Cineversity

What are the media encoder settings? When I change to MPEG4 it limits the resolution?

Hi ColourAndre,

https://support.google.com/youtube/answer/6178631?hl=en&ref_topic=2888648

I pasted the link for your convenience.

Perhaps try to render this directly from After Effects. I tried to reproduce your problem, with the little information I had, and in Cs6 for example, it seems that I can’t get even my UHD material from Ae as UHD output, same as MPEG 4 size is set in CC, here to 352x288 pixels, which is useless for VR.

My tip, go to Export>QuickTime, and then under preset > Match Source

Then go down to VIDEO, and there make certain you have the settings [e.g., Quality slider!] you need. You might set up several options in the render Queue, to optimize the balance between file-size and quality. Match source [in the Video tab] is not helping here.

As Rick mentioned, use the Listed option, so it is not public and check your material, your followers will appreciate your efforts.

All the best

I worked it out. Needed to render as H.264 and match source.
Silly mistake really, as I did know that I just wasn’t remembering.

Nice that it worked, ColourAndre.

All the best

Hi there,

Fantastic tutorials, these are making learning this process great fun and very informative, particularly with my limited VR and video creation knowledge!

My only problem is, I am following the Upload VR video to Youtube tutorial closely, yet my youtube videos are coming out very low resolution. I have some text items in my model which are un-readable.

I noticed that my render is not particularly high resolution when zooming in on the main objects, so have tried doubling the render size, which is slightly better but this still comes out quite pixellated. I’m currently rendering out a huge image (I think @ 13653 x 7680), which is 18hrs in and around 75% complete.

I get a feeling that perhaps this isn’t the problem? Are there best settings for after effects? I tried changing the bit rate (do I increase the maximum rate or target bit rate?), maybe I can render the video out in a different, higher quality format which youtube will render in better quality?

I have followed the Oculus cube map tutorial which works perfectly, but uploading these to youtube in the highest possible quality would be a great next step.

Thanks again for the great tutorials and for any help on this!

Hi Shaun Davies,

I couldn’t find anything on YouTube or Google/G+ that would confirm any option above 8K - well 8192x8129 and not 7680x4320. Typically UHD or a 5120x5120 was mention in the past, check YouTube support, if those values are of any interest.

https://support.google.com/youtube/answer/6178631?hl=en&ref_topic=2888648

What I think is perhaps a typo, is any resolution with an un-even number. Check your resolution.

Anyway, if you go with typical VR viewing options you might have problems to find anything that would even allow for a single eye resolution above HD, specking of final delivery and options given to the broad audience. If you go with on screen viewing, single image, you have to check which options you are using, there are MegaPano options available, where the resolution is typically also even in both directions. But you have to know your service, and their specific requirements.

I’ m not aware of any bit-rate number for a single image, as this couldn’t be set anyway, in any export option to my knowledge. YouTube will certainly recompress footage, and low res is the first version that shows up. It takes often a while before even an UHD video is available.

I uploaded on G+ an 8k image in mono and equirectangular, but the quality even then was unacceptable, like sphere without Phong Tag. After some back and fore with the support I left that option alone.

All the best

Thank you so much for getting back to me Dr Sassi!!
Sorry, looking back at my question, it maybe wasn’t very clear of what I was doing (  or it was clear and I’m more confused than I realised   ).

When I mention my render, I’m referring to my c4d render, which using the CV-VRCam with default “YouTube VR 2160s” settings & anti-aliasing set to best, was looking low quality. Where as when I doubled the render size it was looking marginally better in quality, and now, increasing once more we’re getting somewhere, but still miles off.

Here is the original rendered version, to the specs in the tutorial:
https://www.youtube.com/watch?v=rty0VGRKg9s
(I rendered the logo etc separately and faded these in via after effects).

And here is a version with the increased 13653 x 7680 render size (I increased the height value, keeping within 16:9 aspect):
https://www.youtube.com/watch?v=-JK0n2Kn28U
As the render wasn’t quite finished I copied the incomplete render and opened it in photoshop, then copied the old render and dropped it into another layer underneath and resized it to fit to complete the image, so there is a line over the arch and a section behind you which indicates the change in quality.

I am still composing the image into a video file in After Effects, using the settings there as per the tutorial and uploading to youtube with the meta injection.

It just seems like a huge amount of resolution is lost in between the initial c4d render and uploading this to youtube, where as with YT you can upload 4k quality video, and other 3D examples I have seen have been far superior quality. For example, this one I found on youtube (https://www.youtube.com/watch?v=HNOT_feL27Y&t=26s) seems to be a 3d animation, uploaded in 4K, but seems to be far superior in video quality. I understand these guys really know what they are doing in comparison to my early stages of knowledge, but perhaps they had to encode and upload their video in much the same way?

I have a 4K monitor and it really shows a difference when shown on screen.

In the tutorial Rick mentions that his render wasn’t of the highest quality, and you could increase the bit rate in Media Encoder to improve quality,so I wondered if he was using simple settings for quick render speeds and ease of use, or wether there would be optimum setting for both cinema 4d and after effects to get the best quality possible before it goes to youtube.

Thanks Dr Sassi, apologies if my lack of knowledge is making my questions confusing or misguided

Hi Shaun Davies,

Edit: Have a look here
https://blog.frame.io/2017/02/15/choose-the-right-codec/
/edit

Since I’m a RED shooter I visit often reduser.net, and 8K was since a while a main theme there, so here is an example with 8K. Even I get a error currently, but you can see that the resolution options are given in the cogwheel-icon settings options. Not that I’m aware of any 8K VR from YouTube yet. But perhaps one day. If so the format would be something else:
My hint was a little nudge to your numbers, OK, I will be clearer, 13653x7680 is not a number that I would use. First it seems not to follow any dividable and secondly there is a 3 as last digit, a clear problem. The hight seems pure fantasy, perhaps I miss something, since this is a fast changing subject.
https://youtu.be/sLprVF6d7Ug
The producer/Cinematographer is Luke Neumann

Anyway, if 4K and then it is UHD, which is 3840x2160 and that will be recompressed. I use normally the highest lose less compression I can get uploaded in time. You have to make tests. Not all compression methods are equally good, some might be nice for some footage, then it might suck (really) for something else. Again, if YouTube gives you 4K and thing larger is a waste of time. If they have to sample it down, it might be not as good as After Effects could do. Yes, quality gets better when sampling down, but only if one knows what and how to do it. YouTube is not equipped with a human controlled system, it is certainly all automated.

Since the footage is for motion, but without motion blur, a high AntiAliasing creates less effective compressions. If you work with data rates, those crisp details will take space in the file. So, perhaps reconsider what really needs to be super sharp.

The discussions what works and what not will differ, and I will not even pretend to get it done the first two times in most cases. Trial and error. Learn the compressions-philosophies, e.g., is it run-length or mpeg based, etc. Form there you might have a first evaluation. Or just render different versions out. But never with a loss. Youtube can’t repair anything that is loss.

You see it on a 4K monitor, but the key questions is, how is it finally seen. My iMac has a 5K and when I look at any Facebook image I think of crumble cake, as they place a patter over it, grain makes an appearance of sharpness, to me it looks horrible, compared only to G+ which is also not the best.

As a subjective note: I tend to have my images (practical/still) in 26+Kx13+K/16bit Adobe RGB at the moment and my footage (practical/motion) in capture 6K/16bit down-sampled to deliver 4K/UHD, and it shows (camera UHD-movie stuff is more 3K, if at all, but we all know that), most people will see it on an iPhone - at least so far I get from my “followers at G+” (close to 300M views currently), and there goes the whole beauty of hi-res. But I digress.

Well rendering, if done well is better in quality, and there we need to soften sometimes areas of little interest to get the size (MB-wise) down.

All of that, just suggestions. Keep for now the the VR-Cam render sizes and render more—upload even more —than getting slowed down by 15K stuff. IF you get more savvy with the compression and which image content suffers less from which compression format, you might explore larger sizes. At the moment the bandwidth is certainly not ready for that, nor the majority of devices. Well, that is my idea about. Toss it or explore it, that is just up to you.

Besides, test different frame rates:
https://support.google.com/youtube/answer/1722171
The more frames, so far I got it, the better the quality. But I might be wrong, as things have changed constantly, as ai Said above already. Perhaps 60fps is better than 24fps. Test, test, test. Testing means to use all kind of devices, as nothing is more annoying than using a device that is overwhelmed with the data flow, and your head movement is not followed.

My assumptions:
The less compressed, the higher the bandwidth needs to be
The higher compressed, the better the decoder must be (faster)
The higher the frame rate the more natural motion will show

In anyway, I hope you get your work showcased and find people who admiring it. Isn’t that what it is all about?

ENJOY.

Dr. Sassi

I do not found a video that show me how to put together different scene files into a video to show on any Source available (youtube, tv, etc).

If we have multiple scene files for a short animation; how can this be converted into a video.

Best Regards

Yolanda Guerrero-Rebollo

Hi Yolanda Guerrero-Rebollo,

Different scene files can be merged with the Merge function or with XRef. However, I believe this is not your question, but I wanted to share it just in case.

To render out the scene, you can use the camera, set to spherical. Typically the suggestion is UHD (16:9), but as Google/YouTube will tell you, they recognize many more formats.

Google suggested in their Help to use these tools (no guarantee from us about these tools)
https://github.com/google/spatial-media/releases/tag/v2.0

If that is an animation for each clip, please be aware that editing for 360º/VR is technically the same process as standard film/video editing. However, since it will immerse the audience into the space much more intensively, be careful with the edit. It is easy to give people motion sickness, and that will ruin everything.

Find the most attractive point and match that in the edit. IF the clip is in motion, try to match that motion.

I hope that will answer your question.

Cheers