|
|
This topic comprises 8 pages: 1 2 3 4 5 6 7 8
|
|
Author
|
Topic: HFR - the reviews are in and it "didn't look so good"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Chuck McGregor
Film Handler
Posts: 47
From: Bremen, ME, USA
Registered: Mar 2012
|
posted 08-21-2012 03:44 AM
Ironically, I'm glad to hear that the "Motion Flow" (interpolating frames) doesn't pass muster, your word alone being certainly definitive enough to convince me.
I edited a paper about interpolating between two sonic impulse responses. The mathematics required were not your high school algebra. (No. I didn't edit that stuff, just the text.) From that perspective I'd say this was a trivial task compared to interpolating images. This is because interpolating images involves more parameters and these have to be applied to all those pixels. This has got to significantly raise the complexity of the math.
"Motion Flow" is probably a good example of technology tripping over itself. Digital opens the door to doing this "because we can." When music CDs first appeared the sound quality generally sucked. That's because the various choices of math first used for the Nyquist filters did exactly what they were designed to do: provide a steep cut-off at the Nyquist frequency. However it really screwed up the information below, like the stuff we could hear! Eventually it was realized that people actually had to LISTEN to this filter.
Similar problems exist with digital processors to this day, especially those for loudspeaker management. Set two filters on two different brands of processors to exactly same values and I can virtually guarantee they will measure differently, sometimes significantly at higher frequencies (Nyquist filter raising its ugly head). It's all in the math.
In the digital world it is easy to come up with math to do something with video or audio. However once the nuances of human perception are imposed upon that math, history has and continues to show that some super geek fresh out of MIT isn't going to get the math right, no matter what's in his/her books. Extremely reputable companies on the audio side still don't get the math right and we're about 30 years into this.
The engineer who wrote the aforementioned paper had to actually invent his own DSP algorithms to create some audio filtering he needed for loudspeaker processing (patent 8081766). He had read all the relevant PhD+ level books on digital signal processing and none had the answers he needed. Where human perception is involved the math for digital processing can get very non-trivial.
If your interested here is what the filtering he invented does:
Loudspeaker DSP
That this new "Motion Flow" interpolation kinda sucks does not surprise me one...so to speak...bit, let alone the mega bits involved.
Moral of the story (I guess for now): If you want more frames, shoot and display more frames
Just my 2 x 1 cent worth
| IP: Logged
|
|
|
|
|
|
|
|
|
|
All times are Central (GMT -6:00)
|
This topic comprises 8 pages: 1 2 3 4 5 6 7 8
|
Powered by Infopop Corporation
UBB.classicTM
6.3.1.2
The Film-Tech Forums are designed for various members related to the cinema industry to express their opinions, viewpoints and testimonials on various products, services and events based upon speculation, personal knowledge and factual information through use, therefore all views represented here allow no liability upon the publishers of this web site and the owners of said views assume no liability for any ill will resulting from these postings. The posts made here are for educational as well as entertainment purposes and as such anyone viewing this portion of the website must accept these views as statements of the author of that opinion
and agrees to release the authors from any and all liability.
|