No Script, No Fear, All Opinion
RSS icon Home icon
  • Q&A – Video Editing

    Posted on July 19th, 2013 EEVblog 27 comments

    Dave answers some question from his 5 hour live show about getting help with editing video for the blog, a co-host, slow Youtube uploads, and big productions like the Ben Heck show.
    FYI, the latest Kodak camera teardown was 6.1GB of raw video data.

    Be Sociable, Share!
     

    27 responses to “Q&A – Video Editing” RSS icon

    • I totally agree with you wishing to do your own editing. Some things just cannot be done by someone else. You are your own best critic and like you say “understand your intent.”

    • The first problem can be overcome by uploading a low resolution copy for editing, then simply applying the editing marks to the full quality video at a later time. It doesn’t make a difference with the other problems, but it would allow fast transfers.

    • For faster uploads you may be able to proxy your connection through another computer.

      If the problem really is specifically with uploading to YouTube from Australia, passing through a well connected third party should fix the problem. You can rent a server on Amazon EC2 by the hour. Or even better on Google Compute Engine wich should be right inside Google’s network.

      You may even be able to use your existing web server if it has spare bandwidth.

    • Appreciate what you are doing, and the fact that you can self produce is awesome. The details of behind the scenes helps keep it in perspective. Here at work when we have down time, Dave’s nerd porn helps pass the time, and if we aren’t careful we might learn something.

    • To the commenter @ ~ 18:41 re: “Use a Remote Desktop Session to eidt in the NLE”. HAHA! So you’ve done that how many times? Upload a video demo of that please. You’ll probably have it ready by about… yeah I thought so.

      Isn’t it amazing how many experts there are on every possible subject on the “Internets”.

      Much respect Dave, keep doin’ it your way!

      • Thanks for bringing it up, I though I would just forget this, but, I rather like to clarify my point again, just see if it could be helpful someway. If not, it doesn’t hurt anyway lol~

        No, I was talking about non linier edition or time code stuffs, but something relative to the traditional remote desktop that has been integrated in Windows, or frankly, now we can call it cloud computing.

        What I do sometime, is using my laptop to “borrow” my desktop’s performance. Long story short, it’s the same as the traditional remote desktop function. If it’s too hard to understand, just think it as a kind of cloud computing. The client computer is acting just as a remote screen, every thing is done in the remote computer. If you open a file, it’s opened in the remote hard drive; if you run an app, it’s run in the remote’s memory, if you save something, it’s saved in the remote hard drive. All your computer on hand, is just acting like a screen.

        Let’s say a 10GB raw file is stored in a video edition desktop in your office, the editor is on the other side of earth with a laptop. Let’s say the editor uses the laptop to establish the remote desktop connection to the video editing desktop, so, now the editor can see 2 screens. Please just ignore the laptop’s local screen, enlarge the remote one, play it as it was the local system. Now you know the rest, right? OK, if not, then the editor can launch a software in the video editing desktop; load the raw file, which is already in the desktop’s hard drive(remember these are all done in the desktop with bandwidth limit or so), then, view the video and edit it(for viewing a 1080 FHD remotely, you’d need something like >3MB bandwidth, something like a home-use 10M/3M cable or higher can handle it with ease); then render it, TO THE DESKTOP’s hard drive. If the editor has a very poor internet connection, then just use 800*600 or so resolution on the desktop so requites less bandwidth for streaming the screen, as the purpose for this is just to know what’s going on at the editing machine’s screen. After all, upload the video to youtube or wherever directly from the editing desktop.

        Yeah, there may be about 130~200ms lag, which might be annoying for operations like dragging the time position bar precisely, but, 200ms for this case really doesn’t matter much IMO.

        I hope this is clear enough, and can clarify the “actual raw file doesn’t need to be transmitted, only the operating commands and the screen.” Even just transmitting a horrible 800*600 resolution screen, i think the editor can well understand the content and drag/cut the video, as the editor barely need to say anything or see the specs sheets or so :)

        Last, no need for anger or pissed of or whatever, we all come here just to view some videos we like at the moment. If I missed something important, feel free to say so as always been lol~

        PS: if the average size of the raw file will somehow become bigger and bigger in the future, I really think this can solve it. Anyway, C+V a single 10GB file with a HDD is just about 1 or 2min….Those USB3 with >100MB write speed are still too expensive for me lol, with that money, I’d rather like a 1.8″ SATA3 SSD instead.

        • typo correction: the “was” at the beginning of the 2nd P should be “wasn’t” sorry if it’s caused any confusion~

        • Ah, sorry for more typo

          “all done in the desktop with bandwidth” is missing a “no” too….

          and for the internet bandwidth, I was literally thinking M Byte for streaming FHD raw file real time…ouch! Anyway, for doing remote desktop, any broadband on earth should be able handle it.

          • Hi Goophy, My aplogies if I came off a bit condescending, and thanks for the detailed explanation. I currently have about 6 years of almost daily experience with remote desktop in a graphics oriented environment with mostly Windows, some MAC and the occasional Linux clients. I manage the IT for a small organization with 3 locations. While it’s fine for managing servers and such but anything with graphics… forget it. My users run 3D CAD, Machine tool editing and FDM print management apps. When I need to remote in to a workstation where the user left one of these apps on or need to launch a graphics app to verify an update the lag is not ms but tens of seconds to minutes, even without a file to render. This is true even with > 3Mbps upload bandwith at both ends and regardless of screen resolution. There is simply no comparison possible between the speed of a Graphics card output to an LCD (where 8 ms is considerd slow these days) to the graphic content of a remote control stream. Even under the best conditions the remote screen draw is painfully lethargic and that’s true even when the screen is only redrawing the active region. Reducing color depth will help speed more than reducing resolution, but it also renders even basic graphic content useless. Sorry to be the bearer of bad news but the kind of dream that you advocate may be attainable once we have fiber along the full path, but the remote desktop apps will also need to improve as well.

            • Hum, that sounds like some other un-revealed problems happening in your case… Yeah, the local graphic card’s output weights tons, but all the remote user needs is about 30fps of basic screenshots. You know this is true just by thinking, right? And, audio is not a problem.. In fact, this “remote editing” isn’t new at all, people haven been doing this for years…Just google it if it’s seemed new, there are always new stuff to people in different fields :) .. PS: I believe this method should become more popular pretty soon as the progression of cloud computing.

              As Dave or the editor really don’t need to view the HD version to understand the content, a so so quality video + audio would be totally fine.. Doesn’t hurt to try if Dave really wants lol~

              • Yes I understand that people are doing it. That doesn’t mean they are having fun though. I might even do it if someone paid enough (hourly is best for the slow stuff). :)

              • Yes, remote desktop is used by countless people daily, but NOT for real time video editing. If you know about video editing then you will know why.
                Please cite one example of someone who is able to do this, and do it using a standard ADSL2+ connection.

        • It is simply not going to be usable for video editing, which requirements instant seamless performance.
          I spent several years on and off remote desktopping using ECAD. It was horrible. It worked, mostly, but was barely usable.
          Video editing is MUCH more intensive and reliant upon real time performance.
          Please try it for yourself and see if you can do proper video editing using remote desktop, I’d be willing to bet you can’t.
          And I have never heard of anyone being able to do this successfully.

        • “Yeah, there may be about 130~200ms lag, which might be annoying for operations like dragging the time position bar precisely, but, 200ms for this case really doesn’t matter much IMO.”

          This makes all the difference in the world actually, when it comes to editing video.

          How do you account for the audio? I find the audio tracks really drive the editing process more than the video track.

          Have you ever edited video in this fashion?

          • If the “remote desktop” software itself can’t handle audio very well, then use separate thing to handle it, like skype or so to run in both machines background. This only works if the audio lag is due to the software’s performance.. Besides from doing all stuffs in editor’s own equipments, there are planty of 3rd parties doing this, i tried wevideo.com once, works, but needs subscrition fees for rendering HD versions, and they limit the raw file’s size……These web-based services are essentially the same as remote desktop, and pretty much like an advancer browser-based video editor, just not on your own remote desktop anyway(am sure many people have used online mp3 editors(like mp3fiber.com) and online picture editors(like pixlr.com), right? Same concepts)….. The web-based stuffs are more like the “stream a low quality to your screen, then send back the time code” method, they’re easier to use, needs less configuration/treaking… A great example of video online editors is the built-in youtube editor, where you can edit the video after uploading it to a youtube’s server, then the server will render it. Of cause youtube isn’t intending to provide you a advance editor, so you can only do things like adding captions :(

            I’m not in the video editing business, i know this method just because i’ve done it very rately, not even for HD videos. Most of the work is just to watch it, delete some parts(from x second to y second), output it, then upload it to online storage for sharing.. Honestly, all I’ve done in this kind of remote manner is just cutting few parts that I don’t wanna share, maybe you guys are doing advancer stuffs that requite supper lagless control? The most annoying thing i ever noticed is draging the time bar; yeah, apart from that, it works. Maybe just find another software that you can enter exact time for cutting/swaping video parts?

            PS: just unscribed from this thread as it seems to creat too much anger or whatever…There should be many articles on the web descripting this fashion since the revived cloud computing, and the effort microsoft has made in win8… Anyway, it doesn’t effect the fact that i do frequent eevblog and like some of the segments really much(expecially the mail bag BTW, though the recent one(s) contain too much post cards for me lol I may download them for screensaver if they’re availabe in HD in a separate eevblog section, but not willing to spend 10min watching them…).. and, pretty disappointed that Dave kind of dropped the uSuuply project; so far, that’s the only gadget i’d buy from eevblog i think…

            • I don’t think anyone is angry here. I know a little about Remote Desktop, and I know how painful it would be trying to edit a video like that. Keep in mind that Dave has mentioned several times how slow (relatively) his uploads are.

              You’ve mentioned how you can edit videos on YouTube, etc. Let’s say his connection is 10 down, 2 up. Once the video is on YouTube’s server, this might be halfway manageable. But it isn’t – it’s in his house or office where the throughput is a fraction of what YouTube can provide.

              Concurrently running Skype for audio is ridiculous. Before you we’re trying to deal with a laggy timeline bar. Now you are trying to deal with laggy audio that isn’t in sync with the laggy video. I think you’d have to put that editor on suicide watch after the first week. :)

              I’m not mad, I just think you’re wrong. It’s ok.

    • Hi Dave,

      Please keep the things the way they are! The video quality is very good and the info is precise. We don’t need any mass-media like gingles or editing. With them, your videos would loose the “ne bullshit” devise.

    • The bottleneck of most USB sticks is not flash but the controller and USB2. Get USB 3.0 stick and enjoy 44 MB/s writes, get USB 3 HDD and enjoy 80 MB/s writes

    • I think even if the technology would allow you to have someone other editing your videos
      and even if you meet someone which has an eletronics background
      and even this person is also a fun-loving guy which keeps your unique humor in your videos…

      I really like the way you are doing it now(!) with your unique non-scripted overly enthusiastic style!

      So i don´t really know what the benefit of a (co)editor would be..!?

      Even if it would save you time and/or enhance the quality of the lecture/narration…

      I like the way it is:
      - High video quality
      - Practical explanation of electronics
      - A nerdy humor which makes me laugh like no “generic” humor could do
      - spontaneous topic selection
      - vivid presentation style

      And if i don´t understand something i can look for other youtube videos explaining it different or more thoroughly.

      Ok i could rant about this until your database backend cuts out my typing at maybe 2048 chars…

      The point is: Your are doing a damn good job.

    • Dave,
      Is there a copy of the entire 5 hour live show somewhere? I would like to see it.

    • Dave,

      fwiw: Rendering w/ GPU vs. CPU only. Your previous video on constructing an editing machine got me to run some experiments on my own system (intel x980@ 3.33GHz, 24GB ram, with an AMD “high end” graphics card – ATI Radeon 5800 series) running Sony Vegas 12.0. My raw video was from a Canon HG10 AVCHD 1080×1440 file with a running time of 7:38 and 883MB. If I render using the CPU only, which uses all 12 hyperthreads and about 90% of the CPU, it takes 14:30 to yield a 458MB mp4 file. If I enable the GPU (OpenCL), the render time is 3:30, approximately a 4x decrease, file size is 459MB. After handbrake, I have a 134MB file, and a total of 5:30 in processing time.
      So, I don’t know if this applies in your situation, but I thought I’d share my experience, fwiw…

      Bob

    Leave a reply