Michael Ma

Amazing work!  You're quite the renaissance man.

If you don't mind me asking, if you had to do it all over again, what would you buy first for stabilization?  A weighted-stabilizer, brushless gimbal, jib, or slider tripod?  (Or something else?)

Doesn't the mirrorless GoPro Hero3 do 720p at 120fps? 

Yes it does.  Also, the Hero 3 Black Edition does 640 x 480 at 240fps, but I think the 240fps has a very interlaced looked to it.

The industry is definitely at the brink of figuring itself out.  We're buying cameras with grips suited for photography, but to shoot video.  These companies don't necessarily want to overshadow their dedicated video line of products, which usually turn out to be less capable, but more expensive (at least when looking at the value/dollar).  People are not sold on dedicated video form factors like the 4K cameras that Canon just rolled out.  But people are comfortable buying a SLR-type cameras that shoot good video.  

Even though Canon seems to think the logical upgrade is a dedicated video camera, I don't think a dedicated video camera will ever be the next logical upgrade for a lot of users who plan to shoot a lot of video for the foreseeable future.  There is something great with being able to take video with a SLR form factor.  The compactness, the discreetness, and the familiarity of working with the type of gear and lenses we already own.

There are rogue companies like Black Magic that's playing to exactly the needs what people like me are looking for, but they lack the decades long experience with things like firmware, UI, human-device-interface interaction.  The lenses available always seem to come with a compromise.  

The future....Black Magic cameras can only become better, and the price seems right for the market that I find myself in (The prosumer market).  They can build the perfect camera as their products mature.

Canon on the other hand is in a position to offer this functionality at a flip of a switch it seems.  But they refuse to do it because they want to sell different lines of cameras on their own agenda at a price suited for a different market that I'm in.  I don't plan on getting hired to shoot a episode for a major TV network or a film for a major company, so I will never have the budget to spend $10k on a camera.  Canon has all the tech and systems in place to produce the perfect camera, which would be all the specs and price of a Black Magic camera, except with full electronic compatibility with their lenses, better focus system, better battery life, and better UI and controls.  

In a few short years I think this camera will definitely be here.  Whether it's by Black Magic Design continuing to improve, Panasonic by cranking out the most specs that can be had with consumer components inside the camera, or Canon just simply just choosing to assemble the camera and not put in old parts as in ways of controlling the hack-ability of their firmware.  

Black Magic Design appears to be young and brash.  Panasonic seems to be eager to please.  Canon, they just look greedy and condescending.  It sends the message that they are trying to milk us, think they know what's better for us, don't think we're smart enough to see the handicapping they are doing to their own products, or just don't care enough about the this niche market.  

These are exciting and frustrating times...

In 10 years we will probably have this tech in the Panasonic GH14 for about $1500.  I can wait.

Been playing with this all day.  I want to share another way to explain information which may have already been posted.  

4:2:0 to 4:4:4...it works, I tested it, it looks great when compared to other 4:4:4 vs 4:2:0 examples on the web.

As for 8-bit to 10-bit, and looking to gain any dynamic range from that by combining 4 x 8-bit data, it is not possible.

It's not that we're collecting actual light per pixel and combining light from 4 sources into 1.  We can't add these together as if we're adding light photons.  We only have the readings of 0-255 (8-bit) for each subpixel (R,G,B ) for each pixel for one ISO sensitivity.  It knows no information outside of that for shadow or highlight recovery out of those bounds.  Having this near identical data repeated 4 times won't help expand the dynamic range with any kind of formula or workflow.  You can't extrapolate detail from the value of 0.  Same thing for blacks.  When all the RGB values goes to 255, you can only extrapolate shades of gray and any color is lost.

All we can do is just average 4 pixels for each 1 pixel for better color accuracy.

If Panasonic had made the GH4 so that you can flip a switch and each 4 sensor groupings collected data at 4 different sensitivities, then we'd have the data to work with, but then 4:2:0 to 4:4:4 conversion won't be as effective or effective at all because we're purposely overexposing or underexposing sensors for some pixels that would normally be collecting useful information.  

With that said, you still should convert to 4:4:4 10-bit (not 4:4:4 8-bit if that's even an option).  That would give the color grading software more room to work with.  I learned a lot from this thread.  Now I want a 4:4:4 12-bit camera more than ever.   :)  Maybe in 1-2 years we'll have it on the GH5 or the GH6.

Hi, im new in the forum. I understand this can be do too with gopro hero 3 2,7k fotage, more accesible. I was trying do it today without results (in AE). Somebody know a specific way to do it?

I think that Cineform Studio free version (now called GoPro Studio 2.0) can do this.  In step 1, convert your file to CFHD (in the same resolution).  It is inflated to a 10-bit file.  Step 2, select file.  Step 3, export to custom settings, choose MOV container, scale down to 1080p, you can choose Film scan 2, which is the highest quality that can preserve the details of a 4:4:4 footage.

Update: Just tried this with my Galaxy Note 3 which shoots 4K 4:2:0 8-bit at 50Mbps AVC.  Looks a million times better than native 1080p.  I also tried dropping the same 4K file into After Effects CC working in 16bpc.  Created a sequence and exported to DNx 350X or 440X 4:4:4 10-bit...Looks different than Cineform (CFHD) workflow above.  Not sure which is better to be honest.

Increased dynamic range?  I would think yes and no.

Yes for when you are talking about less noise in the shadows, but not when it comes to any stops gained for recovery purposes.

Still fairly new to all this, but I'm guessing 4:2:0 8bit 4k -> 4:4:4 10bit 1080p, there's no gain in dynamic range, but at least big leap in sharpness?

I have a Galaxy Note 3 that does 4:2:0 8bit 4K at the moment.  Can I put that through GoPro Studio 2.0 and convert to Cineform Filmscan 2, and get this result?

Or can I do this in Premiere Pro CC by scaling down to 1080p inside a DNxHD 350X sequence?  Or does it have to be AE?  If someone can share a defailed workflow, that would be awesome.

I'm sure the GH5 will do 1080p at 10 bit 422 in body.  I can wait 1 year no problem.  I think the GH4 could have done it.  I just don't understand why it doesn't.

Literally no one on the face of the planet would ever want to work for you - you have some kind of anger issues. Also, did you just like, copy/paste the exact some post from earlier? I don't understand. You seem to be under the impression that throwing out insults at people makes you seem cool. It actually is the opposite. It makes you appear very small, and a little bit... y'know, insane. Be nice. Grown ups are supposed to behave better.  :)

Oh, he doesn't sound insane at all trying to reject people for jobs they never applied for.   ;)

deang001,

I decided to have some fun and try converting your chicken choppers from 30p video to 24p using the workflow below:

If you're interested here are the results:

You will want to download it to watch it.  not watch it in the browser which looks heavily compressed.

https://www.dropbox.com/s/8xjns3jan6vbkc0/cc24p.mp4

In After Effects CC

1. create new composite with original video and set Time -> Time Stretch -> Stretch Factor (250%)

2. Composition Settings -> Increase "Duration" by 2.5x (so 40.02 seconds became 1:40:07)

3. Enabled Frame Blending

4. Frame Blending -> Pixel Motion

5. Then export keeping framerate 29.97  What you end up with is a stretched clip by a factor of 2.5x at 29.97fps.  The frames are filled in with computer-generated frames.

After that take the stretched file into Premiere Pro

6. Create new sequence with video and set Speed/Duration -> Speed (250%)

7. Set opacity to 50%

8. Then underneath the 24p video layer, I put the original 29.97fps video.

9. Then export sequence to 23.97fps with the Frame Blending enabled

Added note: You can change the opacity from anywhere from 0 - 100% in step 7.  Sometimes the original footage doesn't have much movement so no conversion is necessary.  Sometimes the interpolated footage is perfect, so 100% is fine.  But 50% seems fine across the board most of the times.

It turns out After Effects has built in feature to interpolate frames just like Twixtor.  They licensed The Foundry's Kronos and called it Pixel motion.  You can probably do all this within After Effects although I like this workflow because frame blending seems to work better this way for me.

I think the best 30p to 24-25p conversion you can do is to use Twixtor to increase the framerate by 2.5x to 75p.  It computer generates an interpolated frame that's not there, and it does a remarkable job for the most part.  If you look at a single frame digital videos shot at 24p, you see 2 trailing frames.  The sound of computer generated frames might sound like something you might not want, but since you are blending it together with real frames, any artifacts are really not noticeable unless you are specifically pausing frames and trying to find them.

I've done this for footage shot in good lighting with sharp edges around subjects and objects, and it looks very convincing.  May not work as well in lower night or shallow DOF shots.  It might, but I've never tried.

I want to share my experience with GH3 as a photography tool.

Coming from a canon prosumer series, I was very disappointed with the GH3.  The midtones are flat, so you end up with pasty saturated complexions on people's faces.  At least in Adobe Camera Raw.

So I bit the bullet and bought PS Kiss' GH3 lens profiles which lets me force different camera profiles that are calibrated for the GH3.  And when I apply the 5D Mark 3 profile and bump the exposure by 1/3 stop, the image quality is superb.  Whereas the standard Adobe Camera Raw profile that which is your only option in Adobe Camera Raw, for my standards, makes unusable pictures when it comes to pictures with people's faces being the subject.

So the problem is that Panasonic is not working with Adobe to create calibrated lens profiles, not that the camera is not capable.  But you can work around it for now.  It's just one extra step in the workflow.  Sure the raw files in MFT cameras are supposed to magically work in Adobe camera raw, but they only fix distortion.  Colors, which may be fine for people looking for a different color profile than what you normally get from Nikon or Canon, but are not good when you are looking for accurate color.

With that said, I feel good about shooting photos with the GH3 now that I know how to get the most out of it.

I don't own the Olympus but after seeing some footage, it looks like the stabilization gets lost sometimes and has to readjust itself, which would make that part of footage unusable.  But when it's working well, which is most of the times, it looks amazing

Handheld video while walking is a challenge on the GH3.  

I've been playing with 720p@60p 72Mbit/sec, so warp stabilizer (on Premiere Pro CC) has crisper edges to work with, then converting that to 24p (frame blend in export only) with some descent results if you don't mind mixing 720p footage into your 1080p footage.

I think price is way off.  In 5 months time, 4K will be enabled in a lot of consumer electronics.  Sure they won't be as good as the GH4, but these consumer devices will be demonstrating to the consumer that only minor changes to the GH3 will enable it to do 4K.  Primarily the sensor, data bus, and the processor.  Some of these components would have been a given on the GH3 successor, 4K or not. Is that worth a $2,000 extra?  Definitely not in 5 months.

Who needs stills.  Wow.

After having done only photography for 7+ years, I just made my first purchase as well into the video realm with the GH3.

I considered the following:

Canon 6D...lots of camera features, but not too many people interested in it in the video forums.  Problems with moire really that bad?  My instincts said to pass.

Canon 7D...dynamic range is really unnatural and outdated.  Waited for 7D Mark II announcement for 6 months and it never came.  I had 6 EF-S lenses, but even with that, I had to pass.  

NEX-7...great video, but terrible and unexpected overheating.  It's pretty old, so a successor announcement was likely, but I grew tired of waiting after many months.

NEX-6...No external mic jack, and the video seemed like a downgrade from the NEX-7.

BMPCC...it started shipping 2 weeks later than the time I needed to make my purchase.  But having watched many reviews, although the video quality you can capture is superior to any other cameras listed here, I'm glad I didn't choose it.  As in terms of being all-in-one reliable workhorse, it is the opposite of that.  You need to find workarounds for the battery, sd card incompatibility, focusing, non-tilting LCD, audio, and find a rig that's gonna hold all that together.  And say I got all that, it would have become something I did not want to carry.

Nikon D7100...A few samples I found online looked great, but no one really talks about Nikon in the videography world.  Surprisingly not too many samples on YouTube.  Without further understanding, my instincts told me to pass.

5D Mark III, the dream camera for video.  It is better than the GH3 when it comes to moire, noise at high ISOs, also my dream camera for photography as well.  Sadly out of my budget.  

Hacked GH2...in good lighting, it supposedly even takes better video than the GH3.  It has less problems with moire.  But in harsh lighting, or trying to recover shadows, the GH3 is far superior.

GH3, which I ultimately bought, was always highly recommended.  The argument that finally convinced me was that people professionally can and do rely on it and it always gets the job done.  Having been the early adopter of many first generation products, and having put on hacked firmwares on probably a dozen devices, I realized that hacked solutions always come with some gotchas and I didn't want to start off on the on the wrong foot.

My experience with the GH3.  The GH3 is everything that people are saying.  You can capture cinematic looking videos right out of the box and have some room to adjust the hightlights and the shadows to fine tune your shots.  The battery seems to last all day.  My existing SD card works.  With that said, I had to find workarounds for problems with moire.  Depending on the application, you may never see it, or you may always see it.  As in terms of harshness of moire compared to other cameras, the moire on the GH3 isn't that bad.  Audio is sometimes usable as is, but not always depending on your environment.  A clip-on mic is my next purchase.

Lenses, a lot of people try to cover the spectrum of distances and apertures, and then maybe throw in a prime when they make an initial purchase, but I think it is a guaranteed way to end up with lenses you won't be using later.  So just start with one lens.  There are a ton of tricks to get any shot you want (with some minor compromise) if you have enough space to work with or in post.  And when you've developed your own style, you'll know exactly what lens you need to take your work to the next level.  And most likely it won't be any of the lenses you would have bought initially.

If they are using the footage to study, you probably want highest fps and resolution, which would be MOV 1920p 60p.  That's 50Mb/s which means the file sizes will be huge.  You will see a higher bitrate All-I 72Mb/s option, but that will be disadvantageous if you aren't post processing.  You'll probably need a 64GB Class 10 card with UHS-1 support.  If you smaller files, there are plenty of options for the MP4 or AVCHD container/codec.

Rule of thumb is to have shutter at double the frames per second.  So if you are shooting 25p, you should have the shutter set at 1/50s.  So for 60p, 1/120s should be the shutter.

Focus mode AFC is better than AFS because it needs continuous active focus.

Probably tripod is your best bet.  If you are doing handheld, unlike shooting photos where you are trying to square up and be immobile like a rock, you should have your elbows away from your body for stabilization while moving.  Try moving your entire body for tilting and panning rather than just your arms.

An image stabilized lens is a must.

If you were doing post processing, you should have set sharpness, contrast to -4 and -3, but since you don't want to post 0 would probably be fine.  You will probably find saturation at 0 is still over-saturated so start with -3 and adjust accordingly.

If I had to guess, the reason why it gives you only 30fps for video is so they can add some (undocumented) software image stabilized video.  Hence it is locked onto very limited options for video because of the extra processing.

Also, did no one notice the extreme amounts of flickering?

Interesting that they made that last shot very noisy on purpose.  Maybe you have to own a 5D3 to get the message.

I didn't really see anything that they probably couldn't pull off with the 5D3, unless the only source of lighting was the sun peeking through the windows.  

Yes, but note that if you film in dusk, the difference between car headlights and the rest of the scene is about 4 stops, and at night, it might be something less 8 stops. If you expose to the right, the rest of your scene will be completely underexposed. Also, forget filming concerts, for example, with their big exposure difference between stage lights and ambient light.

 

Blackmagic advertises the camera as "everything you need to bring cinematic film look shooting to the most difficult and remote locations, perfect for documentaries, independent films, photo journalism, music festivals, ENG, protest marches and even war zones." This is certainly not true for the camera as it works now.

The real bummer here is for early adopters who don't have much video editing experience (who've never put on trackers to remove the black dots, or know how to clean up audio in a usable way), or know how but never intended to do processing outside of premiere pro who took a gamble on BlackMagic delivering a bug free product.  Some people need to put out videos daily rather than spend hours trying to perfect a < 1 minute composition.  Cries for prompt firmware updates and recalls may not be loud enough because a big percentage of BlackMagic customers probably know how to circumvent these issues and still make use out of the BMPCC.

Maybe fixable with a firmware update in the future?

I ran this clip through Adobe Audition's noise reduction (after I captured footprint).  

Then you will see that all the unwanted extra noise is coming 5k and 6.3k Hz in the 30 band EQ.  Put that to zero....not for quality recordings but maybe fine for non-professional output.  When you aren't shooting on the fly, a preamp and a mic would go a long way.

https://dl.dropboxusercontent.com/u/64326938/blackmagic%20audio%20noise%20removed.mp3

 

I would still get the hacked GH2 or GH3 instead of this Pocket if the assignment was "small interchangeable lens camera." The Panasonics are sharper/higher res, have an adequate codec, have I would imagine much better lens support, and far better ergonomics and workflow. And cheaper in the case of the GH2. This BMD Pocket has better in-camera color/"look" and a better gamma. But it doesn't relieve you from having to expose the shot properly (undermining the supposed DR advantage) and is nearly as soft as 5D3 H.264 footage but suffering from moire regardless.

 

Thank you for your input.  I'm still trying to decide between a GH3 or a BMPCC.  How would the GH3's dynamic range do in the same situation against the BMPCC?  Would I get a lot more blown out highlights, white skies, black shadows?  Or would I get just about the same results?  I would love to just point this camera in any direction without thinking of harsh lighting or shadows in respect to the sun because unfortunately I won't be directing the shots and the camera will be in the hands of someone who probably has never heard terms like "dynamic range".  As long as she can hold the camera steady, I'm hoping to salvage footage (when needed) that would otherwise be unusable.

Here's my entry.  Color graded through ACR, which in my mind has the best controls to recover highlights.  Second half is the original for comparison.

https://vimeo.com/72125235

Keep your eyes on the woman with the sunglasses on the other side waiting to cross the street.  Her sweater and face are appear to be completely blown out in the original, but are recovered in a very usable way.  

As for noise, this is an excellent example of the dynamic range of this camera.  Some whites were indeed blown out in this shot and were unrecoverable.  And the very dark shadows can be very noisy, and I did apply noise removal in my processing, which seems to work well.  13 stops are great but to get no blown out highlights and no noise in any outside lighting situation, we probably need something like 17+ stops.  

It is still better than anything I would expect out of any DSLR in the same lighting situation.