Jaimebailon
Well-known member
hmm well im not familiar with pal maybe someone who has worked with ntsc and pal can help you out
do you guys want butter? super smooth slow motion
- Thread starter Jaimebailon
- Start date
Ok, I'm looking forward to trying this out tomorrow. One thing i dont fully understand. If i shot in 24p (setting f5) with the 100b, i import the 24p footage as 29.97 from the tape. therefor my timeline is 29.97. Do I convert the 60i to 23.93 and then import into my 29.97 timeline with my 24p footage.... Ahhh this seems so different then what i'm used too.
yzmotoxer807
Well-known member
I really like this idea. People keep complaining about it not being slow enough though, but what if you then exported this after slowing it with the described process and then used a program like Shake that can adaptively retime, using interpolation, 24P video to a slower rate to slow it down more. I mean before I knew about this, I would use Shake to slow down just regular 24P footage to 50% using adaptive retiming interpolation and it looked good so I can just imaging how good it will look if you start with something that is already as slow as this. Just a thought. I'm definitely going to try it.
Jaimebailon
Well-known member
in one of my older postings i show the difference between the 29.97 timeline and 23.98.i would say shoot in 24pa (F6) and edit in a 23.98 timeline. but to answer your question if you are editing in 29.97 timeline you dont have to convert your clip therefore it will not be as slow as if you do conform to 23.98
Jaimebailon
Well-known member
yzmotoxer807 i can't wait to see please do post i would try this but i don't have shake. it sounds like an awesome idea though
eliasjklein
Member
i am waaaaaay to lazy to do all that stuff FC, i just ramp it in after effects
Any luck getting the screen shots up? I've followed all the instructions you gave through slowing both clips to 50%. Now, I'm trying to figure out what to do next -- I'm not sure how to "render" or merge the clips together in my timeline.
Sorry if this is unclear, but I'm pretty new to video editing.
tommetass
Well-known member
To slowmo with PAL cameras:
(I use adobe after effects)
Shoot 50i with shutter at 1/120 or higher. ( I would actually recommend 1/120 seeing as it gives a very nice motion blur)
Then slow it down to 50%.
Here's some samples:
http://www.youtube.com/watch?v=3G3nQaOKL5g&feature=channel_page&fmt=18
http://www.youtube.com/watch?v=DcJdV1DtI7w&feature=channel_page&fmt=18
http://www.youtube.com/watch?v=ryvYnXzVYQY&feature=channel_page&fmt=18 (Fast forward to 2:22)
(I use adobe after effects)
Shoot 50i with shutter at 1/120 or higher. ( I would actually recommend 1/120 seeing as it gives a very nice motion blur)
Then slow it down to 50%.
Here's some samples:
http://www.youtube.com/watch?v=3G3nQaOKL5g&feature=channel_page&fmt=18
http://www.youtube.com/watch?v=DcJdV1DtI7w&feature=channel_page&fmt=18
http://www.youtube.com/watch?v=ryvYnXzVYQY&feature=channel_page&fmt=18 (Fast forward to 2:22)
vx1kOrnot
Active member
This must be twixtor? I don't have twixtor, but i have a hard time believing Adobe after effect's Pixel motion would do this good of a job... This has no distortion around the skateboard's edges which is common with motion compensation...
EDIT: Wow, nevermind, after downloading and watching the original footage frame by frame there is some distortion around his feet and the railing..
Last edited:
vx1kOrnot
Active member
That can be fixed with changing the frame rate and the ramping its self
http://www.vimeo.com/2738068
Last edited:
yzmotoxer807
Well-known member
Finally did some tests!
Finally did some tests!
Okay so I finally did some tests combining the use of the aforementioned method of this thread and Shake. The first thing I did was the complete, "follow the directions" approach. I did exactly as the instructions at the beginning of the thread and the result was okay. The motion itself was buttery smooth with no motion blur however, I could see, as predicted and is inherent in this method, the result was half resolution. This only makes sense as the frames are separated into separate fields and then then these fields just copied to fill the frame via FCP's deinterlace filter. So: Motion=Good
, Resolution=Not so good.
My original speed, 30fps (60i deinterlaced for web) footage can be found on youtube: HERE or as a quicktime download: HERE
PLEASE USE THE DOWNLOAD LINKS. YOUTUBE DOES NOT REALLY SHOW ANYTHING AS THE YOUTUBE MAKES THE FOOTAGE JUMPY AND CRAPPY RESOLUTION.
I sought to figure out a way to fix this problem. After reviewing the steps I found that the problem lay in the deinterlacing. The fact that when you deinterlaced in FCP that it just copies the current field to the empty field's location wasn't cutting it and was halving the resolution. So... what to do. Well directly after conforming to 23.98 in Cinema Tools, I took the footage into Shake where I did the deinterlacing here. The difference is that Shake can interpolate the missing fields in a very smooth and accurate way. After exporting the two files (upper field and lower field) from deinterlacing in Shake, I then continued with the standard process of slowing by 50% in FCP and applying the blink filter to the top clip. As a side note I used a 1/1000th shutter speed as I think that 1/2000th looks too stroby. The final result of this process can be seen either on youtube: HERE or as a quicktime download: HERE (Don't mind the "pre shake" in the file name. This refers to the next step of my testing which is described below.)
So there I was, happy as could be with the 40% speed result (This is the final speed due to the 80% speed from conforming 30 to 24 then 50% after that. 0.8x0.5=0.4). But wait! Before this point I had used Shake to slow standard 24P footage down to, before it started to look a little noticeable, 70% and to a minimum of 50% which is as low as I would go because it looked really choppy. It does this using the same technology as when interpolating for deinterlacing. It can do what Apple calls "Adaptive Retiming" to interpolate the missing frames by tracking the pixels. So I thought, what if I now put the super smooth result of the last test back into Shake and adaptively retime it another 50% slower now brining the end result to 20% speed of the original clip. Doing this with standard 24P footage would surely result in crappy jumpy blurry and artifacty footage however, being that the result of the last test was already 40% speed but still looked as if it were brand new untouched 24P I gave it a try. It doesn't look amazing but it is passable. The resolution is there but it doesn't have the smoothness. However, a slight amount of this could be due to not using the best settings to avoid a long render. It looks better when you consider that you would need to shoot at 120fps to achieve this and that theoretically about 75% of the image information is completely made up and not real. The result can be found on youtube: HERE I will upload the quicktime later.
I will later try to do two more things. One, use the best settings and see if I can get the 20% to look better, and two, use the same technique to get a more reasonable 30% speed which will undoubtedly look better than the now existing 20%.
Let me know what you think.
Finally did some tests!
Okay so I finally did some tests combining the use of the aforementioned method of this thread and Shake. The first thing I did was the complete, "follow the directions" approach. I did exactly as the instructions at the beginning of the thread and the result was okay. The motion itself was buttery smooth with no motion blur however, I could see, as predicted and is inherent in this method, the result was half resolution. This only makes sense as the frames are separated into separate fields and then then these fields just copied to fill the frame via FCP's deinterlace filter. So: Motion=Good
, Resolution=Not so good.My original speed, 30fps (60i deinterlaced for web) footage can be found on youtube: HERE or as a quicktime download: HERE
PLEASE USE THE DOWNLOAD LINKS. YOUTUBE DOES NOT REALLY SHOW ANYTHING AS THE YOUTUBE MAKES THE FOOTAGE JUMPY AND CRAPPY RESOLUTION.
I sought to figure out a way to fix this problem. After reviewing the steps I found that the problem lay in the deinterlacing. The fact that when you deinterlaced in FCP that it just copies the current field to the empty field's location wasn't cutting it and was halving the resolution. So... what to do. Well directly after conforming to 23.98 in Cinema Tools, I took the footage into Shake where I did the deinterlacing here. The difference is that Shake can interpolate the missing fields in a very smooth and accurate way. After exporting the two files (upper field and lower field) from deinterlacing in Shake, I then continued with the standard process of slowing by 50% in FCP and applying the blink filter to the top clip. As a side note I used a 1/1000th shutter speed as I think that 1/2000th looks too stroby. The final result of this process can be seen either on youtube: HERE or as a quicktime download: HERE (Don't mind the "pre shake" in the file name. This refers to the next step of my testing which is described below.)
So there I was, happy as could be with the 40% speed result (This is the final speed due to the 80% speed from conforming 30 to 24 then 50% after that. 0.8x0.5=0.4).
But wait! Before this point I had used Shake to slow standard 24P footage down to, before it started to look a little noticeable, 70% and to a minimum of 50% which is as low as I would go because it looked really choppy. It does this using the same technology as when interpolating for deinterlacing. It can do what Apple calls "Adaptive Retiming" to interpolate the missing frames by tracking the pixels. So I thought, what if I now put the super smooth result of the last test back into Shake and adaptively retime it another 50% slower now brining the end result to 20% speed of the original clip. Doing this with standard 24P footage would surely result in crappy jumpy blurry and artifacty footage however, being that the result of the last test was already 40% speed but still looked as if it were brand new untouched 24P I gave it a try. It doesn't look amazing but it is passable. The resolution is there but it doesn't have the smoothness. However, a slight amount of this could be due to not using the best settings to avoid a long render. It looks better when you consider that you would need to shoot at 120fps to achieve this and that theoretically about 75% of the image information is completely made up and not real. The result can be found on youtube: HERE I will upload the quicktime later.I will later try to do two more things. One, use the best settings and see if I can get the 20% to look better, and two, use the same technique to get a more reasonable 30% speed which will undoubtedly look better than the now existing 20%.
Let me know what you think.
Last edited: