Searcher
Well-known member
If that's not pixel peeping then I'd hate to see what is...
Canon EOS C300 NEW pricing!!!
- Thread starter Bertholt4
- Start date
- Status
digitalfilmmaker
Active member
True, but I just remember the hair pulling and BS when Leo Ticheli posted the first AF100 footage, fair is fair. 
Rick Burnett
Section Moderator
How much resolution do you lose? Definitely something I'd want to test. From what I have seen, it's a pretty significant drop in vertical resolution.
KyleProhaska
Veteran
You definitely lose some...even if it's significant...you're still getting the C300's source footage to start with instead of the FS100 or AF100...AND...you get all the other benefits of using a camera you're already set to record everything else with. Why grab another cam?
If anyone notices the resolution drop...then there's a lot more wrong with the shot then that.
Rick Burnett
Section Moderator
I'm more worried about the blending of the frames with such spatial time differences then the resolution drop. The examples I've seen of people using methods to combine 2 real interlaced frames just doesn't look good to me. Now, granted I've never done them myself, but with the action stuff I film, the movement can be at least fast enough that I feel that the blended frame on the movement portions between both frames would be too great and the motion blur that get's generated by the frame blend just looks wrong to me. I guess my first test if I keep the FS100 for a bit would be taking 720p out of the C300 and upsizing it, then taking 60i and frame blending them in AE and then taking the FS100 at 1080p60 and comparing them.
As someone who is interested in slow motion I'd be very interested in these comparisons.
KyleProhaska
Veteran
There's no problems with this method, trust me. If you shoot at the proper shutter, it'll look fantastic. 1/120 is the proper shutter to use if you want a motion blur similar to 1/48 at 24p. I suggest you try the method out and see how it looks. I've never seen anything in the image besides a slight resolution drop that would cause any concern.
Rick Burnett
Section Moderator
First, thanks for your insights. I think the part I was getting confused at is I am not combining 2 frames, now that I think about it. I am taking one frame at 60i and doubling up the information basically to fill in the holes. To me that would mean the vertical resolution is really half right? You are missing half of the frame. To me, that would seem like a lot?
On cameras where the resolution isn't that close to 1080p anyways, I could see it not being as big a difference, but on the C300, it has a lot more initial resolution due to the design. It would seem to me that having that big a difference in vertical and horizontal resolution might be noticeable.
(What I am saying is if your camera only resolved say 75% of resolution of the frame size it stored at, if you dropped to 50% in one direction, you'd be a difference of 25% resolution whereas with the C300, the resolution drop would be 50% relative to actual perceived resolution). I guess if you did these tests on other 1080i60 3CCD cameras you'd have the same situation.
Well, I'll test it anyways
I'm more curious to see how 720p60 upscaled on the C300 looks to the 1080i60 done as you suggest more than anything. As I agree, working with one set of footage would be easier, without question.
KyleProhaska
Veteran
By all means test...I would like to know myself how it holds up. Your delivery format could be considered here as well, because unless you're seeing the footage in the best way possible like projected for instance, nobody is likely to notice for those handful of slow-mo shots you have. The resolution drop is significant, but don't forget it's only vertical so there's plenty of detail left to satisfy.
Lots to consider here, but for me...it's the answer to the C300 not having any 60p options. That is of course, unless the 720p is that good that the upscale can best it, but somehow I doubt it. You're enlarging pixels as opposed to losing detail that's already there. Enlarging all the pixels vertical and horizontal vs. losing some of the existing vertical seems pretty one sided to me. But...you're more then welcome to prove me wrong.
Lots to consider here, but for me...it's the answer to the C300 not having any 60p options. That is of course, unless the 720p is that good that the upscale can best it, but somehow I doubt it. You're enlarging pixels as opposed to losing detail that's already there. Enlarging all the pixels vertical and horizontal vs. losing some of the existing vertical seems pretty one sided to me. But...you're more then welcome to prove me wrong.
Rick Burnett
Section Moderator
I definitely don't know if it is better or worse, that's for sure!
KyleProhaska
Veteran
I think what you'd be likely to see is slight aliasing in the image if you did the 60i to 60p method (although there's some really good tools out there that can do a good job) and the resolution drop sorta giving the appearance of softness to the image. So it'll lose some punch for sure. The alternative even if the 720p on the C300 is gorgeous and sharp...means blowing up all those pixels so an inherently softer image across the board...BUT depending on the source and the method used to uprez, it could look great.
If you're going to DVD, neither of these is a bad option. For me SD delivery is the majority so either or would work fine and the viewer would never be the wiser.
If you're going to DVD, neither of these is a bad option. For me SD delivery is the majority so either or would work fine and the viewer would never be the wiser.
Excuse me....would some one mind preparing me a plate of crow? I just might have to eat some....
Thanks for posting/sharing!
bruceallen
Well-known member
With a good algorithm, such as RE:Vision FieldsKit / Twixtor... there is motion detection. It goes and looks at the frames before and afterwards.
So... if you have a shot of a landscape and there is very little change from frame 1 to frame 2, then there is no need to deinterlace. Just use the data from both fields.
If there is a change (over a percentage that you specify), then it discards one of the fields and interpolates. That's it - there is no temporal blending / weirdness going on at all.
Thing is: at 24p the moving stuff often has so much darn motion blur that you don't often see the resolution loss.
Worst case scenario is something like a camera wobbling up and down while trying to film a sign. There you lose 50% vertical resolution. But if your camera was wobbling... chances are there was motion blur and you weren't getting the full 1080p resolution anyway. So, you have low-res blur... instead of high-res blur. Oh noes!
Best case is a locked-off shot with a small amount of motion. There you lose very little resolution.
Only fly in the ointment is if the C300's 60i mode has a little bit of vertical blur added digitally anyway to avoid interlace flicker. EG so even if you pointed it at a stationary test chart, its 60i wouldn't resolve as well as its 30p or 24p. Some cameras do that.
Well... to me yes... but then again I can clearly see the difference between the AF100 and the FS100 footage in Barry's test - to me, the AF100 looks like it lacks luma resolution when you download the file and view it at 1080p... but then most people don't seem to mind.
Again, yes, I can see it and it bugs me - because editors do that sort of stuff all the time and sometimes it sneaks its way onto broadcast TV. General public doesn't seem to spot it often though.
Also, bear in mind that a lot of times you're going out to a 1080i broadcast. So your vertical resolution may be getting screwed with anyway.
Anyway, it will be interesting to do tests. Though probably just when you've got a nice test done, Canon will come out with a C350 or whatever that does 48/50/60p
Bruce Allen
www.boacinema.com
Last edited:
The difference in resolved detail between the Af100 and Fs100 to me is pretty obvious, but at the same time it was strange to see just how similar they were in other areas as far as that test was concerned.
J Davis
Veteran
from what I have heard, the c300 is the front of a range,
edit:
but y'know it doesn't even have to be hearsay, canon nearly always do this. Like ... how many vixia's are there now?
Last edited:
Rick Burnett
Section Moderator
Well given I owned an AF100 and FS100 at the same time an evaluated both, I can say I felt the FS100 had more perceived resolution, but, to be fair, I dialed down the sharpening setting of the AF100 quite a bit because I didn't like what their sharpening algorithm did to the noise.
The problem with a locked off shot at 60i is that, well, that defeats the purpose of using slow motion for us!
For instance, with the AF100 I shot a car crash for a film, and it was FANTASTIC. It had A LOT of movement. With the FS100 so far, I've mostly done a little fighting and some epic gun shooting scenes where a character is spraying bullets across the FoV and I used 60p to really sell the epicness of it. In both, there is quiet a bit of movement.Most of the Twixtor stuff I have seen is taking something like 24p or 60p and slowing it down further, making independent frames that have NO data in them to start, whereas with the 60i, you have half a frame
But yeah, now that I have researched even more, with how some of the algorithms do use the frames before and after, you can get some aliasing due to the fact you really do have half the resolution. We'll just have to wait for testing! And pertaining to Barry's example between the AF100/FS100, I think it's a great test and shows how close the cameras can be.
Given most of our content gets released on YouTube, I am not horrible worried about 720p for that stuff. What I do want to see is examples of what the internal processing is doing to make 720p. Down they downsample the frame REALLY matters to me. I'm sure it will be better then they do on their DSLRs which is absolutely horrible, but I've just not seen any examples yet.
But even looking forward, I have one scene in my film I wrote that I am producing this year that needs a good slowmo shot, and the more I think about it, I think I might just rent for that one. I think it needs to be at least 120fps to get the look right (splashing water).
This guy did a great job. Made me laugh out loud more than once.
Jim Martin
DVXuser Sponsor
Dr-
On the pricing, it looks like we've been pushed to the later part of next week for the final pricing......but I'll post as soon as I know......
Jim Martin
Filmtools.com
On the pricing, it looks like we've been pushed to the later part of next week for the final pricing......but I'll post as soon as I know......
Jim Martin
Filmtools.com
Love the video review... but the high ISO stuff really doesn't look good. Neither do those stopped down night shots. Looks very video like. But, again, I guess it's not terrible for ISO20k.
Also, shallow DOF at night with bokeh looks great... always will. I don't know why people always need these gimmicks like highISO and now "stopping down" at night to pretend like their footage is somehow more special than others' now. I think the stopping down to f5.6 was a good practical example, but the bridge stuff looked straight-up terrible...
Then remember when the AF-100 came out and people weren't pleased with the higher-ISO performance? There was a whole thread about how obnoxious high ISO shooting is and why "professional" shoots don't need high ISO because they have lighting. Now, all the sudden, since the C300 is pretty lack-luster in the features department... highISO is now the holy grail of performance. There's is just some serious mental gymnastics going on for people to justify this high-priced professional stigma.
Source:
http://www.dvxuser.com/V6/archive/index.php/t-236143.html
Also, shallow DOF at night with bokeh looks great... always will. I don't know why people always need these gimmicks like highISO and now "stopping down" at night to pretend like their footage is somehow more special than others' now. I think the stopping down to f5.6 was a good practical example, but the bridge stuff looked straight-up terrible...
Then remember when the AF-100 came out and people weren't pleased with the higher-ISO performance? There was a whole thread about how obnoxious high ISO shooting is and why "professional" shoots don't need high ISO because they have lighting. Now, all the sudden, since the C300 is pretty lack-luster in the features department... highISO is now the holy grail of performance. There's is just some serious mental gymnastics going on for people to justify this high-priced professional stigma.
Source:
http://www.dvxuser.com/V6/archive/index.php/t-236143.html
Last edited:
- Status