Sony EX-1/EX-3 and Final Cut Pro, What’s Your Workflow?

Late last year we added a new camera package to our offerings at Edit Creations. In part because we wanted to diversify the services we had to offer, and also because it played into a spinoff company we’ve been working on. That new company would be a lot more production based then Edit Creations currently is.

We looked at all the options out there in the sub 10K price range, and after weeks of research ended up with a Sony EX-1. That really surprised me because when we started the search I was pretty much set on the Panasonic P2 format and the AG-HVX200A model. Not only do I have experience with that camera, but almost everything else we do is shot in the DVCPro HD format, so we have a nice workflow in place. More then that though, I can’t stand editing in the MPEG-2 format. It’s fine for shooting and can capture great images, but once you get it into the edit system you can be assured you will rendering more then ever before. But in the end, the Sony EX-1 won hands down in image quality, built in features, and price. There was just no denying it.

So, what to do with the workflow?

On the first several projects we tried multiple different ways of attacking the projects. Everything from working in the MPEG-2 format natively to converting everything to ProRes and editing with the converted files instead. We also have extensive experience working with the Sony software, which includes XDCam EX Clip Browser and XDCam Transfer.

In addition to all of this trial and error, I was training a client on how to handle the workflow for his own project (he was renting our camera). But it seemed every time I met with him I was saying “OK, I know I said do it like this, but now there’s a whole new way to do it”. Nothing like learning on the job to keep everybody on edge!

So here we are now, about 6 months later, with what I believe is a solid workflow. So if you’re using EX-1 (or XDCam footage of any kind) you can learn from our mistakes and start off on the right foot.

In The Field

If you’re new to using a tapeless format, it can be very unnerving to spend an entire day shooting and have no tape to show for. To avoid disaster, redundancy is the key.

A simple but important note – label your media cards!!! Use a sharpie or a nice printed label, but make sure each card has a unique name/number. This will avoid confusion in the field when you drop your cards into a pile of other cards and don’t know which ones have footage to download and which are empty. Seriously, this one little step can save you in the field.

Assuming you don’t have enough SxS cards to make it through an entire day of shooting, you’re going to have to offload the cards and erase them in the field so that they can be reused (see, you’re already freaking out). Here’s how we handle that task.

In addition to the producer, camera person and possible audio guy, it’s ideal to have a AP or junior tech person with you. The handling of the media cards is very important, and getting it wrong can be disastrous. Don’t underestimate how important this is, just because it’s easy. We call this person the Media Assistant (MA from here forward).

The MA has the job of offloading all cards in the field, checking to be sure files were successfully transferred, and then erasing the clips from the card so the card can be reused. The equipment used for this is a MacBook Pro and two portable FireWire or USB drives. You can do the same thing using the PC but we’re all Mac here so I’m unable to provide any details on that side of things.

As each card is filled by the cameraperson, it’s passed off to the MA. The MA will then load it into the Express 34 slot on the MBP and it will mount on the desktop as an external drive (make sure you have the proper drive installed on this machine so that the card will mount on the desktop properly). Then, using a program called Shotput EXPress, the media is copied, simultaneously, to two different external drives (you can do up to 3 different locations. While you can do this manually in the finder as well, Shotput EXPress makes it more automated and does a nice data verification as it copies everything over. It just adds another layer of security.

Once the media has been copied, Sony’s XDCam EX Clip Browser software is used to view the clips on the drives and verify that everything looks and sounds as it should. At this point we’re just spot checking a random sampling of clips. Once the MA is satisfied, the inserted SxS card is erased by selecting all clips and hitting the delete key (this is also done using the Clip Browser software). The card is then given back to the producer or camera person for use again.

Once the shoot is over and the producer is back at the office, the footage is all copied to 2 different full size (and less expensive) drives. One is used during the production of the project by the producers and editors, and the other is a backup, and is stored with the client. It’s essentially equivalent to the box of beta tapes the client would have received in the past. After the media is successfully copied to the larger drives, the mini-drives are erased and ready for the next shoot.

As a side note, don’t cheap out on your hard drives, especially the big ones that are the final destination for your camera originals. A good drive does not cost that much more then a cheap one. Just ask yourself how much your shot day is worth, and if you would be willing to pay to reshoot everything. I highly recommend buying drives from G-Technology. While they’re not the cheapest, they are backed up with the best warranty in the business, they run quite and most of all, cool. On the flip side, STAY AWAY FROM LACIE DRIVES! I can’t tell you the number of editors and clients who have horror stories about drive failures, myself included. We used to have a ton of them but sold them all on eBay just to avoid any additional disaster.

Preparing for the Edit

While the free Sony tools can be used to screen and prep for editing, we have opted to go the route of having a FCP work station for the producer. After doing it both ways, we feel this is the most efficient for both the producer and the editor. (We’re currently looking into setting up a FCP Server workstation to streamline this process even more).

At this point loading, logging and transferring footage is the same is using P2 footage. Using the Sony FCP Logging plugin outlined in a previous SuiteTake post, the footage is loaded into the Log & Transfer window and clips are marked and labeled. As each clip is identified, it’s added to the import cue and converted to QT movies that FCP can edit with. The files you end up are basically MPEG-2 files wrapped into a .mov file. (For faster logging, check out our post on log and capture shortcuts.)

While going through this process, the producer can import multiple clips from a single recorded clip, add clip names and log notes, and organize footage into multiple bins as desired. This is all typically done to a drive designated as the “project drive”, so what you end up with is a single drive that has the FCP logging project with all of the bins and clips, and all of the transferred and transcoded media. This is what is handed off to the editor at the start of the edit.

Once the editor receives the drive, the project file that the producer created is opened, and all bins/clips are copied and pasted into the current working project (or the template project if we’re starting one from scratch). The project the producer creates is treated as a log project only, we do not edit inside that project.

With the new bins/clips pasted into the working project, we then use media manager to move all of the logged footage to our local edit drives and re-link them. We never work off of client drives, and instead opt to always have the footage on our own system. Since we have dedicated raid systems on each edit system, it’s a lot faster and it’s easier to keep things organized on our end. Once everything is copied over and re-linked we eject the drive and give it back to the producer.

Editing


So the main drawback to editing with the XDCam/EX-1 footage is that it’s saved in the MPEG-2 format. Because it’s a long GOP format, there’s not actually a full frame of video for each recorded frame of video. So when editing, FCP has to recreate those missing frames. While on the fly editing it’s usually seamless, the moment you add a transition or anything else that requires editing FCP has to create these new frames (FCP Calls this “Conforming MPEG-2 Video”), and THEN render. It adds a whole new layer of processing and even on a fast machine it’s a slow process.

So, what to do? There are three basic options.


Transcode to ProRes

While logging and importing footage inside of the FCP Log & Transfer window, you don’t have the option to transcode the footage into any other format. However, once that the footage has been imported and wrapped into QT files, you do have the option to convert everything to ProRes using compressor. What you’ll end up with are iFrame based video files that look as good as the original, but are more “edit friendly”. You’ll be able to edit quickly in a ProRes sequence with minimal rendering, and be able to do compositing without watching the image degrade right in front of your eyes. The drawback is you’re adding another step to your process, and depending on the amount of footage you have and the speed of your machine, it can be a very timely step.

Edit Natively in the XDCam EX MPEG-2 Format
For some very simple and short projects, this is actually a good option. Just drag your first shot into the timeline and let FCP match the timeline to your raw footage, and everything will play and edit in real time and you’ll be pretty happy with how it all comes together. But there’s one trick to make it go smoothly.

After you’ve let FCP match the sequence settings to you footage, open up the settings for your sequence and go to the “Render Control” tab. Make sure all rendering is done to ProRes instead of Same as Sequence. This way whenever you do need to render it will take less time and you won’t be compressing back into the MPEG-2 format. This essentially removes the conforming step that FCP would otherwise need to perform, at least while editing.

Exporting your final sequence will go slower then you might be used to, but it’s usually tolerable unless your project is very complex with lots of effects/render files. Because you’re exporting back into the XDCam EX format, it has to convert everything back into the Long GOP MPEG-2 format, which is very slow even on a fast machine.

Edit in a ProRes Timeline
You can also setup your timeline as a ProRes timeline for whatever frame size/frame rate matches your source footage. Once you do this, you can edit with your footage and for the most part play it back in real time (except for some effects) with FCP’s RT timeline. The advantage to working in this way is that since you’re working in the ProRes codec, your footage will continue to look great even if you throw lots of layers, graphics and effects at your project.

The huge down side to this is everything, and I mean EVERY-THING must be rendered at one point or another. Even if you’re able to play things back in real time while you’re editing using the RT features of FCP, when it comes time to print to tape or export a master QT file, you will have to render everything and it can take a very long time, even on shorter projects. You will end up with a very nice, high quality ProRes master file, but if you’re on a tight deadline this might not the way to go.

The Best Way?
After trying all of the above approaches on multiple projects, I don’t think this is a single solution that fits every project. I think before starting each project the best thing to do is look at the details of that project and decide. For simple projects that are short, mostly cuts and dissolves, few graphics and no real layers, staying in the MPEG-2 format is by far the fastest way to go. But if you have a more complicated project with layers, keys, graphics and so on, I would work in the ProRes world so that the image quality holds up though all of the processing.

At least for the majority of projects we do, transcoding everything to ProRes before we start is not really efficient. It takes too long and just adds an extra step that we just don’t have the time for. So in my opinion the it’s just not worth the time penalty.

The SuiteTake Take?
The EX-1/EX-3 and the higher end XDCam cameras have all proven that you can create some great images and record them to the MPEG-2 format. MPEG-2 cameras have really matured over the years, and they’re far superior to the baby brother HDV format cameras.

But understand that if you go this route there are tradeoffs when you get to the postproduction stage. The tradeoffs are not nearly as bad as they used to be and there are now multiple ways to handle the challenges – but they are still tradeoffs.

When starting any project, you should always consider the entire work flow start to end before shooting your first frame of video. If after doing this the EX-1/EX-3 fit your production needs, I say go for it.


About the Author

author photo

Thomas Tomchak

Thomas Tomchak
President and Chief Technology Officer of Edit Creations, Inc.

Tom has been working in the broadcast industry since 1987 and has extensive experience editing corporate, broadcast and documentary programs. Tom has worked with a long list of nationally recognized companies and agencies including, Young and Rubicam, J.Walter Thompson, United Airlines, Better Homes & Gardens, Sears, Lions Clubs International, and Warner Brothers. You've seen his work on CNBC, United Airlines in-flight programming and major market television stations across the country.

Today he continues to seek new business opportunities while working with Edit-Creations' ever-expanding client base.

See All Posts by This Author

There Are 27 Responses So Far. »

  1. Awesome post.

  2. Please clarify an editing question: Can I edit in ProRes timeline with footage that's not transcoded (i,e, MPEG2) and then transcode final program to ProRes? Also, how long to transcode footage to prores on Duall Quad Mac Pro?

  3. Yes, you are quite able to take a non-ProRes video clip and edit it into a ProRes timeline. At the core of it, this is what the multi-format timeline in FCP is all about. You choose what you want your output to be, and have your source footage conform to it. At least this is how we approach it.

    The exception to this would be if all of our footage came to us in a single format, we would just edit natively in that format most of the time. We do this all the time with DVCProHD. But more and more often, our projects are a mixture of SD/HD, tape and disk based, different frame rates and in some cases PAL clips (when everything else is NTSC). So in cases like this, we just keep all source clips in their native format and edit them into a ProRes timeline.

    As for your compression time question, I don't really have solid numbers other then to say it's so long that we avoid it at all cost. It's just not practical in most cases.

    Hope this helps. Tom

  4. So if I took files directly from EX-3 into a prores FCP timeline I could edit without having to render basic effects and then export as ProRes or whatever I wanted? Thanks for your help.

  5. That is correct.

    In fact, I start a job tomorrow morning that is just that. Everything was shot on the EX-1 (same files as the EX-3). I'm brining in all of the footage through the log & transfer tool inside of FCP, and from there it's all being edited into a ProRes timeline. Things like dissolves usually play fine without rendering, but much more then that and you will have to render.

    I normally work through the entire project this way, and set the auto-render to 1 minute. This way, whenever I take even a short break, or switch to any other program to do something else the timeline starts rendering. This cuts down on the perceived amount of rendering that has to be done.

    If you're at all doubtful, I suggest you just grab a few shots and give it a try.

  6. I don't have the camera yet, I am considering buying it. Someone else suggested the following workflow:
    Keep sequence native using EX easy setup. Set Sequence Render to Apple Pro Res.
    This will allow fast editing and rendering without GOP conform.
    Transcoding to another codec can't make something look better. When you're done editing simply use export using Quicktime with the current settings as you would any other sequence.

    Does this make sense? What are your thoughts on the pros and cons of this approach? Thanks.

  7. The workflow you outlined is also covered in this article with my thoughts on it. For certain types of projects it's a great way to go.

    Transcoding does not improve quality, you are correct. But the way it's discussed in this article is purely to get out of the MPEG-2 format for faster/easier editing.

    If you can handle longer then usual export times (something that is problematic on our end due to deadlines) then staying in the native format is probably a good way to go. You can edit quickly with reasonable render times (you still do more rendering though), and just export to QT keeping the settings set to match the timeline. Maybe I wasn't totally clear, but this is what I was suggesting for the “staying native” option.

    I think once you start editing with the footage you'll be more clear on the tradeoffs. It took me a few months of trying each way to finally know which approach to apply to which project.

  8. Thanks for all your advice. I bought the EX-3 last week and very happy. It looks great. I also found that most of what I need to do is fine in the 'staying native' set up. Most of my final product is delivered on DVD. New question: I usually export to Compressor and use the preset Best Quality 90 min. than burn a dvd from DVD Studio Pro. IS this the best way to preserve my image quality or can you suggest something better? Thanks.

  9. That is usually how we handle it as well, and the results have been very good.

    I've recently started looking into using Episode Pro as a replacement for compressor, because I don't think the quality from compressor is always the best it can be. Programs like BitVice do a much better job when it comes to MPEG-2, but that program is more of a one-trick-pony.

    Enjoy the camera. The more we use it, the more we love it. For the money it's hard to beat.

  10. I have read this article and dialogue with interest as I have just shot a dance show with three EX-1's (my first time using this format) and am about to start editing in multicam. I had read this article before shooting and figured I'd stick with the native sequence settings in FCP. However, I hadn't factored in the extra load editing in multiclip mode adds. I have just started experimenting using the native codec in the sequence settings and dropping in a multiclip. It seems very processor heavy and I had to reduce playback quality to low for it to play back without dropping frames. I could probably get used to that until the edit is finished and I have to start colour correction and so on but is there a better workflow for multicam editing with XDCAM footage?
    Also, my client asked me to “run off” a DVD of the show from the back camera so I dropped that onto a separate sequence – again set to native codec – and exported a non-self-contained movie using “Export Quicktime” with current settings and imported this to DVD Studio Pro, encoded and burned a DVD. I was disappointed that whenever there was any reveal or zoom, however subtle, the horizontal lines in the picture would become watery and stay watery until the zoom/reveal ended. Otherwise the pic quality was okay. So my next question is, what is the best route to an SD DVD from an XDCAM timeline?

    Thanks in advance for advice,

    Éamon

  11. I never even thought about staying native for a multi-clip project, but yes I can see how that would really tax the machine. It’s doing a lot of decompressing of multiple streams on the fly. Even a new Mac Pro would be a bit challenged. If you know you’re going to be going the multi-clip route, doing a conversion to anther format might be a good idea. And this would be a good time to ask yourself if you NEED to have your final output in HD. If not, transcode to SD and really help your cause.

    I think your DVD problem might be that you’re doing the encoding inside of DVDSP. While they say it uses compressor in the background, I’ve always found I have much better results using the “best” settings inside of compressor. Then, bring the already compressed elements into DVDSP and burn your DVD. I would be shocked if you didn't see a huge difference.

    Hope this helps. Tom

  12. Thank you Tom,

    Actually, I do intend to go to SD but only after the edit as there is one
    back camera with which we were deliberately conservatively wide and I hope
    to be able to punch in a little on the scale at certain times in the show to
    mask out unwanted things at the edge of frame and bring up the size of the
    action. Nothing huge but I had supposed that there would be plenty of
    latitude for this given the ratio of pixels HD:SD-wise. If I converted to SD
    first and then did this scaling afterwards I know this would not work. So I
    am wondering id there an intermediate solution (ProRes, Apple Intermediate
    Codec etc.) which would make my workflow easier.

    An interesting thing I have noticed is that when I put my multiclip down on
    the native sequence (and this happens whether or not I set the render
    settings to “same as sequence codec” or ProRes), turn off my external video
    and set my RT to Unlimited and Playback to Dynamic or Low, I get a green bar
    across the top of the sequence. This allows me to edit away in multiclip
    once I set the Playhead Sync to Open. However, if for any reason I render a
    part of the sequence or, as happened, I go away for a while and come back to
    find that auto-render has rendered the whole thing, I then have a purple bar
    above the sequence but playback in multiclip no longer works, no matter how
    I mess with my settings afterwards. It took me a long time to figure out why
    it was no longer working – nothing in the manual hinted at this. I had to
    make a new sequence and copy and paste the entire sequence into it to regain
    my green bar and allow myself to continue editing in multiclip mode. Now I
    have auto-render off. It’s troublesome because soon I will have a client on
    my shoulder who will need to see the thing playing in multiclip in order for
    us to make fine adjustments, yet wherever I do punch in scale-wise on a
    camera, I have to render in order for the picture not to be horrible. Should
    I “media manage” the whole thing into a different project with a different
    codec?

    Thanks again,

    Éamon

    LittleVision

  13. If you know that you're going to SD already, here is how I would handle it (and how I have in the past).

    Everything we shoot is HD now, regardless of final output. So there are several times when we just need SD on the output. What we normally do now is create a NTSD (or PAL) SD sequence, set it to 16×9 anamorphic (so we can retain the wide screen aspect ratio) and make sure the sequence codec is ProRes (non-HQ in this case). Now, as we edit our HD footage into the SD timeline it is automatically scaled to fit. We can edit fast, most things will play back in real time (although I have to admit we have pretty new Mac Pros and fiber raids) and if we ever need to do any scaling up of images there is all the room in the world to blow things up and have them look great. When we're ready for output, we render everything and export a self-contained QT movie that's already in the SD format.

    There are a few other benefits to going this route. If you have any graphics to create and render, you only need to do them in the SD format. DVD compression goes much faster since you're dealing with SD files and your SD image will look awesome. We've had great experience with this workflow.

    As to your multi-cam question, I have to say I can only tell you what I would try myself. I have not actually had this problem yet though. Given everything you've said, knowing I'm not only going to SD but I'm also in need of doing multi-cam with a client in session, I would almost certainly have transcoded my footage to ProRes and worked in ProRes instead of native. Again, I say that without having actually done it myself, but that would be my first thought.

    If somebody else has a better (maybe even first hand) experience, please feel free to chime in here.

  14. I recently found your site since I just finished shooting on EX-1 for the first time. Thank you for your articles and constant feedback.

    My question is probably a “duh” question but here it goes.

    I was able to use the “Log and Transfer” to import all my footage into the FCP project file. Now my question is when you start a new project what settings do you use? I know it varies for each type of project.

    We shot Sony XDCAM 1080 30p

    Under easy setup I found and am trying — FORMAT: Sony XDCAM HD – USE: XDCAM EX 1080p30 VBR – RATE: 29.97

    Is this the correct setup or do you recommend a different one or way?

    Our final output will be to dvd for the client. Which obv. wants the highest quality they can get. (Should I follow your previous post/comment as my answer?)

    Thanks,

    Josh

  15. Really? You are letting FCP do a downconversion for you?? FCP is awful at conversions. Edit native in ProRes (That's a big reason this codec was created, to work with “HD” in realtime) and downconvert with whatever means you have upon output. (NOT FCP)
    My 2 Cents,
    Joey

  16. That works great when you capture your footage directly into ProRes. But when you shoot tapeless often your only option is to transcode all of your footage after the fact to ProRes. This is not a fast process, and at least in my world it's never an option budget or time wise. In a perfect world you would just capture straight to ProRes – it is the great universal codec inside of FCP – but that's just not the reality for many people, myself included.

    And my experience with letting FCP size HD images into a SD timeline has been pretty darn good. I'm pretty picky about image quality and signal path, and I have not been disappointed yet. So that's all I can go on. In fact, just recently we had some SD anamorphic footage that was scaled up to 720p inside the timeline, and I was shocked how good it looked.

    We did a lot of testing before proceeding with this workflow, and it works for us and our needs. But everybody needs to do their own test and make that decision for themselves.

  17. I wonder if it's possible to finish the edit then put the programme back on the SXS card. then the camera can play out SDI or composite to any format vtr? Is this a possibility?

  18. Hi,

    I don't know if you are still monitoring this post, but I wondered what your archive process looks like. Right now we pretty much store entire projects to our server which can be quite large. Disk space is pretty cheap, but I'm always looking for a better way to work.

    Steve

  19. I would suggest that you take a look at a post I did on the LTO drive we use. In addition to covering the LTO, it shows how we maintain a long term archival system for everything that we shoot and edit.

    Hard drives are not good for the long term, take my word for it!

    http://www.suitetake.com/2009/04/06/nab-preivew

  20. Why does FCP state in the project window that my XDCAM Data Rate is 4.3MB/sec? Isn't XDCAM HQ 35MB/sec? When I go to capture scratch it says the clips are 35MB/sec so why does FCP do this? And why when I bring in ProRes422 HQ does it list its data rate at 22.5MB/sec?
    Thanks
    Andrew

  21. Why does FCP state in the project window that my XDCAM Data Rate is 4.3MB/sec? Isn't XDCAM HQ 35MB/sec? When I go to capture scratch it says the clips are 35MB/sec so why does FCP do this? And why when I bring in ProRes422 HQ does it list its data rate at 22.5MB/sec?
    Thanks
    Andrew

  22. My guess, I'll chime in,

    XDCAM HQ is 35 Mb / sec (megbit, not megabyte).

    MB = megabyte

    Mb = megabit

    Have to be aware of that

    To get megabytes from megabits, divide by 8, so 35 / 8 = 4.35 MB, so this is totally correct

    ProRes HQ is 220 Mb, 220/8 = 27.5 MB

    ProRes is 145 Mb, 145/8 = 18.12 MB

    From what I hear and have seen in the industry concerning ProRes 422, it is hard to discern ProRes HQ and ProRes. I just use ProRes.

    macguitarman@mac.com

  23. Also, just my 2 cents, but with all the formats out there, why use XDCAM?, which is MPEG-2 Long GOP

    And why buy this camera at 10K

    To me the camera to get is clearly a Panasonic HPX-170, DVCPRO HD, with HD-SDI out. $6K

    Then get the new secret box, the AJA Ki Pro, and record Apple Pro Res 422 natively on a 250 GB HD

    Several issues solved:

    - Goodbye MPEG-2 Long GOP as an acquisition codec, and hello Apple ProRes 422, a full raster codec, very important

    - Goodbye long transcodes in Final Cut Pro from MPEG-2 (a playback codec, an obsolete one IMO)

    - Goodbye expensive P2 cards or SxS cards, a joke really. The Ki Pro 250 GB hard drive is 500 bucks or so, but beats the heck out of $700 16 GB P2 card or SxS card.

    To me this is the new, great workflow, HD-SDI out, Apple ProRes 422 recorded in the KI Pro, walk over to your Mac and drag the ProRes files into a Final Cut Pro ProRes 422 timeline and start editing. Output to H.264. Done. Can even play the outputted ProRes file from the Ki Pro itself to a large screen, in a digital theatre, no more expensive film transfers. IMO this is where the industry is not only going, but already started.

    6K + 3K for the KI Pro

  24. Hi Thomas love your post, as for card labeling this something I to advice to my clients that I train using XDcam EX workflow.

    James Wood
    HD Productions

  25. Cheers, great overview of the process, just have a new EX3 in my hands and getting my head around workflows, your insights are great.

    Tony

  26. What if you’re exporting an edited short video with choma to be keyed in Adobe After Effects?
    What’s the best edited format to deliver it to After Effects for keying?

  27. Uncompressed or ProRes. Always go uncompressed if possible when keying, it
    can make all the difference. The EX-1 is especially good at shooting for
    keying if you record the output via SDI uncompressed.
    On Mon, May 9, 2011 at 12:51 PM, Echo <
    js-kit-m2c-LBPM2L3EI1S8K1764MRJ7118UFOCG8UDPLF7P5BJBIMGH31TJPE0@reply.js-kit.com

Post a Response

You must be logged in to post a comment.