Personal View site logo
GH4 4K Panasonic video camera, official topic
  • 3230 Replies sorted by
  • @caveport

    Oh, that I can see. Which is why I've looked at some of the negative comments here and thought "hey, like, Dude! Really! ... get a life!" ;-)

    It's not only going to be a major upgrade in HD video, being able to shoot 4k has actually a big upside for those of us who do both stills & vid of the same subjects/clients: we can shoot something in 4k, grab any frame for a "still" shot, and then transcode the thing down for video processing if we choose and get the best of both worlds.

    Requires one to shoot very righteous in-cam WB and exposure so you can do stills (at the level we expect here) without shooting raw, but certainly can be done. Oh, yea, there's things I could do with that cam ...

  • @rNeil I don't know how you edit and which software you use but it helps to use proxies for editing. I have 4-5 HD tracks to edit and do so on a duo core iMac. using original or optimized media it's really sluggish but I can go a long way with proxies. I don't think I could add another stream of 4K though :-)

  • @rNeil Hard drives and storage is an issue big time doing documentary work.

    On our last feature, I had two arrays with 3TB drives in them. I thought that I could run a mirrored array and have enough space, but unfortunately, I discovered that we had about 8TB of optimized footage (and we archived to another array the cards from the cameras). Therefore, I put the FCPX project on one mirrored array and the other was spanned to 6TBs. I was editing at times without an immediate backup (I'm not sure how having the project mirrored would have helped me, I only thought it did).

    So, one of the advantages to AVCHD was that it produced smaller files. But once optimized they were quite large. With this new 4K, I can only imagine the file sizes.

  • depending on your computer you might not need to optimize and can always use proxies, even 4k doesn't necessarily need optimisation depending on codec

  • @DailyFilmFix: What do you mean by "[AVCHD files] ... once optimized they were quite large"?

    The video editing software I use (probably like most others) just keeps references to the original material and will apply any optimizations/effects either on the fly in the preview display or while rendering the final output. At no time, there is a need to keep files larger than the original AVCHD files on permanent storage devices.

    Thus I'm not at all concerned by 100MBit/s recordings from a GH4 - they will be the same size as the 100MBit/s recordings from the hacked GH2. 4k will certainly take more CPU cycles to process, though, but that is a different story.

  • @killagram

    The 7Q won't take 4K through HDMI, but someone may make a cheap 4K HDMI to SDI converter box. The YAGH is probably the best option right now, though.

  • Will we see a AF200 of sort at NAB 2014?

  • What a chunk!

    image.jpg
    640 x 359 - 62K
  • One question. The body maybe is bit different to GH3, but weight and chassis is very, very closed, maybe chassis is te same, and batteries too are the same of GH3. And at least in prototypes it doesn´t have a remarkable higher cosumption. Nice all gear from GH3 will be able to this camera.

    In the interview with Eduardo Angel he explains with a 64 Gb card it has 29-30 minutes of 4k recording, and each official card will cost... about 150$ each. Good optical gear is indeed necessary in 4K.

  • @karl I suppose that editing natively is an option. With the GH3, the .mov can be edited nicely natively without trancoding.

    But AVCHD is notorious for causing problems in post when attempting to edit natively. Before quad core machines, AVCHD could not be edited natively at all really. For example, my old G5 tower could not edit the footage unless I transcoded the files to Pro Res of AIC using an Intel machine.

    I'm no expert, I suppose, but I don't know anyone who edits AVCHD natively, even with a monster machine. I will ask my wife what she does in Premiere, but her Canon cameras do not use the AVCHD codec.

    Consider the workflow I employ: 1. Archive at the end of the shoot (on location, I use a MacBook Pro and FCPX); 2. Bring the footage into FCPX by importing (I create optimized media, with my old i3, I used to create proxy files as well, which are in the nature of Pro Res LT); 3. Sync the footage (now with multicam instead of Plural Eyes); 4. Edit.

    My reason for creating Pro Res transcodes at the beginning is to have clean files that are all the same and ready to go. I want no issues once I start editing. It is time well spent. I do move the archives to a separate drive to save space on the array, but having everything in the same codec is ideal.

    What NLE are you using? After some problems, the new FCPX is pretty good.

  • I've just used the mov 1080 50m files from my GH3 mostly ... for a few things "dropped" to the 720 mp4 as that was still overkill. Used the AVCHD only a few times early on, and decided as there wasn't any quality help to that format for what I was doing vs the mov, and the mov was easier to move around that I didn't need to use AVCHD at all.

    Now ... I do know a number of folks that edit from cams that primarily produce AVCHD material (some of the pro level Panny's and all?) for commercial work. Who use PrPro and work natively without transcoding and without a problem, day in day out. One of them points out that his isn't a "monster machine" for processor/cores & etc. but is 1) very well balanced between the various subsystems, 2) his hard-drive setup is really hefty with multiple spanned RAIDS for speed and 3) his video outs are set to run four things: program controls monitor, program material monitor (the playback screen for what he's editing), a professional scope, and a pro broadcast monitor. The latter two set up so that "Window's doesn't know they're there and doesn't mess with them", all heavily calibrated to bars.

    I still think that's a rather major editing set-up, but compared to other pros he knows, he thinks his is kinda so-so. I would love to have it ...

  • @JFKa

    In the interview with Eduardo Angel he explains with a 64 Gb card it has 29-30 minutes of 4k recording, and each official card will cost... about 150$ each. Good optical gear is indeed necessary in 4K.

    The 64 GB cards are already down to $105 on Amazon. Using faster cards than this probably will not have any benefit. The camera is designed to work at no more than the minimum speed of these cards. Getting a faster card than that won't change the maximums that have been imposed on the camera.

    Faster cards had benefits when cameras were hacked because we were basically removing that max write speed limit. Since the GH3 hasn't even been hacked yet I don't anticipate(or desire) a hack for the GH4 anytime soon. I suspect that once several cards with the right specs are out on the market the SD card price will not be a significant deterrent.

    http://www.amazon.com/dp/B00I3BQJNA/ref=wl_it_dp_o_pd_S_ttl?_encoding=UTF8&colid=3GAGRWZRDL55J&coliid=I32GI30ORADTZ8

  • @DailyFilmFix: I never considered editing the AVCHD material problematic - even a very old 2-core Athlon X2 machine of mine was fast enough to decode AVCHD somewhat faster than realtime, so the only times it made a subtle difference in speed to use one less complex compressed formats was when "rewinding" while viewing, then you could notice a small delay from the software having to decode a few frames before reaching the one frame you wanted to "rewind to". Even back then, I considered transcoding not worth the effort (and that is even though the NLE "kdenlive" that I use offers to create/use proxy clips comfortably).

    Today, I mostly edit on a core i7 CPU, and decoding speed is not any issue of concern there.

    Out of curiousity, I just imported the Panasonic "Lights of the Yucatan" 4k-promo-video into a test-kdenlive-project, and just as expected, working with that as a clip was not a problem, including when scaling or using only a crop from the original for a 1080p output rendering.

    There are certainly codecs that are not implemented as optimized as of today, e.g. this sample 4k video encoded in VP9 is not one that I would like to work with other than after creating proxy-clips.

    The one thing that I am still concerned with is that I currently watch the material on a 1080p monitor when editing, and there is not 4k monitor that I would like to buy at the moment, and especially no 4k-display notebook that I could take with me.

  • I currently edit in AVCHD in PremPro without problems. I have a 27 iMac which I bought in July.

    Btw, an ExTele 4k crop mode on the GH4 would be just AMAZING...

  • @flablo, The GH4's 4k and 3.8k UHD modes are Extra Tele Conversion crop mode, in effect. They take pixels from the sensor one-to-one with no scaling. Refer to the first post of this topic for crop factors.

  • I have edited native AVCHD in premiere since 2012 (on a 21.5 inch imac i5 - 2012). It CAN be problematic for post work if the client / editor has not cleaned up his timeline properly. Often something goes wrong with the xml / file linking.. For instance I´ve had problems with wrong clips in the timeline, which can be very hard to spot - but that exclusively happens when the export comes from multiple tracks (often multiple clips on top of each other). A clean edit is a must - only individual clips can be used (multi-cam-edits need to be broken down into their original clips). If you have a clean edit, not only will you much more easily spot any problem - you are very unlikely to have a problem.

    When there´s risk for problems and no budget / no time I ask for a prores 444 output of the edit which I re-cut and import into AE. I do not see any reason to transcode files before editing unless you work in ancient software or unless you have a corrupt file that needs saving or some file format that doesn´t work.

  • @balazer I see. Didn't get they are one-to-one, as 4096 x 2160 correspond to 8.8 Megapixels, while the sensor is 16 Megapixels. So the 1.2x crop (referred to micro43 sensor, not FF) applied to 16MP results in 8.8MP? Seems strange to me. Not arguing, just asking :)

  • @DailyFilmFx re: "I don't know anyone who edits AVCHD natively." Everyone I know edits AVCHD natively. You take the footage from the Cam, drop it on the timeline in PP, and if your $50 Nvidia card is installed, you edit natively and apply effects, all in real time, and render the final project not only quickly but using superscaling for pan, zoom and crop. It is not only an enormous time saver, it is an enormous space saver--if you are running a dozen cameras, you can fill up a 4TB drive if you resample to an intermediary codec. Editing natively saves money on hard drives. The more cameras you use in your array, the more space, time and money you save, so the benefit scales to the size of the project.

    This saving of time and money will increase exponentially with 4K, even with the presumed introduction of H265.

  • Didn't get they are one-to-one, as 4096 x 2160 correspond to 8.8 Megapixels, while the sensor is 16 Megapixels. So the 1.2x crop (referred to micro43 sensor, not FF) applied to 16MP results in 8.8MP?

    Crop calculation is at http://www.personal-view.com/talks/discussion/9542/gh4-4k-panasonic-video-camera-official-topic/p1 , as for 16Mp, check ratio at this resolution :-) all numbers match.

  • Maybe I'm too careful, but having everything in the same Pro Res codec makes sense to me.

    This allows me to sync and pre-grade with no concerns. My machine is about a year or so old, a 3.1 Ghz quad i5 27 inch iMac with 1TB of VRam. It seems to work better with everything transcoded (even the .mov footage from the GH3, which is clearly overkill).

    I have never tried to edit AVCHD natively, and I have never spoken with an editor here that has. But most of the folks I know, shoot Canon and DO edit that natively (like my wife). I will ask around.

    Now, the real issue here is that most people on this forum (like the world) seem to use Premiere Pro. My wife upgraded from CS5 to CC and loves it. I stayed with FCP, and it is working well for me now. I think that if I don't transcode everything, I will have problems in FCPX. Also, if I ever figure out how to use Resolve, my feeling is having a multicam project in the same codec is probably a good idea.

    Finally, when we all leave the AVCHD world for the mov or All-I, transcoding will not be necessary. If I were shooting with two GH3s, I would not transcode. But because I'm mixing the G6 with the GH3, I feel like I need to transcode to Pro Res. I have two video streams and a third two channel audio stream going when I sync in multicam, and this can bog a computer down.

    I wonder if the GH4 in 4K will play nice with the GH3 1080 in FCPX multicam? Think about it. If your final output is 1080 and you have a wide (say 14mm) in 4K and a closeup (75mm or 45mm) in 1080 (GH3), you would have lots of options with the 4K footage in that environment. Very cool...

  • On while you have to see 30min limit:

    In EU VAT (about 20%) and expensive "marketing initiatives" (read - stealing money without much result) already add quite a lot. But camcorder tax is around 14% on top.

  • @rNeil

    I think of shooting 4k or 2000mb data streams, and go ... oh baby, I need to buy my own hard-drive factory first ... :-)

    Unless you go for H.265 :-(

  • One of the things I would be most interested in regarding the GH4 features: Will there be a useable "continous auto focus" while recording video, possibly "thanks to the 'Depth from Defocus' feature?

    Did any source mention anything about this, yet?

    Having no useable continous autofocus (as in: "clips being destroyed by sudden episodes of 'hunting desparately for focus'") is one of my biggest issues with the GH2 when shooting larger animals under water. On these occasions, I was missing my much older Sanyo camera which despite its "obsoleteness" was quite good at keeping things in focus while recording.

  • @DailyFilmFix

    I edit all my GH2 footage as AVCHD, not transcoding a thing - I've never seen a need to spend time converting to another format. Playback isn't an issue (Windows, Core2 Quad, Premiere CC, nVidia card).