Have a 60fps HDMI frame capture card/box?

[HobeSoundDarryl]

We're having a debate over in this thread: https://forums.macrumors.com/threads/youtube-app-can-now-play-videos-60fps.1893965/ . Several guys think that the relatively new YouTube app on the TV3 can now play 720p 60fps video to their HDTVs. Most of the support is subjective- people claiming they can see 60fps via select YouTube videos that do indeed have 60fps options available to computers with enough horsepower to play 60fps versions. I'm skeptical that TV3 can output 60fps, believing that even if it's fed a 720p60fps video (as I've done myself), it is dynamically cutting 60fps to 30fps and that's what is pushed out to our HDTVs.

The key here is a hypothesis that the new YouTube app might be doing something with software that is beyond the hardware specs shared by Apple- a video fps hack if you will, squeezing more fps out of TV3 than previously published and/or tested. So we need a new, objective test. Does anyone have a frame capture box, card or some kind of fps meter capable of receiving 720p 60fps video via HDMI? If so, could you hook up your TV to your box and try a few of the videos referenced in that thread to see if it actually is outputting 60fps? We'd all love to know that the TVs we already own have this capability. Thanks!

 

[lostless]

Just tested myself. Youtube app on the apple tv, clearly 60 fps. No doubt. Streaming from the iOS app, 30 FPS. Google really needs to enable it on iOS.
The limitation of the apple tv is bit rate, not resolution or frame rate. Ive always been able to run 60FPS material on the apple tv without it being down converted to 30 using my own videos. I think Apples specs are very generous so people that try to push the limits, won't complain that the apple tv can't keep up.

 

[HobeSoundDarryl]

You tested it with hardware (objectively) or eyes (subjectively)? Apple's specs say 30fps. I've fed it 1080p60fps video that I shot myself and know for sure is 60fps and, even with relatively low bit rates (well below Apple specs), it chokes if there is much on-screen movement, especially camera panning. Of course 720p60fps could be different and that's what we're trying to know.

We need it hooked up to a HDMI framegrabber or some kind of fps meter to confirm or refute what is actually coming out of the TV. Did you do that?

 

[lostless]

I originally did the test with my eyes. I can easily see the difference between 60 and 30 FPS. But I did do an "objective" scientific test to test your hypothesis. I don't have a HDMI capture card, but i do have an iPhone 6 that shoots at 60 FPS. Started the apple TV youtube app, started a fast moving 60FPS video of Mario Kart, where the frames will be easy to tell the changes. Captured my TV screen at 60FPS. Imported the video to my computer and scrolled frame by frame. Every one of the 60 frames was uniquely different. The apple TV is definitely playing back youtube at 60FPS. Still streaming from iOS to the apple tv from the youtube app, is still only 30.

 

[Billy B.]

You tested it with hardware (objectively) or eyes (subjectively)? Apple's specs say 30fps. I've fed it 1080p60fps video that I shot myself and know for sure is 60fps and, even with relatively low bit rates (well below Apple specs), it chokes if there is much on-screen movement, especially camera panning. Of course 720p60fps could be different and that's what we're trying to know.

We need it hooked up to a HDMI framegrabber or some kind of fps meter to confirm or refute what is actually coming out of the TV. Did you do that?

What are your subjective tests telling you? Do 60fps vids on your Apple TV look smoother than 30fps vids? I know some people cannot see this difference.

 

[HobeSoundDarryl]

What are your subjective tests telling you? Do 60fps vids on your Apple TV look smoother than 30fps vids? I know some people cannot see this difference.

I have a pretty good eye for quality myself. I tested the video I referenced the other day in the other thread. I made a 720p 60fps version and a 720p 30fps version. To my eyes, they looked about the same via TV. Thus, I suspected that feeding it a 720p 60fps video has it dynamically down-converting to 720p30fps.

As you may recall from the other thread, I also rendered a 1080p60fps file and TV tried to play that but then choked. Why? My hypothesis is that, unlike a 720p60fps file, there is not enough horsepower on board to down-convert a 1080p60fps to 1080p30fps file for playback. I uploaded that same 1080p60fps file to YouTube and then played it on the same TV and it played flawlessly. That could mean that the YouTube app is doing something better than the Apple (stock) player app so that it can play 1080p60fps files too... OR that YouTube is feeding a request from an TV for a 1080p60fps uploaded file something the TV can actually play smoothly (something I presume is < 1080p60fps). That same concept- that files claiming to be 720p 60fps on YouTube are not necessarily what YouTube is feeding to TVs trying to watch such videos- adds to the skepticism.

With all due respect to lossless in this thread, his "test" simply adds another score for those judging 60fps vs. 30fps with their eyes. Eyes can see magician's turn their assistants into tigers. My own eyes looks at that apparent 8K YouTube clip referenced in the other thread and it both looks super sharp and super smooth but I know it's neither better than 1080p and highly likely not 60fps.

I've laid out a completely objective test which is why I created this thread. Hook up a HDMI frame grabber to TV, play these videos you and others reference in the other thread, (objectively) use the frame grabber to identify the fps it is receiving from TV as those videos play. The hardware test cannot be deceived. It will "know" if it is getting >30fps out of the TV. That will be completely objective.

I created this thread because even some hollywood talent will visit this part of the forum. Many of them are going to have access to such hardware and only one of them needs to do the test to confirm or refute the hypothesis. As I shared in the other thread, I (too) hope it's >30fps coming out of TV but I am skeptical until some hardware confirms it.

If 100 more guys come into this or the other thread and claims their eyes (too) can tell that they are seeing 60fps, I won't lose my skepticism... just as the whole audience's eyes can witness the magician's assistant being turned into a tiger. That hardware test will not be deceived. That kind of hardware costs less than $200. Somebody will have it + an TV and do the test for us, so we can all know for sure... which is all we are after.

 

[Billy B.]

I have a pretty good eye for quality myself. I tested the video I referenced the other day in the other thread. I made a 720p 60fps version and a 720p 30fps version. To my eyes, they looked about the same via TV. Thus, I suspected that feeding it a 720p 60fps video has it dynamically down-converting to 720p30fps.

As you may recall from the other thread, I also rendered a 1080p60fps file and TV tried to play that but then choked. Why? My hypothesis is that, unlike a 720p60fps file, there is not enough horsepower on board to down-convert a 1080p60fps to 1080p30fps file for playback. I uploaded that same 1080p60fps file to YouTube and then played it on the same TV and it played flawlessly. That could mean that the YouTube app is doing something better than the Apple (stock) player app so that it can play 1080p60fps files too... OR that YouTube is feeding a request from an TV for a 1080p60fps uploaded file something the TV can actually play smoothly (something I presume is < 1080p60fps). That same concept- that files claiming to be 720p 60fps on YouTube are not necessarily what YouTube is feeding to TVs trying to watch such videos- adds to the skepticism.

Ok, now on your Mac or PC how does that 720p 60fps vid look compared to the 720p 30fps? If you can see the difference.. THAT is smoothness that I'm seeing on Apple TV. There's no mistaking it.

I don't have a capture card but I did another test. I went and took a 720p 60fps video and spliced it into 20 short videos. I then converted those same 20 vids to 720 30fps. Then threw all 40 (20 60fps - 20 30fps) into iTunes, created a playlist and mixed up the order so they were totally random. I then Airplayed them to my Apple TV. I could easily see the difference. So much so, that my "guesses" of 60fps or 30fps was 100% correct! Just to make sure, I went and mixed the playlist up again... Again 100% correct. Maybe it's just an absurd amount luck and I should probably just head out to Las Vegas ASAP!

But what is more likely is one of three things:
1. You can't see the difference between 30fps & 60fps.
2. Apple TV is now doing some sort of very natural looking frame interpolation, but this frame interpolation only works on 60fps video.
3. Something in your setup or settings is causing 60fps videos to only display @ 30fps.

 

[HobeSoundDarryl]

The counter to your hypothesis (you have the hypothesis that it is able to play 720p 60fps) is that human eyes are getting fooled. That's what makes tests based on human eyes subjective. Again, if the magician turns the assistant into a tiger for 100 shows in a row and you attend every one of those and witness it 100 times, did the assistant actually turn into a tiger? If 1,000 sets of eyes attend each show and they all witness the assistant turn into a tiger, 1000 sets of eyes can't possibly be wrong. If the magician does 2 shows a night for 300 nights per year (600 shows) all packed with 1000 sets of eyes in the audience and all 600,000 sets of eyes witness the assistant turn into a tiger, who could possibly refute such overwhelming evidence witnessed? She certainly turns into a tiger in each show. No doubt about it.

Your new test adds something to your hypothesis. If it was random and if you had no way of knowing which clip was which, and you had perfect success picking which were 60fps vs. which were 30fps, it adds to your case. However, the key variable at play is your eyes making the judgement. It's still subjective vs. objective, unlike when someone hooks up any kind of hardware that can read FPS out of TV and leave no doubt about what is coming out of there.

Perhaps in converting 60fps to 30fps clips from your 60fps master, something else changed that made them look less sharp. For example, since the master file was 60fps, maybe it passed through the original quality for the 60fps clips but re-rendered for the 30fps, suffering a generational loss of quality? Perhaps on playback, the TV's approach to dynamically converting a 60fps file tossed at it yields a better looking 30fps output than the rendered 30fps clips you made. That could be a plausible explanation why your eyes were able to pick out superior-looking clips from rendered. And, since you want to believe your original hypothesis, you are biased to the outcome. That is why we need a truly objective test.

I know that the magician is fooling my eyes (I know he doesn't really turn the assistant into a tiger). However, if I wanted to believe the magic was real, then maybe I could convince myself that he actually is turning her into a tiger. All I need to do is believe what I think I'm seeing enough and then it really is as I see it (to me).

As I've shared before, I hope you are right. I hope it is actually 60fps video coming out of TV. But the test that absolutely proves it can't rely on someone's eyes thinking they are seeing something... especially if those eyes are already somewhat biased in what they think they are seeing. Apple's specs are clear: 30fps. There's not a confirmation of this anywhere ELSE on the whole Internet- especially among the AV enthusiasts who would be VERY EXCITED about this. The only claims that this is true are 2 threads on one website: MacRumors.com.

We need objective proof if we want to prove this with certainty. If we want to make a pretty good case but leave room for skepticism, mission accomplished. Even I hope it proves to be true. But I want the certainty to be fully convinced.

 

[Billy B.]

If Apple tells you the Apple TV can display 60fps vids, will the motion of your vids magically look better? That's strange. My eyes can perceive 60fps motion without someone telling me what my eyes are seeing. If you record a live MLB game, which is obviously 60fps and made a 30fps recording of the same vid... could you tell the difference using just your eyes? I know I could. This comes down to either your system isn't displaying 60fps, or your eyes cannot perceive the difference between 60fps & 30fps.

 

[lostless]

With all due respect to lossless in this thread, his "test" simply adds another score for those judging 60fps vs. 30fps with their eyes. Eyes can see magician's turn their assistants into tigers. My own eyes looks at that apparent 8K YouTube clip referenced in the other thread and it both looks super sharp and super smooth but I know it's neither better than 1080p and highly likely not 60fps.

But the test was not done with my eyes!???????? I did do an "objective" scientific test with a 60FPS camera pointed at my tv and the result is a 60 fps video captured. EVERY FRAME from my IPHONE CAPTURE was uniquely different. It may not be an hdmi capture, but there is no way that frames on captured video just magically appear. If it was 30fps, each frame of my capture would repeat the same image on my tv screen for 2 frames. I say do the test yourself if you have a go pro or iPhone 6 that captures at 60FPS. I did some more tests and ran into the 1080P 60 FPS issues you did. Dropped frames and huge skips. It started at 60, dropped to 30 or worse and tried to come back to 60. Even tried lowering the bit rate in handbrake and same issue. Dropped that 1080P to 720P. NO ISSUES. played back a 480P 60FPS. NO ISSUES. So the apple tv can clearly play 60FPS, but does not have the hardware to keep up with 1080P 60FPS video. It doesn't convert the video as you say, to 30FPS, It just gives up drops frames. And when it can't keep up, It just drops many seconds of frames.
Im assuming that the youtube app is only sending the 720P 60fps feed

 

[Billy B.]

Hey lostless, I just tried your test and you're 100% right! This definitely proves that Apple TV (and YouTube on ATV) are playing @60fps!! I recorded a 60fps Mario Kart 8 video on my iPhone 6 from the Apple TV, same as you. I played it back in Quicktime Player on my PC. Between each second was obviously 60 frames, and all 60 frames were unique. I tried the same thing was a 30fps video. This time just every other frame was unique. Undeniable proof right there! Thank you!
I knew Apple TV could play 60fps vids. I Airplay videos from Giant Bomb all the time... and since at least June of 2013 their videos have been 720p 60fps... and they playback perfectly @ 60fps.

 

[HobeSoundDarryl]

It seems like Lostless found an objective way that didn't need to lean on a hardware frame grabber to verify the hypothesis. Good job Lostless! I also added the same to the other thread.

And way to go Billy B, for being perhaps first to notice this with enough conviction to actually write it down in a post.