Forum Replies Created

Page 21 of 51
  • Erik Lindahl

    December 10, 2009 at 3:29 pm in reply to: How is this Gillette ad made

    Looking at the spot again you might be able to do a similar trick simply with AE and Mocha. Depends on the amount of “curved 3D objects” you’re working with.

    ————————
    Erik Lindahl
    Freecloud Post Production Services
    http://www.freecloud.se

  • Erik Lindahl

    December 10, 2009 at 3:27 pm in reply to: How is this Gillette ad made

    Looks to me like they’ve done a 3D track of the face, recreated a wire-frame effect in 3D and then comped it together in comp-app X (After Effects, Shake, Nuke, Flame or what ever). I have a hard time seeing you do that effect with out some kind of 3D back-end. You MIGHT be able to do it in Flame since I think it has a more proper 3D engine in it with a 3D track or / or frontal projection. Possibly Nuke could manage this as well, not sure about it’s 3D capabilities. Fully doable with something like Buju, Maya and AE if you got the know-how.

    ————————
    Erik Lindahl
    Freecloud Post Production Services
    http://www.freecloud.se

  • Erik Lindahl

    December 4, 2009 at 2:41 pm in reply to: SNEAK PEEK: Adobe Mercury Playback Engine

    Interesting, clearly. I was hoping to see this happen in FCP7 but we weren’t that lucky. CS5 might be when I jump and try a new editor (Premier CS4 doesn’t quite feel ready for prime-time, at least not on the Mac).

    I wonder if this technology will suffer from the same limitation say Motion has. For some projects the GPU acceleration is amazing, for some it’s just a limitation (i.e. horrible rendering times). I like the above poster would love to see this technology also merge into AE as much as possible.

    It’s a bit sad they go with CUDA as their code-base but I guess OpenCL wasn’t really ready at the time this started to be developed.

    ————————
    Erik Lindahl
    Freecloud Post Production Services
    http://www.freecloud.se

  • Erik Lindahl

    November 26, 2009 at 11:17 pm in reply to: Exporting format for Pro-tools?

    Using MPEG4 or h264 sounds a bit iffy in my book unless it’s just used as a transport codec. There can also be quite iffy in terms of frame accuracy.

    I’ve always sent Photo JPEG (half PAL res) or DV-files (full PAL res) to all sound studios that use ProTools for a video-reference, a AIFF-file for the a sound-mix reference and an OMF for the RAW sound-files. Older ProTools liked MJPEG when they used Igniter cards but asaik those aren’t used anymore and the AVID hardware prefers something like DV.

    ————————
    Erik Lindahl
    Freecloud Post Production Services
    http://www.freecloud.se

  • Erik Lindahl

    November 22, 2009 at 10:49 pm in reply to: viewing RED 25fps with Kona 3

    I haven’t tested native RED-file output through the Kona 3 with FCP 7 but I have experience with FCP 6.

    1. I haven’t NOT be able to get full 25 fps output to work with native RED files in a ProRes timeline as adviced by AJA in FCP (correcting the values to 25 fps instead of their 23.98, so I’m breaking the system a bit)
    2. Working with 2K ProRes files for Offline works fine but that’s a totally different cup of tea.
    3. Working with again the native RED files in COLOR works but I haven’t gotten realtime playback there either.

    The above is on a FCP 6 / MacOSX 10.5 system.

    I had a talk to one of the AJA guys about this problem at IBC this year and his comment was “I’m using RED wrong” in the fact that people don’t expect 2K to be 25 fps, 2K is 24 or 23.98 fps. I’m stunned this is the case seeing PAL countries 25 fps would be far smoother in every aspect, but I might have missed something here.

    His recommendation was to convert to ProRes and stick with that.

    After this I have however at home been testing FCP 7 on Snow Leopard and well there’s a HUGE difference in regards of native RED performance there. Either FCP7, Snow Leopard or the newer RED-driver makes something spin, not sure what, but the difference is staggering.

    1. Full 25 fps playback in FCP. Given this isn’t output through any Kona card, I don’t get this at the studio with or with out video output enabled.
    2. Working with 2K ProRes ofc works as before.
    3. Working in COLOR with Native RED files now feels like I’m working with ProRes media almost.

    What I don’t think you’ll get is any of the Kona 3’s 2K special features in 25 fps. You can output a cropped 1920×1080 frame from the 2K plate, which most of the times is good enough, but you don’t get the crop / scale features I don’t think.

    What also surprised me a bit is that Blackmagic does offer 2K support for 25 fps media, given I can’t say what they actually support in this regard. They haven’t been known for superb scalers in the past.

    ————————
    Erik Lindahl
    Freecloud Post Production Services
    http://www.freecloud.se

  • Erik Lindahl

    November 22, 2009 at 12:56 pm in reply to: More infor on McSmoke

    Awesome, thanks!

    ————————
    Erik Lindahl
    Freecloud Post Production Services
    http://www.freecloud.se

  • Erik Lindahl

    November 22, 2009 at 12:31 pm in reply to: More infor on McSmoke

    Link doesn’t work, boo 🙁

    ————————
    Erik Lindahl
    Freecloud Post Production Services
    http://www.freecloud.se

  • Erik Lindahl

    November 18, 2009 at 11:41 am in reply to: Multiprocessing Settings

    I’d say it’s hard or impossible to say “setting A or B is the best and most optimal” since this is very dependent on the scene and how well AE internally can spread it across multiple cores with one instance of AE. I guess if you had infinate RAM, more cores would always be better with this setting. But, realtiy is most of us are sitting with 8 or 16GB of RAM, hence we’re bound around 1-2 GB per core on an 8-core machine.

    I just did some tests to show the above. This is a relativly simple scene where we’ve replaced a filmed paper-card with graphics. The scene is very standard in this type of regard and uses amongst other things:

    – Pre Comps
    – EPS footage (using AE’s rasterizer)
    – QuickTime Animation Footage (SD PAL)
    – Uncompressed Audio (AIFF-file)
    – Text layers in AE
    – Filters in AE (Fast Blur, Directional Blur, Noise, Reel Smart Motion Blur, Corner Pin)
    – 8-bit project

    MY FINDINGS

    Mutliprocessing OFF
    Render time: 57 sec
    CPU usage: 350%

    Multiprocessing ON
    Use 8 cores, min 0.5GB / core, medium setting Longer vs Faster Rendering
    Render time: 31 sec
    CPU usage: 800%

    Multiprocessing ON
    Use 4 cores, min 0.5GB / core, Faster Rendering
    Render time: 31 sec
    CPU usage: 800%

    The above was done on a 8-core, 2.8 Ghz MacPro with 16GB of RAM with an externa 8-drive RAID 5, but it shows a rule I reckon one can follow:

    Render a scene with no multi-processing on. See how much CPU is used and figure out how many more instances will be needed to maximize your system.

    100% useage would need 8 instances
    200% useage would need 4 instances
    400% useage would need 2 instances

    The “mulitprocessing” setting is a bit vage since it’s not correct in saying “actual CPU’s that will be used”. What it is actually saying is “auditional instances of AE that will be used”. The degree one instance of AE can tax a system varies a lot as does the RAM requirements. In the end the wins of using multi-instance rendering will also vary a lot, in some cases you might be i/o or RAM limited. I would however say the above is the only way to “optimize” your settings and that will then be on a per project basis.

    If you want a general “max my machine” setting I’d go with something in lines with 4 cores and up the RAM a bit. I would also say this is an area where Adobe could improve AE a lot.

    ————————
    Erik Lindahl
    Freecloud Post Production Services
    http://www.freecloud.se

  • Erik Lindahl

    November 18, 2009 at 10:30 am in reply to: Media Manager Failure

    First off, big thanks for all the response.

    Second off, yes the people who started up this project should seriously reconsider how they work in FCP. Even disregarding the slip-up of NOT adding reel-names to a lot of the media, I do not understand how they could ever work efficiently in this project… It was spread out over multiple, removable FW drives, in about 4 different codecs and a mixtures of file and tape-based media. So yeah. getting the online in my lap I thought I’d best sort this mess out before I start with that…

    One of the main issues it seems was the fact a few of the clips DID NOT have a reel-name associated to them. This is human error in regards to the HDV-tapes that didn’t have this. However, all the “digital created media” won’t have this unless you manually sort it out. I still reckon this is something either FCP or the MM should alert me about or even be able to sort out. I don’t understand why it can’t reduce the amount of media just cause this single line of info isn’t available. Yes, it’s probably trying to be clever and see TC X to Y from reel A vs B and so forth but then you should have a different option in the MM to for instance “reduce media on a per clip basis” also or something. I mean, if I work with animation / pure digital created content, I won’t have reels either and the TC will likely be from 00:00:00:00 in every clip I import.

    Now, finally, I’ve at least reduced the media amount from 4 hours to 36 minutes and it’s currently being transcoded to ProRes. It will be interesting to see how that goes to say the least…

    Valued lesson taut though! All the “normal” things one does in FCP is done for a reason and people not understanding this shouldn’t be allowed near the production work flow. I am however a bit surprised at the short-comings of FCP / MM or perhaps it’s stubbornness / unhelpfulness. I presume if one works with a lot of non-tape created media has to manually remember to add a reel to these files? The given project has quite a few animations in it, and I just assorted them reel “file 1”, “file 2” etc which seems stupid but did cut down the amount of media needed to be transcoded by quite a lot.

    ————————
    Erik Lindahl
    Freecloud Post Production Services
    http://www.freecloud.se

  • Erik Lindahl

    November 16, 2009 at 5:11 pm in reply to: Media Manager Failure

    First problem is this project was spread out over 3-4 disks with about 10 “untitled” media files so to be honest I’m not 100% sure where everything is or I would have copied in from the finder. Yes, whom ever off-lined and esp.

    Still, and more importantly perhaps, I’m still amazed FCP can’t consolidate the media of the given sequence. I had to copy 2 hours of footage where my project might be using 20 minutes if I exaggerate. I don’t get why FCP can’t handle this after almost 10 years as an editor�

    The problem lays in the fact if you capture 1 hour of media, edit 10s of this file and then want to media-manage these 10s, FCP will think “oh, okey, I can sort that for you!” and it will want to copy the full 1 hour of original media. This is a huge irritating limitation of how the “media manager” works (or doesn’t�)

    Is FCP 7 any better in this regard or is there a solution for this in FCP6? If I exported an EDL and reimported everything from tape it work fine, but the media manager can’t manage what I’d expect anyone would want to do when moving from offline to online, esp. now that we work with a lot of file-based formats.

    I guess I’m just lucky I don’t work on long-form offline / online workflows on a daily basis, or have I fundamentally missed something here?

    ————————
    Erik Lindahl
    Freecloud Post Production Services
    http://www.freecloud.se

Page 21 of 51

We use anonymous cookies to give you the best experience we can.
Our Privacy policy | GDPR Policy