Jump to content


Photo

Upcoming 2D-to-Stereo3D Conversion Software needs your input


  • Please log in to reply
2 replies to this topic

#1 Mehmet Gokyavuz

Mehmet Gokyavuz

    New

  • Basic Members
  • Pip
  • 2 posts
  • Digital Image Technician

Posted 07 February 2011 - 01:59 PM

Hello All,

I am part of a small group of graphics researchers working on an imaging algorithm that performs fully automated "One-Click" 2D-to-Stereo 3D conversion from 2D images, 2D video and 2D film footage. Fully automated conversion means that there are no manual 3D conversion steps like motion tracking, masking, layering, rotoscoping, 3D stereo rendering required. The algorithm analyzes 2D film/video frames automatically without needing any user input and creates Stereoscopic 3D frames from them, again without needing any user input:


2D Source Footage -> Automated Mathematical Analysis -> Automated 3D Conversion -> Stereo 3D (Anaglyph) Footage


We are not new to the subject. This year is our 4th year of Research and Development on the subject of 2D-to-3D conversion and we are not terribly far from having a "Production Quality" or "Production Ready" 2D-to-3D software solution. We are a few months to maybe 1 year away from that point.

When we have our 2D to 3D conversion software ready to use (hopefully this year, within 2011), we want to use the following model for making money from it:

1. The software is free to download and test/play with

2. The trial software is the full product without crippled features, watermarks or other restrictions

3. Users of the software pay a conversion fee only when a 3D converted project is ready to be exhibited/distributed/broadcast/webcast.

Examples: (prices aren't final or real... just examples)


5 min Music video for viewing on Youtube/Dailymotion/Vevo = 500 dollars 3D conversion fee

1 hour documentary for TV broadcast = 1,000 dollars 3D conversion fee

1 hour TV series episode for TV broadcast = 2,500 dollars 3D conversion fee

2 hour long Indy Film for DVD/BluRay dist. = 5,000 dollars 3D conversion fee

High Budget Hollywood Film for Cinema = 50,000 dollars 3D conversion fee


The whole thing would be based on an "Honors System" where we trust production professionals to look at the price list, find the category for their project and pay the appropriate one-time 3D conversion fee for the project in question (online via Credit Card, PayPal or bank transfer et cetera).


I would like to ask the users of this forum 2 questions about this approach:


1) Would most production professionals honor the "trust system" and pay us the appropriate fee when a project has used our software to Stereo 3D convert and the project is finalized?

2) What do people think would be appropriate/realistic prices to charge for conversions like these

30 second commercial = ???? dollars

5 minute Music Video = ???? dollars

10 - 20 minute Short Film = ???? dollars

30 minute TV program = ???? dollars

1 hour TV Documentary = ???? dollars

1 hour TV Series Episode = ???? dollars

2 hour Independent Film = ???? dollars

2 hour Hollywood Film = ???? dollars


These are the questions I would like some forum user input/opinions on.

Unfortunately I cannot post any images or videos of our 2D-to-3D technology yet because we are in the middle of patenting and copyrighting the various mathematical methods used in the algorithm.

All I can say is that we will keep working hard until the Stereo 3D conversion results are "Top Notch" quality.

Thank you for reading this, and hoping to hear people's opinions on whether the "pay a one-time conversion fee when your production project is finished" I talked about above can work well.

Best Regards,

M.Gokyavuz
Software Developer
Graphics Researcher
  • 0

#2 Chris Millar

Chris Millar
  • Basic Members
  • PipPipPipPip
  • 1642 posts
  • Other

Posted 07 February 2011 - 03:26 PM

The algorithm analyzes 2D film/video frames automatically without needing any user input and creates Stereoscopic 3D frames from them, again without needing any user input:

Except when it goes wrong, which begs the question>> how far along is your one button UI developed for us to correct it ?

We are a few months to maybe 1 year away from that point.

Sweet


Define the distinction between indy film and hollywood film - legally :blink:
  • 0

#3 Mehmet Gokyavuz

Mehmet Gokyavuz

    New

  • Basic Members
  • Pip
  • 2 posts
  • Digital Image Technician

Posted 08 February 2011 - 11:21 AM

Except when it goes wrong, which begs the question>> how far along is your one button UI developed for us to correct it ?


Valid question.

Imagine that 20 short mathematical calculations go into creating a Stereo 3D frame from a 2D source frame.

Each of these small calculations contributes about 5% of the overall Stereo depth or "3Dness" of the converted 3D frames.

Now you get a really awkward shot (unusual lens, perspective, framing, lighting, DOF, zoom, motion, color; copious amounts of fog, dust, smoke or rain in the frame) and 8 out of the total 20 calculations fail to work properly.

Put as a percentage, this means that a whopping 40 percent of the 3D conversion process fails or produces no usable results.

In 99% of cases this won't create a "strange" or "messed up looking" shot. It will create a shot that is merely "flatter" and more "2D looking" than it could have been, because the algorithm was able to extract only about 60% of the 3D depth that is in the 2D shot.

On the spectrum going from "2D looking" to "3D looking", your problematic frame or shot will land on the 60% percent line:


2D looking >-------------------------------------------> Problem Frame <----------------------< 3D looking


With time, we will hopefully find better mathematical tricks and ways of identifying and dealing with these "problem frames".

From our testing though, these problem shots and frames aren't very frequent and are often only visible for a fraction of a second.

Example: A galloping horse passes so close to the camera lens that its body blocks out almost the entire frame and all that is there to analyze for several frames is a massive amount of gray-white motionblur.

Most viewers won't care that there is no eye popping Stereo 3D effect in the 1/2 second that the horse takes to gallop out of the frame and out of sight.

Define the distinction between indy film and hollywood film - legally :blink:


The distinction is in the budget. E.g. Under 2 million dollars = Indy Film
Over 2 million dollars = Decently funded film
Over xx million dollars = High Budget Hollywood Film/Blockbuster

Where you place the exact "budget cutoff" for each category is an interesting problem of course. We will need to do more research on film budgets to determine this.
  • 0


The Slider

Technodolly

CineTape

Visual Products

Metropolis Post

Broadcast Solutions Inc

Opal

CineLab

Aerial Filmworks

Media Blackout - Custom Cables and AKS

Wooden Camera

FJS International, LLC

Willys Widgets

rebotnix Technologies

Rig Wheels Passport

Paralinx LLC

Tai Audio

Abel Cine

Gamma Ray Digital Inc

Ritter Battery

Glidecam

FJS International, LLC

Paralinx LLC

Aerial Filmworks

CineLab

Wooden Camera

Rig Wheels Passport

Ritter Battery

Gamma Ray Digital Inc

Media Blackout - Custom Cables and AKS

Tai Audio

Opal

Broadcast Solutions Inc

Technodolly

The Slider

Visual Products

Metropolis Post

Willys Widgets

Glidecam

rebotnix Technologies

Abel Cine

CineTape