Second time out with the gimbal setup over the bank holiday weekend, switched from 2.7K 60fps to 4K 30fps for this footage and the results are much better. Ideal conditions to capture the fantastic trails at BMCC, I really cannot rate them highly enough!
For those of you who saw my last video at FoD, I'd be interested to know if you can see the difference in quality in this video. I also managed to get the gimbal tracking straight for this footage...
[img]
[/img]
That was a fairly rapid decent! The gimbal looks to be working well.
That was a fairly rapid decent! The gimbal looks to be working well.
Thanks, there were a few bits that were quicker on other runs but this was the best top to bottom run overall, apart from the bit where I come up short on the double at the end of Moto! haha
But yes the gimbal worked great now I actually mounted it straight so the gyro could centre properly!
Case in your face .
BMCC is so damn good, we're back there soon, can't wait! Nicely ridden too, though ironically the gimbal's (IMO) gone a bit too far in the other direction, I'd have liked a bit of shoogle and lean in there- not a lot, just enough to connect it
I always get lost looking at film of bmcc, the trails seem to keep changing.
Can't wait to get back there, and good riding
Very nice 🙂 much better with the quality and straightness of the gimbal 😉
That high speed whine from knobblies when your tyres hit maximum velocity, the sound of summer! Can't believe how dry it is there!!
What do you like about the change in settings?
Case in your face .
I need a lawyer for that one right? haha
Nicely ridden too, though ironically the gimbal’s (IMO) gone a bit too far in the other direction, I’d have liked a bit of shoogle and lean in there- not a lot, just enough to connect it
Thanks, I did think that myself actually, I'll have to explore the gimbal settings (of which there are many) and see if there's a way to unlock the horizontal axis but limit the extent of it's movement somehow.
I always get lost looking at film of bmcc, the trails seem to keep changing.
That's the first time I've been so I was getting lost regardless, but you can't fault the variety! I like how all the runs start from only 4-5 different points so you can mix and match as you go down depending on how you feel.
That high speed whine from knobblies when your tyres hit maximum velocity, the sound of summer! Can’t believe how dry it is there!!
I'm super impressed with the sound quality of the Hero4, the addition of the nice clear tyre sound really makes the videos in my opinion compared to the mushy rumbles of my old Hero5/6! I was definitely surprised how dry it was too, every single trail was absolutely bone dry!
What do you like about the change in settings?
The main improvement for me was the increase in clarity of the 4K over the 2.7K (which was downsized to 1440p for YouTube anyway), especially when things get bumpy it helps reduce that mushy blur that can occur at lower resolutions.
I'm not able to do a direct comparison to 4K 60fps as the Hero4 can't film that fast, but I feel like the 30fps makes the video flow slightly better too, could just be my imagination though.
What POV setting you using?
Very nice, you rode that much, much faster than I managed to last summer.
What POV setting you using?
Just the standard 'Wide' setting, I've gone off Superview for this type of footage as it flattens things out too much.
Very nice, you rode that much, much faster than I managed to last summer.
Thanks man, these trails are really well suited to my riding style, which is basically flat out but flowy.
Just the standard ‘Wide’ setting, I’ve gone off Superview for this type of footage as it flattens things out too much
i've been using wide. never tried superview.
Nice. And you’re certainly not hanging about! Much greater sensation of speed with the chest mount, high res, gimbal and no music, just proper live sound.
Never heard of the place, but sure looks fun !
I too have never heard of the place, but it looks magic. Footage is fantastic too. I feel a road trip coming on 😀
Did you get any footage of the larger jumps (below) OP? - edit tried to make the video appear in the post but this new forum...it's...poop!
i’ve been using wide. never tried superview.
I found Superview works really well when I've been karting for example because the terrain is flat anyway, but for mountain biking you really do lose the sense of height/steepness in the footage.
Nice. And you’re certainly not hanging about! Much greater sensation of speed with the chest mount, high res, gimbal and no music, just proper live sound.
The sound was the reason I switched to the Hero4 from my Hero6. Seems to be fairly common among YouTubers to still be running the Hero4 as there just aren't any suitable external mics that will work with a gimbal on the Hero5/6 to remove the mushy sound. The sound of the tyres buzzing is one of the best indicators of speed for me!
Never heard of the place, but sure looks fun !
I've had a few people say they've never heard of BMCC which surprised me as it has some really fantastic trails!
Did you get any footage of the larger jumps (below) OP?
I didn't have the balls to attempt the 'Full Moto' line featured in that video on my first visit, but the jump section at the top of 'Tidy Darts' looks really fun and is halfway between my runs and the video you posted size-wise, so I'll try to get some footage of that next time!
Yer I had to use some custom code to add a hyperlinked photo in my post as the forum won't allow YouTube embedding, grrr!
The main improvement for me was the increase in clarity of the 4K over the 2.7K (which was downsized to 1440p for YouTube anyway), especially when things get bumpy it helps reduce that mushy blur that can occur at lower resolutions.
I guess this is just a youtube thing. They crush 1080p to death sometimes. You upload the same content to Vimeo and it looks amazing. They also seem to vary it by views/subscriptions/etc.
I guess this is just a youtube thing. They crush 1080p to death sometimes. You upload the same content to Vimeo and it looks amazing. They also seem to vary it by views/subscriptions/etc.
2.7K is just not a standard resolution, so YouTube will always downsize it to 1440p. So to avoid ropey YouTube compression I downsized that footage to 1440p during export/encoding myself.
YouTube does have a higher quality encoding option called VP9 which is only used for higher traffic/subscriber channels. I think Vimeo uses a higher quality encoder by default, but it has VERY restrictive upload limits.
For example as a free user I think you have a 500MB per month upload limit, the video I posted above with 3 minutes of 4K footage was 1.25GB alone...
The main improvement for me was the increase in clarity of the 4K over the 2.7K (which was downsized to 1440p for YouTube anyway), especially when things get bumpy it helps reduce that mushy blur that can occur at lower resolutions.
I’m not able to do a direct comparison to 4K 60fps as the Hero4 can’t film that fast, but I feel like the 30fps makes the video flow slightly better too, could just be my imagination though.
Without more detail what you're probably seeing is a mix between between interlaced and progressive scan and the lack of global shutter on Go-Pro's.
The problem with Go-Pro's is they have a rolling shutter, which means the camera reads each pixel then moves onto the next one, so from one end ot the other the image is ~1/framerate seconds delayed. Better cameras have a global shutter which takes an image of the whole sensor, then waits 1/framerate before taking the next. You can see this really badly in dashcam footage, it make the whole image wobble (the technical term is actually jello) as the engine shakes the camera at a similar frequency to the framerate.
60fps is presumably 4ki, 30fps will be 4kp. Which is better is subjective. An interesting effect is that films are 24fps, entertainment TV like Soaps are 25fps and the news is filmed at 50fps (we're not in America remember, everywhere else in the world films at 25 or 50fps not 30 or 60), which is enough of a difference that your brain subconsciously associates one with fiction and one with fact. So even if you did invest in a camera that could record a high frame rate with global shutter, it's not always ideal, a lot of DoP's still like 25p because it creates that slightly 'film' look even when they're shooting factual content.
For all the BMCC-not-knowers, it's just down the road from Bike Park Wales really and a similar distance from the Forest of Dean and they all offer really different days out. I really like BPW but after our visit to BMCC last year I just thought... Why would you ride BPW instead of this? The riding is a level above, and what it lacks in facilities it makes up for in tractors.
But those 3 so close together pretty much makes for the best long weekend's riding in the UK, imo. Not the best riding, but being able to uplift it and do 3 big days of fantastic riding is pretty much unrivalled.
YouTube does have a higher quality encoding option called VP9 which is only used for higher traffic/subscriber channels
if you upload at 1440p and higher it encodes the videos with vp9. I record mine at 1080p60 then interpolate to 1440p60 for upload.
Without more detail what you’re probably seeing is a mix between between interlaced and progressive scan and the lack of global shutter on Go-Pro’s.
I hadn't put too much thought into the science of the different filming modes beyond the framerate but I'd be interested to know if each of them is i or p now you've mentioned it, but none of the online specs seem to clarify the point.
I've noticed that jello effect on my dashcam before and thought it was purely just vibration, so it's interesting to know it's caused by a specific hardware type.
But those 3 so close together pretty much makes for the best long weekend’s riding in the UK, imo.
This thought did occur to me when I realised you have to practically drive past FoD to get to BMCC and that Merthyr Tydfil is only 25 miles further down the road. It would indeed make for a really fantastic 3 day uplift roadtrip!
Also your comment of "what it lacks in facilities it makes up for in tractors" is a perfect description of BMCC! haha
if you upload at 1440p and higher it encodes the videos with vp9. I record mine at 1080p60 then interpolate to 1440p60 for upload.
I thought they closed that loophole due to that exact reason, because people started encoding videos above their native resolution to trick the system into giving them access to the VP9 encoder? All the new YouTube upload guides now reference the fact that VP9 is based on a view/subscriber count.
Great video OP, got the angle of the camera just right. and I'm another who's only vaguely heard about the place, certainly on my list as well.
none of the online specs seem to clarify the point.
Just looking at the gop-pro site and it does imply that it's p60 to be used for slo-mo, (but in the same section states 1080p240 but only refers to 4k60 so maybe not).
There's also things like shuttering and motion blur to consider, if riding in bright sunlight you can stick some ND filter in front of the lens, this forces the camera to expose for longer (closer to 1/25 rather than shuttering down to 1/2500 or whatever it can achieve) which can reduce the shuttering effect between frames by blurring them together giving a better impression of speed.
IME anything faster than p25 or i50 makes the image look like someones set the TV up badly and turned the motion smoothing up too high which makes objects appear quite flat as the edges become too sharp (it's knows as the soap opera effect for the reasons I mentioned earlier).
Seems to be fairly common among YouTubers to still be running the Hero4 as there just aren’t any suitable external mics that will work with a gimbal on the Hero5/6 to remove the mushy sound.
The other reason is I've not yet seen any way to put timecode on a Hero 6, so if you have multiple cameras, or need to sync with a seperate sound recorder it becomes a faff to edit.
Great video OP, got the angle of the camera just right. and I’m another who’s only vaguely heard about the place, certainly on my list as well.
Thanks Nick, took a bit of trial and error at FoD (around 5 runs) to get the correct angle, but I was then able to rock up at BMCC flick it to the desired angle and film straight away with no fiddling so it was worth the research time!
The other reason is I’ve not yet seen any way to put timecode on a Hero 6, so if you have multiple cameras, or need to sync with a seperate sound recorder it becomes a faff to edit.
Generally accepted (but very basic) solution to that seems to be to just do a loud double clap at the beginning of the footage, then just line up the footage based on the sound spikes in the audio channel, but it's a bit fiddly and not very accurate.
That's interesting about the frame rates though as all the guys I subscribe to on YouTube like BKXC or Jordan Boostmaster seem to use a minimum of 30fps despite being American, Jordan especially seems to throw in a fair bit of 60fps footage too.
Generally accepted (but very basic) solution to that seems to be to just do a loud double clap at the beginning of the footage, then just line up the footage based on the sound spikes in the audio channel, but it’s a bit fiddly and not very accurate.
That’s interesting about the frame rates though as all the guys I subscribe to on YouTube like BKXC or Jordan Boostmaster seem to use a minimum of 30fps despite being American, Jordan especially seems to throw in a fair bit of 60fps footage too.
There's a similar method used on the motorbike vlogging community, with a beep of the horn used to sync cameras/external audio recorders. Not saying that's suitable but if you've got a bell... Ding!! 🙂
All the big US youtubers use maxiumum 30fps, 4k - BCPOV, loam ranger, BKXC, Singletrack sampler, Daily MTB Rider, Seths bike hacks all use either 24 or 30fps.
Jordon has done the odd 4k 60fps video but it looks so slow vs his normal stuff 🙂 ND filters are also getting a lot in biker vlogs, increases the motion blur as they're normally riding in bright sunshine, not dark forests so need to reduce the light coming onto the sensor to increase the shutter speed.
60fps is great for stuff you want to be crisp and clear, like gaming - but for stuff you want to have a more filmic qualty - 24/25/30fps is where to be.
There’s a similar method used on the motorbike vlogging community, with a beep of the horn used to sync cameras/external audio recorders. Not saying that’s suitable but if you’ve got a bell… Ding!!
I think its one of those things if you do it yourself it's fine but if someone else has to edit it then they make it an issue (especially if they have to do it on every clip and/or someone is paying to ingest it). If your software can read it you can also get away with just playing some LTC on your phone at the start of the clip, but some will only sync those frames not the whole clip.
Anyway........ That's getting geeky and a bit OTT!
That's all pretty handy info actually, even if it is getting a bit geeky. Because I may see if I can get hold of one of the small GoPro Session cameras and strap it to a fork leg facing backwards or something, just so I can cut between a couple of different angles.
I thought they closed that loophole due to that exact reason, because people started encoding videos above their native resolution to trick the system into giving them access to the VP9 encoder? All the new YouTube upload guides now reference the fact that VP9 is based on a view/subscriber count.
works on mine.
Seems to be fairly common among YouTubers to still be running the Hero4 as there just aren’t any suitable external mics that will work with a gimbal on the Hero5/6 to remove the mushy sound
I use a separate sound recorder. just do three claps and its piss easy to sync the sound. doesn't add that much time to the editing. I use a sony as30v with a Rode smart lav+ (need a 3.5mm TRRS to TRS adaptor - I've got a rode SC3)
All the big US youtubers use maxiumum 30fps, 4k – BCPOV, loam ranger, BKXC, Singletrack sampler, Daily MTB Rider, Seths bike hacks all use either 24 or 30fps
they can't do 4k60 though on their 4 black's though only the 6 black does.
works on mine.
Seems that you are correct about VP9 being auto applied to 1440p and 4K content now. It's just whether your older content gets re-encoded that depends on views which I confirmed by checking a 1080p video which I uploaded a year ago but now has 1500 view and it's been converted to VP9, while similar videos with less views from around that time are still in AVC.
they can’t do 4k60 though on their 4 black’s though only the 6 black does.
Indeed, which I thought long and hard about before my 'downgrade', but having uploaded similar videos a few days apart, one in 60fps and one in 30fps, I'm yet to find a single person that rates the 60fps footage higher than the 30fps at any resolution. So unless you want slow mo then it's no big deal.