Jump to content

Andrew - EOSHD

Administrators
  • Posts

    15,720
  • Joined

  • Last visited

Posts posted by Andrew - EOSHD

  1. There's definitely a place for low-fi image quality.

     

    Kendy's video shows why hiding stuff is just as important as what you show. The 550D / T2i was a perfect choice for what he wanted to do. I have used that Sigma 30mm F1.4 and it also has character. Nice lens. Not too perfect, not too clinical.

     

    Says Kendy...

     

    The whole film was done over 3 days, post production included. I used my small Canon 550D, everything is natural, no additional lights. I think it is much more atmospheric. I don’t like it when you can see everything clearly on screen because it looks fake. I like the idea of showing just the shapes or the shadows of the characters like in a comic book. I wanted something organic so I added some clean 35mm grains to my footage, because I like the visual render of the old cameras such as the Bolex. For the strange flares effects, I use a broken glass in front of my lens and shot with only one hand. It’s easy with a small camera like the 550d, but could be a pain with an Alexa because of the weight. Because I make everything with After Effects (including the edit) I work very fast. I don’t waste my time by switching between several softwares or video tools.”

     

    Talented guy and without doubt the 99k plays on Behind The Move are there because of his sensitive use of the camera aesthetics, the sensitive handling of music and content, and the content itself including that amazing girl, fantastic dancer, great story. It all matters. ALL of it.

     

    In the case of Behind The Move the low-fi visuals suited the content but what about when they don't? What if you want to film a rich sunset in memory colour and see every blade of grass in a field? The 550D works best with the shallow DOF and medium close-ups seen in Behind The Move because it hides a lot of the flaws, like a lack of detail, moire and aliasing, compression, muted colour, etc. There will be a LOT of content that these issues don't suit and it helps to know about it before getting all inspired over someone else's content and picking a camera that might be completely ill-suited to your own.

     

    Also shooting style has a lot to do with it. Here the 550D is on a relatively slow-shutter speed and all handheld, with some great framing going on. Kendy's camera work here is like a high wire act. If you don't always get the framing emotionally connected with the mood and the subject matter, the slight of hand will be revealed the magic vanishes. If your shooting style is locked down, on a tripod, the 550D is the last thing you'd want to use, believe me. If your shooting style is more dynamic and flowing, then you'd probably be better off investing in drones and rigs rather than cameras anyway. Everyone is different.

     

    I want epic scenery, anamorphic, striking visuals, super rich colours and rich details. Sometimes I may want handheld camera-work, even black & white, a grungy low contrast low resolution feel. Why not start off with as high a quality image as you can get your hands on and rough it up? The glass in front of the lens in some shots in Kendy's movie is an example of that. Making stuff look less clean, less 'real', more magical and otherworldly.

     

    The technology needs to keep improving and the shooter needs to stay on top of it.

     

    I am already imagining ways where Kendy's video could have been improved in terms of lenses and cameras!

     

    It doesn't need to be improved though because the muted, low-fi look is perfect for his subject matter... the isolation, loneliness and overwork of the dancer, being stretched this way and that in a big city away from home.

     

    If the film was longer and we followed her back home and there was a loving embrace with her mum or something... maybe you'd not want the same look for that scene.

  2. I downloaded Luke's raw files for that test. It was a good test but not a definitive final one as I am sure he will be first to admit. I am sure he has more planned. The stuff I found...First the GH4 has digital sharpening applied because it is set to 0 in-camera, the middle of the sharpness scale rather than -5 and that separates it from the Epic and 5D because they both have no sharpening applied in post on the raw data. None of that changes Luke's conclusion that the Epic is softer though, he's right. It is. The GH4 seems a little overexposed in the test which lends a harsher look to highlights in the image, like glinting specular highlights on the roof and the gradation in the sky suffers a bit because of it. The 5D Mark III looks overexposed to me and boosted in post to ISO 800 to match the native ISO on another camera. It should have been shot at ISO 200 with no exposure boost in post necessary on the raw file. Exposure to the right may have helped it. The Epic looks filmic to me, but I am surprised how little extra detail it is giving compared to upscaled 1080p to 4K from the 5D Mark III!

  3. Wow, this thread is fired up.  I really think the GH4 looks good, maybe surprisingly so, but then so does the GH3. Both cameras can hang with the 5D3R raw if you put them in the right setting and use them in the right way.  I see the GH4 as a nice improvement over the last generation, but it's lack of raw support means it's not for me. 

     

    Resolve 10 actually now does a better job than Adobe at debayering DNG. Resolve 9 was ugly and had no noise reduction capability at all. You need at least a very light colour noise reduction going on, which ACR applies by default. When you convert the 14bit raw files to 10bit ProRes, you will not notice a drop in image quality (as long as you grade in 14bit before the conversion). 10bit is enough. Most of our screens are 8bit. 8bit as an acquisition format, now that's different... it isn't enough for everyone, which is why the GH4 has the 10bit 4K output over HDMI for the Atomos Shogun. Problem solved. Granted not yet... but by the time the A7S is out the GH4 will be shooting 10bit 4K compared to 8bit 4K from the Sony. A considerable advantage but not as noticeable to the viewer as the effect a well handled full frame sensor has.

     

    I'm planning a more in-depth test between full frame 5D3 raw and 4K GH4. First the 2.2x crop needs to be tested more, without Speed Booster. Secondly more variety of subject matter needs to be shot including people, but with a variety of camera positions and focal lengths. It will be interesting to see if full frame loses it's edge when we go to a longer focal length and tighter framing, because here the smaller sensor is actually an advantage.

  4. It's a variable bitrate codec and it probably leaves 20Mbit in reserve for a particularly challenging few frames to avoid corruption.

     

    Try a sudden whip pan of a forest of trees at F5.6 or a sudden flare up of the entire frame whilst shooting something with tons of detail in it, that will really push the codec to the limit!

  5. The GH4's resolution is mind blowing there. Really impressive. I think it will be amazing for green screen work specifically. Cinematically, sharpness is not exactly a look I would always want to go for, but you can always use lenses or post to unsharpen or add more character. Vica versa is not possible (i.e.: adding detail where the resolution isn't there) so definitely not putting down the sharpness. Amazing detail out of the box, no hack.

     

    As for the "video" look people talk about... what video cameras are people comparing the GH look to? I think often its either 30p, bad lighting, stock lenses, using all auto settings and ultimately a basic point and shoot handheld style that give these cameras a "video" feel. The whole "video look" comments are quite ridiculous really.

     

    The article wraps up my sentiments perfectly. The single hesitation I have is that its not full frame. FF is not a gimmick - it allows for wider shots to be used with lenses performing at their intended focal view. Better low - light. Shallower depth of field. I like all of these features quite considerably... making my expectations very high at the upcoming Sony A7s.

     

    I think whoever is not really needing a new camera would be sensible to wait 6 months. Now that the GH4 seems to set the standard for sharpness and detail... who knows, maybe the next 4K beast will feature a global shutter. That would be a real winner.

     

    Whatever the case, at $1700 this is very very tempting. Lets just cross fingers for the A7s as being a good FF alternative. 

     

    Yes that's pretty much my single reservation as well, that it isn't full frame! A7S is definitely high up my list to try out.

     

    As for sharpness, the GH4 at -5 sharpness is actually quite soft in 4K at 1:1... when the image is down sampled to 2K certain downsampling methods can make it look a bit too sharp. Simply add a slight Gaussian blur in post. It will be interesting to use my LOMO anamorphic on it wide open too.

  6. What do you think of this gh4 versus a blackmagic pocket? 

    I have the impression that blackmagic has a more cinematic look but maybe it's something you can have in post with the gh4 or it's always have a more "video" look more suitable for documentaries than fiction?

     

    The Blackmagic is practically designed from the ground up to deliver a cinematic look out of the box, whereas with the GH4 it is more at the discretion of the user. If you mishandle it, then it is easy to end up with an image which is too sharp, too flat and too lacking in character. If handled correctly the GH4 is as cinematic as just about anything else out there if not more so.

  7. If one had to make a film out of that footage I would pick the GH4.  Beautiful crispness!  However, I don't feel any of that footage showed the low-contrast look possible with the 5D3 RAW, and not the GH4.   In the scenes where there was wide dynamic range, like when you shot under the arches, or into shadows, it seemed both cameras were similar.  The GH4 did look more contrasty, but not by much. 

     

    I believe, if you shot indoors the difference between the cameras would be more pronounced.  Outside, the un-modified brightness just overwhelms both cameras.

     

    Thoughts, Andrew?

     

    A low contrast look is not superior. You can make the GH4 and 5D raw look as flat as a pancake if you want to. Question is what is the benefit? If your monitor had a low contrast look you'd return it to the shop! I don't know why people think a flat look is superior. For grading raw has no look. It isn't flat. It is raw data straight off the sensor that describes colours, whites and blacks. Why compress all that into the mids and get grey? I blame stuff like CineStyle for the confusion.

  8. Funny how certain people elicit oohs and aahs about sharpness and then put 40 or 50 year old (sometimes uncoated) lenses on these cameras. 

     

    Nobody wants to see ultra sharp footage of people's faces, especially when the skin tones are produced by a GH series camera. 

     

    There's nothing wrong with the skin tones on the GH4 at all. They benefit from 10bit luma at 2K for a start. 10bit everything at 4K if you are going to use a Shogun. The colour reproduction is good and skin tones grade well especially with a boost to the red channel in post.

  9. For the neumanfilms comparison the Epic and 5d3 looked out of focus.  He says in the comments he was focused as good as possible, but he doesn't say on what object so I don't believe they were.  So that particular video is useless as far as a comparison.

     

    As for Andrew's, the shots really do match together quite well. Obviously more detail in the GH4 so I think it would be superior for documentaries.  There are a few shots where the 5d3 seemed to have a little bit more dynamic range, but not by much at all.  I guess the best person to make the comparison is the person grading the footage.  Which camera can get to the image you want the easiest and most reliably, that is the only way do decide because in this test the images are equal.

     

    It was a good test in my opinion. It clearly shows that the Epic has a stronger OLPF and a softer result ungraded. Ungraded raw is very different to GH4 4K where a certain level of sharpening is already applied in-camera even with sharpness all the way down at -5.

     

    Yes the Epic looks surprisingly soft, almost as soft as upscaled 1080p from the 5D Mark III raw footage.

     

    But with a little sharpening in post it would have been much closer.

     

    Am I the only one surprised at how well the 5D Mark III blow-ups look under the microscope at 4K? I expected there to be a bigger difference. My opinion is that in terms purely of detail, 4K vs 1080p depends on how you're going to present / show the end footage to the audience and even on the eyesight of the audience, but clearly taking a full pixel readout of the sensor in the GH4 for 4K has resulted in overall image quality leaping forwards, not just in terms of resolution. Said this all along.

  10. This isn't a centre piece for Panasonic's promotion of the GH4, it is just what Nick Driftwood wants you to believe. Because he presumably thinks there's some kind of career reward in being nice to big companies. The thing is there really isn't. Panasonic just want to sell cameras and make money, they're not really in the best position to be picking sensitive artists to give demo units to. I just walked past a DPReview gold badge on a GH3 box today, just randomly on a store shelf. I realised if it wasn't for me writing the GH3 video page in that review and putting my argument across to the editor that it deserved a gold reward because of the video capabilities (DPR felt it merited a silver, because as a stills camera it lacked any real outstanding spark) that badge wouldn't exist. I felt a bit sad that in some ways my review and my views were being commercialised to sell cameras, without any reward for me, and nor should there be as it would cause bias. That's why I prefer to get minimally involved with stunts like this and marketing. If I make a short film with the GH4 it will be because the camera suits the material, not because I want to score attention with Panasonic, because like I say, there's no real reward for doing so and it is not really what I want. I want attention from other filmmakers and the audience instead. You have to question Nick's motives for making this to be honest and trying to pass it off as a wannabe Panasonic promo video.

  11. It all depends on what you want to do.

     

    For some stuff I prefer the 5D Mark III and certainly raw video on it is mind-blowing for work involving rich colour saturation, heavy grading and low light. I also love the full frame rendering of my lenses.

     

    The GH4 in terms of 4K and usability of course is a big win, but the smaller sensor and the stills are not up to 5D Mark III level. Still pretty good though... how much stills quality do you NEED? That's the question to ask, not so much whether the GH4 is rated that way or this way.

  12. I feel I have the knowledge and temperament, though I'm not always around work depending, but I'd be happy to assist :)

     

    I agree. You are now upgraded to mod status sir  :)

     

    Now then... if all the mods fall out... who will mod the mods? ;)

  13. http://nofilmschool.com/2014/05/first-short-film-shot-on-4k-panasonic-gh4/

    Check this out & read the comments. It seems it's common knowledge now that the GH4s image is very video-y and incapable of producing filmic images. I am stunned by the amount of comments saying this. I know it's pure crap and in the right hands it can produce a filmic image, maybe harder than say a Blackmagic but still do-able.

     

    This was shot on film...

     

     

    Therefore I have decided film is incapable of producing a filmic image!

     

    The first comment on nofilmschool kind of sums it up

     

    If I was the CEO of Panasonic, I would buy this film and immediately take it down. Then tell these “filmmakers” to stop using the GH4.

     

    Give a hack a camera, and a hack of a film comes out...

×
×
  • Create New...