Jump to content

Andrew Reid

Administrators
  • Posts

    15,422
  • Joined

  • Last visited

Everything posted by Andrew Reid

  1. At the moment heat is not the bottleneck. Let's try to stay on topic please.
  2. I agree eBay is now heavily biased towards the buyer, some people I've dealt on eBay treat me like a shop, one day I will have to don a badge and uniform and deliver the camera in person with a cheesy smile. It isn't much fun working for eBay!
  3. Try sharpening this one, it is less compressed for the interwebs...   http://www.eoshd.com/uploads/d5200-sharpen-this.jpg
  4. You are trying to sharpen a heavily compressed JPEG downloaded from Wordpress.   I'll see if I can find the master MOV clips and upload. But I am quite busy on testing the 5D Mark III raw at the moment and finishing off the Sony RX1 review.
  5. Consider this - the hacked 5D's 2K raw image is better than the 1D C's 1080p.   We are being diddled.   The more that comes out the more I am sure of it.
  6. It's funny that these guys were laughing at the DSLR crowd a few months ago, kind of ribbing us about how easily people get excited about raw when what really matters is the script, etc. etc. etc. etc. etc. etc. etc.   And now the joke is on them! Remind me what the C300 shoots again... Erm is it 8bit mud!?
  7. No, that theory is discredited unfortunately.   If you convert compressed H.264 to 10bit ProRes, it is the same to grade as the master files. Transcoding gains no extra image quality. I once read a very good blog showing this but unfortunately I forget the source. Unless you're trying to fix a compatibility issue like 5DToRGB did with GH2 footage and Quicktime on a Mac, you're best off dropping the original native clips straight into the NLE.
  8. The CF cards are certainly cheaper than spending $8k on a 4K recorder for the C500. Well, it's less than that now thanks to the Odyssey 7Q but you get the point.   However, the fact remains that the 1D C does its image a huge disservice by offering zero changes over the 1D X for double the price, other than the image itself. Zero. Nothing. Put that image in the C300's body Canon please. Next time they may get it right.   And whilst they're at it, please sort out the low end and DSLR video. It is a bloody mess. More video features are being reversed engineered out of the 5D Mark III than Canon's own engineers added in the 3 years between the 5D Mark II and 5D Mark III. They could have 2K raw video, maybe also 3.5K MJPEG 1:1 crop in that camera at the drop of a hat. I have no idea what the hell is stopping them.
  9. Which frame are you referring too? Is it this one -   http://www.eoshd.com/uploads/d5200-sharpened.jpg   It surprised me too.   Just try downloading the full JPEG and sharpening it in Photoshop, you will get the same result that I did.   There's not much in it between the GH2 and D5200 but the GH2 does still have an edge for resolution and the D5200 is much better in low light with a cleaner codec and larger sensor.   By the way the D5200's sensor also ranks very highly for stills at DXOMark, basically the best APS-C camera out there for image quality. Toshiba really did do well.
  10. You can't uncompress compressed video, the data is lost forever.   Upscaling is a different thing. DVD players used to do it. Standard definition DVDs to 720p. Sony's 4K TV will do it with 1080p. I don't know of any plugin or workflow for Premiere, Resolve, Final Cut Pro, etc. that does it. I find that a glaring omission in the editing world!
  11. Canon can't take legal action if they tried because it isn't illegal, nor does it infringe copyright. Canon should take the approach of Panasonic. Welcome the project, then build the features into the next model.   Just the fact that they have added DNG and 30fps raw bursts is incredible enough in itself.   To get raw video would be truly beyond anyone's wildest dreams, beyond even a Philip Bloom April fools.
  12. eBay fees are truly crazy. 10%! You are welcome to sell the GH3 here, good luck with the sale.
  13.   The Cinema EOS line would not suffer. Which pro with $$$$$ to spend needs a hack? They need reliability. They need ergonomics. They need standardisation. Image quality is important but only one factor. Canon are wrong if they think DSLR video will cannibalise their C300 and C500. Maybe the 1D C would suffer... But I can't see someone like Philip Bloom selling his C300 and going back to a DSLR + Magic Lantern just for image quality.   For the creative bandits like us who just want world beating image quality and are willing to be brave, the hack is perfect.
  14. I don't see how they can intervene other than encrypt firmware on new bodies leaving the factory and even then they shouldn't, because if this raw video hack takes off sales of 5Ds will go THROUGH THE ROOF.   Indeed what is the point of the 1D C I have asked myself that many times!!
  15. The image is processed before it gets to the HDMI port so unlikely to get the raw feed off the camera that way without dramatic changes to the way HDMI works.   I think it is more likely we'll see a crop mode of some kind first, at lower resolutions, written as DNG frames to a fast CF card. But I'm speculating there. We'll just have to see what Magic Lantern comes up with. Exciting times indeed.
  16. Wild Ranger - that's the million dollar question. Nobody knows yet.   Here's a 3.5k sample frame I just shot in silent stills burst mode at 30fps...   http://www.eoshd.com/uploads/ml/56420194.DNG.zip   You should be able to open it with Photoshop if your Adobe camera raw is up to date.   Image quality is stunning, no signs of moire. I think it looks like a 1:1 crop of the sensor. It isn't the full frame, more like a 1.3x crop horizontally and a much more severe crop vertically.
  17. I could do, it works just like the Nikon V1 burst mode.   Just need to transfer Magic Lantern to one of my faster 1000x cards as only getting 15 frames per burst on the 600x card.   The End Trigger mode is interesting - this appears to capture endlessly until you let go of the shutter release, then it writes the last few frames to the card.   BTW - with getting 3592 x 1320 raw burst with a certain combination of settings - is this documented normal behaviour?
  18. Download sample 2K DNG file and see the image quality for your own eyes! Recently Magic Lantern had the breakthrough discovery of beautiful pin sharp 2K sensor feed in raw format on the 5D Mark II and 5D Mark III. This completely turns our understanding of DSLR video image quality on its head. Canon claimed the 5D Mark II line skipped on the sensor to produce video. I don't understand how that can possibly be the case. [url=http://www.eoshd.com/content/10250/canon-5d-2k-raw-feed-update-1920x720-possible-on-1000x-card]Read the full article here[/url]
  19. Ah yes Joel, the guy who has been sending me abusive emails for the past 3 weeks.
  20.   If one fell in my lap I'd be more willing to put up with the sheer travesty of $12,000 as well.
  21.   The CF card I used in the 1D C does 150MB/s so not far off, but the write speeds are probably a lot lower than that.   Alex of Magic Lantern says he has got burst mode working - it can write 30 frames in 1 second at whatever the live-view FPS is (24-50fps) but flushing the buffer to the card takes half a minute for those 30 frames!   So yes it is a bit like the Nikon V1 burst mode.   In continuous mode the FPS is much lower. I've heard various reports between 4fps and 14fps.   For video I think the 2K would need to be cropped significantly (720p or pseudo 2x anamorphic 2048x540) or compressed. The internal processor might not be able to handle the data rate at 24fps even if the sensor can output that fast.   In Alex's latest update he seems quite pessimistic on the chances of raw video:   Don't hold your breath for raw video. 24fps at 1080p is unlikely IMO, even on 1000x cards. - HDMI out doesn't help. Here's why: raw image buffer vs recorded with hyperdeck shuttle uncompressed. Ignore the overlays. - In continuous mode I've got 4fps of raw data (not DNG, just the raw image buffer) on a 266x card. G3gg0 got 12.5fps of YUV422 video on the Lexar 1000x card. - I was able to record around 30 frames in burst mode, at LiveView FPS (24, 25, 30, 50, whatever). It takes roughly 1 second of "recording" to RAM and then half a minute of saving the DNGs to card.
  22.   I don't know. I've lost interest in it as you may as well just use the internal IPB codec and save the hassle.
  23.   So it is the camera not the recorder. Good to know.
  24.   It isn't about 'Canon bashing' at all, it is about making an informed decision on a $12,000 camera.   If you don't like the bad news, using that straw man argument of being creative with a camera is as relevant as talking about camera technical capabilities on an actor's forum.
×
×
  • Create New...