Motion Cadence

[jonpais]

Aren't online videos Long GOP? How can you tell the difference between these cameras by comparing them on YouTube? Don't know for sure, just asking...

[mercer]

That is a good question I don’t have the answer to but I could tell by looking at them on my computer after using the cameras. I assume just capturing it affects the motion as recorded and anything after that wouldn’t change it?

[jonpais]

We're both in over our heads.  

[mercer]

Yeah, honestly I don’t even care why. I just notice what I like and what looks better to my eyes. How it gets there... whatevs... lol. 

[Django]

shutter angle, all-i/ipb & global/rolling shutter probably affect motion cadence the most.. hence cameras like digital bolex winning in that dept imo.

[kye]

7 hours ago, jonpais said:

Aren't online videos Long GOP? How can you tell the difference between these cameras by comparing them on YouTube? Don't know for sure, just asking...

I don't know for sure, but I suspect it's about maintaining quality throughout the workflow.  For example, if you were to have a long-GOP capture format then it will 'bake-in' motion issues, and then those issues may be worsened by intermediate processing steps (and of course each time you round-trip you're baking things in) and then the final delivery format gets put over the top of all of that.

If the imperfections in the way that the long-GOP behaves are worsened by similar issues in the input then you might find that the motion issues compound, or even worse, are converted by some other limitation in the format into some other type of secondary issue.  

 

[jonpais]

Up until now, all this motion cadence conversation has been anecdotal. I was hoping to see some real life comparisons, just like we regularly see for rolling shutter artifacts. If it were truly an issue, we'd be hearing reviewers talk about it; manufacturers boasting of their superior motion cadence; we'd see some technical papers published about it; directors would all be mentioning it; websites would be testing for it; prominent filmmakers would certainly have brought it up at the Zacuto shootout; the comments section of mirrorless camera reviews on YT would be overflowing - but the fact remains, none of this has happened. And if you do a Google search, EOSHD is pretty much the only website on the planet that regularly brings it up. Which is why I'm inclined to believe it's the EOSHD server. 

[sam]

did you try searching for "judder" ? that might help. millions of posts about that. the link I posted earlier in this thread where none other than Geoff Boyle weighs in, should at least convince you its a real thing. my opinion is it only really comes into play if you are trying to emulate film on the acquisition side. on the monitor side, have you ever watched a plasma screen?  lcd(oled) has very different motion. videophiles prefer plasma based on all the posts I could find, and i do as well. the vague terminolgy combined with an extreme amount of variables  makes it hard to define. 

I think for most viewers, its a subconscious thing.  I show friends and family back to back samples of  my 1dc, F35, Fs700raw4k, and a newer iPhone without letting them know which is which. then ask which they prefer, and to also describe the look and feel. I've tried it on different screens as well.  

the F35 has always been picked as favorite and usually something to the effect of "oooh that one looks like a movie"   not scientific in the slightest, but,  what is it people notice?

[jonpais]

 

@sam Do you have a side-by-side comparison? Otherwise, it’s just anecdotal.

[mercer]

Do you have any side by sides to prove it doesn’t exist? You brought up the topic to prove that motion cadence doesn’t exist, so the burden of proof lies on you.

[IronFilm]

3 hours ago, sam said:

I think for most viewers, its a subconscious thing.  I show friends and family back to back samples of  my 1dc, F35, Fs700raw4k, and a newer iPhone without letting them know which is which. then ask which they prefer, and to also describe the look and feel. I've tried it on different screens as well.  

Friends and family must love you!

[cantsin]

The term "motion cadence" does not exist in any scientific or engineering literature, not even any published work on cinematography or videography, but only as an urban myth on camera forums.

[cantsin]

Did some historical digging into the web. The earliest occurence of the term "motion cadence" is from 2006, in this forum posting, where it refers to the look of After Effects' slow motion effect as opposed to that of Vegas: "From my own personal perspective, the slow-mo material I've encoded using Vegas looks no worse than the sort of slow-mo segments one sees in broadcasts. Other HDV camera users have found that After Effects has provided the sort of slow motion cadence that they required."

The next mention is in a 2007 DVXuser review of the XH-A1 camcorder: " 24F is Canon’s simulated progressive scan system.  24P is Panasonic’s true progressive scan system.  24F results in a resolution drop of something like 15% on the Canon; 24P delivers the full resolution of the Panasonic chipset.  But in practical terms, when you get right down to it – it doesn’t make much difference.  Both deliver film-style motion with proper 24fps motion cadence."

So here, "motion cadence" is simply a synonym of frame rate (in a time when  24p video recording was a luxurious rarity).

The next mentions are in a 2009 forum posting on video projectors: "When characters in a movie become actors on a set, frame interpolation has gone too far. I used to play with frame interpolation (Trimension) on my pc outputting to a 106" projected image, it certainly was a neat effect, but the artifacts, motion cadence inconsistencies and soap opera effect were ultimately deal killers for me".

Again, "motion cadence" just refers to frame rate.

In the same year, the term turns up on DVXuser again in the context of frame rate conversions: "A more advanced way of converting 60 frames to 24 is using an "intelligent" processes that will use the actual 60p frames information to guess how each frame would have been if the footage was actually shot in 24p, this might be called frame interpolation or frame resample.  Any of this methods result in a better motion cadence, due to the regular time separation between the resulting frames."

In 2010, the term resurfaces on REDuser , in the same context of frame rate resampling.

I'd argue that any argumentation of the quality of motion rendering (to put it into less esoteric terms) really is about shutter speeds relative to frame rates and rolling shutter effects. Every other discussion of "motion cadence" probably has to do more with the fact that most people on this and other fora shoot 24p but do not have native 24p displays (and native 24p graphics card output) so that motion will always look quirky and be influenced by one or more motion interpolations in the playback chain.

 

[sam]

sure. it's a made up word. when in Rome, no?  regardless of the terminology and what causes motion(put your favorite descriptive word here..... blur, judder, etc...) it doesnt change the fact that truly world class Dp's are discussing "it" in a thread with cadence in the title if you have not visited the link in prior posts.

  regarding my own fiddling, which is already admittedly unscientific in every last way, the F35 was always chosen. why? I attributed it to the ccd vs cmos but who knows?     

 

2 hours ago, IronFilm said:

Friends and family must love you!

the love comes to an end quickly!

[sam]

I set up these cams (although more carefully aligned in the actual video) Details and settings are in the YT description.  

[BTM_Pix]

Great test.

I'm not sure I'm any the wiser about motion cadence but after about 55 seconds of watching that pendulum I definitely felt that I could've been prompted into giving up smoking and getting over my fear of flying

 

[AaronChicago]

Theres definitely a big difference in the F35 and S8+. 

[sam]

What surprised me, even though the clips were synced, there appears to be drift by some of the cams. And not the kind where one or more of the pendelum shots either pull ahead or fall behind the others due to a tiny difference in frame rates like 23.98 vs 24. The clips seem to be drifting minutely ahead in one frame and then getting slightly behind in another and so on through out the clip, but they start and finish all together.  

What's easiest to see is the impact of rolling shutter. And we've all seen at least one of the multitude of rolling shutter tests involving cmos sensors.  In the discussions that revolve around these tests, someone usually chimes in with "well the Alexa has a cmos sensor" and then the conversation turns to  "faster read out speed" This, http://tessive.com/mechanical-shutter-comparison/ , which many of us have seen before, at least outlines a difference, stark or otherwise.  Whether this is part of what people notice, or not? 

   

 

 

 

 

 

[sam]

 I forgot to mention the lighting was natural light from the sun. No other light source to possibly interfere.  

[jonpais]

The Samsung looks funky.