegas. So I still run Vegas 32bit.
Finally, yes, it takes a long time to export in HD. However, you can speed up your editing/exports by upgrading to Platinum 9. AVCHD was almost re-written for v9, so it’s faster to decode. Also, use my tutorial to setup the project properties (something most people don’t do properly), and export in HD: http://eugenia.gnomefiles.org/2007/11/09/exporting-with-vegas-for-v…
Many thanks for the overview. When I had tried Vegas myself a couple of years ago, I had found video navigation quite awkward. I couldn’t find any shuttle controls, (where you can play the video at variable speed in both directions) and scratching (moving the time line pointer) was quite slow with a noticeable lag. Navigation is much more intuitive and responsive on both Premiere Elements and Final Cut Express. Have things improved in this respect?
Thanks!
Edited 2009-06-03 13:46 UTC
Actually, Vegas has these controls and it’s faster to respond that Premiere or FCP. It’s just that you didn’t read the manual it seems. So, to control frame operations you can use the JKL keyboard buttons, and/or the numeric keypad.
Now, regarding the timeline pointer. You did the same mistake as many other people who come from other editors: you clicked and dragged the “head” of the pointer rather than the line itself. In Vegas, the head only scratches the video (so it feels like lagging), while the line’s movement is actually free and super-fast (faster than in most editors).
Please give it a try.
This is a little off topic. But, the mention of AVCHD and Vegas Pro is timely for me because I’ve been trying to figure out how get my clips off my camera for viewing.
I have an hg20 camera. I use a Mac and Linux. Simply viewing my video clips has proven to be a difficult task.
I can archive my video for long term / future use by imaging my camera’s disk. But, I don’t have a way to watch my clips without encoding them first, which takes a tremendous amount of resources. So, copying AVCHD to something like DVD is appealing to me. I’d like to be able to dump AVCHD files onto a DVD (like mp3’s) and let the hardware decode it. But, that’s proven to be not so straight forward either.
I’d be happy if I could just create DVD quality video from my clips. Currently, I’ve done it with iMovie and iDVD. But, it takes hours to create a few minutes of video and the quality sucks, too.
There must be an easier way.
Any suggestions?
Edited 2009-06-03 17:52 UTC
The main problem with the Mac is that the Apple editors won’t load AVCHD clips to be loaded “as is”. They have to be imported from a camera, otherwise the editors refuse to load them.
With Linux, it’s as bad: both ffmpeg and mencoder have massive bugs with interlaced h.264, and they think that a 60i video is actually 60p. This destroys motion. There are some long ways to deconstruct the streams to elementary streams and then feed them to ffmpeg, but these ways will take even more time than burning a DVD.
So, I’d suggest you install Windows and Vegas Platinum 9, and do all your conversions in it. I know that you might not want to touch Windows, but you can definitely have a PC station *just* for video editing. That’s what I do too (my laptop is running either Linux or Mac but my video station runs Vista 64 because of the kinds of video utils available for it).
If you are not happy with that, you can just continue burning DVDs from within iDVD, and archive the original .mts files on DVD discs (as files) and hope for better support in the future.
Thanks. I was afraid of that.
I’m new to the world of video formats. I don’t really want to be a pro at it. I just want to video my son playing baseball, then play back the clips for teaching/learning.
I own a Canon HF11. I have found many ways to encode/view the AVCHD files in Linux, using mencoder, ffmpeg, mplayer and ffplay. I have also tried Handbrake and several of the front-ends to ffmpeg and mencoder, with some success.
For me, the two biggest problems were:
1. making a clean, quick “pull-up” (pull-down removal) of the Canon AVCHD 24p footage;
2. making sure that the transcoded footage will work with whatever NLE (non-linear editing software) is to be used.
There are lots of ways to “pull-up” the 24p Canon footage with mencoder. The quickest and easiest way that I found is this command:
This command makes a clean 24p Quicktime file from the AVCHD, with full 1920×1080 resolution, and it often encodes faster than realtime playback. The “-fps” argument must be set to “60000/1001” (or 59.94, or 60), because of the AVCHD framerate bug in mencoder (and in ffmpeg). Unfortunately, the audio sounds perceptibly different from the original.
Here is the same command using the H264 codec:
Two Final Cut Pro users have failed to put the encoded files on the FCP timeline (however, these results are not conclusive, because these FCP people were not very technically savy). The Perian utility http://perian.org/ might be required to use/view these files on Macs.
I also read somewhere that mencoder’s lavf settings can make it difficult for some video players/NLEs to interpret the files.
Here is the pu-avchd “batch” script, that uses the first command above:
Of course, you should change the “.mts” references to your AVCHD file extension.
Here is a more time-consuming method to achieve the pull-up that definitely works on Macs with Perian loaded:
http://www.avsforum.com/avs-vb/showpost.php?s=ea20de2f2cb829f2821af…
For non-24P footage that doesn’t need a pull-up, one could probably just change the output framerate (“-ofps”) and remove the pull-up settings (pullup, softskip) to get good results. Trying a “non-lavf” demuxer/format might also be worthwhile with non-24P footage.
Furthermore, without the need for a pull-up, one can use ffmpeg or something like Handbrake.
Here is a post about the interactive hdffxvrt batch script, that uses ffmpeg:
http://www.avsforum.com/avs-vb/showpost.php?s=ea20de2f2cb829f2821af…
It takes longer to encode, but yields nice, deinterlaced video and audio, and one can select the final resolution.
Note that this script is set to work with the PAL camcorders (25fps). Here is a post about a modified version of hdffxvrt, set to work with 60i cameras:
http://marks.org/avchd/hdffxvrt-mov1-8-09
Here is the same modified script, set to create AVIs: http://marks.org/avchd/hdffxvrt-avi1-9-09
There is a much more involved script that uses xporthdmv, ldecod, yuv4mpeg and ffmpeg. It is described throughout this thread:
http://www.avsforum.com/avs-vb/showthread.php?s=ea20de2f2cb829f2821…
To get AVCHD files to play in mplayer, I use:
This allows me to actually “see” what is in the file, but there is a lot of stuttering. I get similar stuttering when using ffplay (no “-demuxer lavf -vc ffh264” arguments).
A lot of experimentation and tweaking remains to be done in finding the best mencoder/ffmpeg settings for AVCHD. Hopefully, the framerate bug will be fixed soon.
Hope this helps.
Edited 2009-06-03 22:27 UTC
Unfortunately, this is a pain in the bum solution. I have written a number of tutorials myself too about AVCHD pulldown removal http://eugenia.gnomefiles.org/2008/01/04/canon-avchd-24p-pulldown-r… but I am afraid the free tools are too much of a pain. If you want the job to be done _right_ and really fast, you might need to buy Cineform’s NeoSCENE Windows utility for $130. It costs, but it does the job right.
Exporting in mpeg4 has major quality loss btw, you should try Huffyuv or Avid’s DNxHD. Ffmpeg now supports DNxHD.
A single mencoder script that batch encodes faster than real-time playback is “a pain in the bum?”
There are many solutions given in the above post; some are simple and some are more involved.
I have heard that, that is why I gave an H264 version. However, I have yet to notice artifacting or a loss of quality in the highest bitrate mpeg4 footage transcoded from my 25Mbs AVCHD.
Couldn’t get Huffyuv to work properly, but the ffvhuff/YV12 flags worked. However, I was unable to make mencoder yield a Quicktime file using lossless codecs.
Thanks for the info. I’ll try it soon.
Edited 2009-06-03 22:49 UTC
Having to deal with 60p and mpeg4 is not ideal. Personally, I wouldn’t consider this a solution. It’s just “the best we can do under the circumstances”. But it’s not acceptable by my standards. When I process a file, I need it to be in a good shape, so I can edit it, color grade it, etc. For that, I need the right utils. If these cost money, I buy them. Unless a free solution really works as it should, I can’t consider it.
For example, if I want to resize a picture, GIMP is good enough for it. But if I want to work in 16bit, then I buy Photoshop (even the latest GIMP doesn’t work properly in 16bit). So I just use the best tool for the job, free or commercial. If I don’t have the money to buy the best tool for the job, then I might just not shoot in PF24 mode, or just not use an AVCHD camera in the first place. Everything, all the aspects of it, has to work as I expect it to and not lose quality, otherwise it’s just not a good solution.
So, in your case, maybe you should have not gone for an AVCHD camera, and maybe you should not be shooting in PF24 in the first place. I rarely do (and I even own Cineform’s tools that are able to remove pulldown properly and save in an intermediate format that is really fast and has no quality loss!).
Using the the pu-avchd scripts, I don’t have to deal with 60p, and using the H264 version of pu-avchd I don’t have to use mpeg4. Both scripts yield true 24P Quicktime files, which can be edited, color corrected or processed just like any other video file.
Again, I haven’t seen much of a quality difference between the highest quality mpeg4 files and the raw, AVCHD files, even though a couple of people have told me otherwise.
I haven’t spent much time trying a lot of other codecs/containers, but I would imagine that most of them will work in mencoder and ffmpeg. Handbrake-gtk gives most of the same capabilites, with a gui (but it has limited pull-up options).
It takes me a long time to figure out the best settings for mencoder/ffmpeg, just because I am experimenting blindly to find the ideal settings. However, once the best settings are found, one merely types the name of a script in the command line, and a whole batch of files are automatically transcoded.
I have tried to give files encoded with lossless codecs to FCP users, but they don’t seem to try very hard if the format is not something familiar.
By the way, for those who want to edit their photos in 16 bit and who don’t like the Gimp, try Krita. Or, for anyone wanting to edit photos in 32 bit, try Cinepaint — it has had 32 bit capability years longer than Photoshop.
In regards to choosing not to shoot PF24, the decision is ultimately that of the director (and/or producers) — I am just the cinematographer. My urgings to shoot 30P mean nothing when those running the show have the notion of an eventual film transfer. Most of them cling to that notion.
Regarding your comment about perhaps it would have been better if I had not purchased an AVCHD camera, I am happy with the HF11, but, of course, I will be happier when I can merely handover raw AVCHD files to the clients.
H.264 is mpeg4 btw, so it’s a lossy codec. For intermediate usages you should be using lossless codecs.
Well, yeah, H264 is mpeg4 Part 10, technically.
However, AVCHD uses H264, so AVCHD is technically mpeg4.
However, the images produced by my mpeg4, AVCHD camera are gorgeous, even with significant movement and whip pans.
So, although it is often ideal to use lossless codecs, I wouldn’t define mpeg4 as “lossy,” per se. A lot depends on the degree of compression/bitrate used.
One can use a codec that is not lossless, but that is also not necessarily lossy. Such is the case with AVCHD and with transcoding AVCHD footage to maximum bitrate mpeg4.
The part you don’t understand is that re-encoding again in h.264 is massive loss of quality. Same for mpeg2. It’s one thing for the camera encoding in h.264, but re-encoding yourself in that format for editing, is a bad idea. Just export in an intermediate lossless codec, then edit, and then at the very end, export at h.264 your final masterpiece.
I understand what you are saying, and I appreciate the warning. You are the third person to tell me.
However, I just can’t see the problem… literally.
When I first heard of it, I spent almost an hour comparing the original AVCHD footage with the mpeg4 and H264 footage, and I did notice a very slight difference, only occasionally, in the mpeg4 footage encoded at maximum bitrate. I even went through the clips one-frame-at-a-time.
I couldn’t tell any difference at all with the H264 clips encoded at 24Mb/s.
I can certainly encode to an intermediate codec, and I am definitely going to try ffmpeg’s DNxHD codec. For FP24, I’ll have to figure out how to do the pull-up in mencoder/mplayer and pipe the raw video into ffmpeg for encoding to DNxHD. If FCP can use the resulting files and if they aren’t too unwieldy in size, I’ll probably switch to a batch script version of that method.
I have used Sony Vegas myself for some video editing and it’s indeed a very decent editor, especially for the money. When I tried out the demo versions, I found the Platinum version actually better than the Pro version, because it’s less cluttered. While using Vegas 8, I had a couple gripes with the software, it would be great if you could comment on those with regards to Vegas 9:
* Apparently no AVCHD smart rendering. I would expect the software to copy unmodified content between keyframes and only re-render the changed parts from or up until the next keyframe. The content was 1080i from a Sony SR-7.
* Very poor thumbnail caching. With lots of RAM still available, the software keeps throwing away already decoded timeline thumbnails. It appears to keep thumbnails for about 2 or 3 times the current view range. When you scroll further out of that range, it reloads the thumbs, although it could have cached them from displaying them just a little earlier.
* Very poor project loading times. I have written video editing software myself, and this software could open projects with about 25000 individual playback items stored as files with attributes (in BeOS) in about 1.5 minutes on a Core 2 Duo @ 1.8GHz. Yet Vegas takes several minutes to open a project with about 40 clips in it on my Core 2 Quad @ 2.4 GHz. I have no idea what it’s doing during project opening.
* Unneccessary identification of clips for drag&drop operations or right-clicking in the import panes. Just select a couple of clips there, right click them and watch the software freeze completely for almost a minute while it decides if the clips shall go into your context menu.
The above items are in no way meant as Vegas bashing. As I said, I use it myself and I think it’s got a lot of great and elegant features. I would just like to know if there has been any improvements on these not so good points I mentioned.
Thanks for a great article, BTW!
Yes, Vegas now can smart render AVCHD, as it can do HDV and DV.
Thumbnail cashing is the same.
I don’t have a problem with project loading times. If it happens for you, maybe you have a broken project file, or maybe Vegas needs to recreate the audio caches for it.
I am not sure what you mean with the fourth point.
The re-generating of audio peaks could indeed have something to do with it. Is there a way to avoid that?
No, it’s needed I am afraid…