[Aaron Star]

Your sample video has been removed by the user, so it is hard to see what you are getting.

Also you do not list anything about your workflow from camera/source thru to render. Listing your project settings and “media info” on your edit source media would be helpful to the group as well.

Rendering for YouTube is not that hard, you could try rendering to XDCAM-EX.mxf which is a variable bitrate codec with set parameters. YouTube will convert this media for you automatically.

[Aaron Star]

I like the keyframing in this example. Did you find this somewhere or was there a way to automate the key frame generation?

[Aaron Star]

These sort of issues seem encoder related. If you are not adverse to CLI apps, get a copy of FFMBC and try converting one of your 20min clips using the following:

ffmbc -i “input file.mp4” -target xdcamhd422 -an “output file name.mxf” -acodec pcm_s24le -ar 48000 -newaudio -acodec pcm_s24le -ar 48000 -newaudio -map_audio_channel 0:1:0:0:1:0 -map_audio_channel 0:1:1:0:2:0

This will convert your MP4 source clips to XDCAM422-50mbs with PCM audio. During the encode FFMPEG will report issues internal to the file. If there are no error in the streams, it will just show a FPS conversion counter. XDCAM will edit much smoother in Vegas, and with less stability issues.

FFMBC is basically just a version of FFMPEG that has templates/targets for converting to broadcast codec formats.

[Aaron Star]

I downloaded the .TS file and had the same issue as the rest of you, that being Vegas seeing the HP recording.TS as 100+FPS.

FFPLAY played the file back just fine, but I was having a hard time getting the file to re-encode via FFMPEG to and .MOV, AVI or .MXF that would play in Vegas.

Using FFMBC and converting the .TS file to XDCAMHD422 worked, and the file loaded into Vegas as normal. FFMBC can be googled and downloaded for free. The Command is as follows:

ffmbc -i “input filename”.ts -target xdcamhd422 -an “output filename”.mxf -acodec pcm_s24le -ar 48000 -newaudio -acodec pcm_s24le -ar 48000 -newaudio -map_audio_channel 0:1:0:0:1:0 -map_audio_channel 0:1:1:0:2:0

Using FFMPEG I was able to convert to ProRes -Profile 0 (Proxy) which also loaded in to Vegas normally, but with a lot of files you will hit the quicktime limits. That command is:

ffmpeg -i “input filename”.ts -vcodec prores -profile 0 -acodec pcm_s16le -ar 48000 -ac 2 “output filename”.mov

[Aaron Star]

Digital Clock font and frames for each change?

There is a plugin for Timecode, and if set on the preview window, you can then change the timecode of the timeline to what ever you want it to display. For a countdown run the rendered clip in reverse. Use a cover mask to clip the numbers you do not want to see.

A still image of what you would like to see and still of what you have so far would be helpful.

[Aaron Star]

The adobe blog post says that AE5 had an opencl supported raytracer, which was swapped out for a Cuda in AE6 due to stability issues. It does not mean AE will not work AMD, you just will not get the bump with that feature. You would have to know how much of the 3d text and particles you use in your workflow.

[Aaron Star]

The difference I see in Skylake(z170) vs x99 are:

x99 has 4 channel memory using the same memory as z170. This means greater through put. See the block diagram difference.

X99 has more PCIe lanes than Skylake for things like, an SSD intel 750 operating in NVMe vs AHCI. NVMe is the future of things. Verify the benchmark configurations of the tests you are referring to. https://www.tomshardware.com/reviews/samsung-sm951-nvme-versus-ahci-sata,4137.html

X99 supports the 5960X which is 8 cores and not 4, 16 threads vs 8. This makes it closer to a Xeon class machine.

Skylake is fine too due to the lower costs.

RAID 10 will work, or spreading you media out across multiple Raid1 configurations would work too. If you use quality SSD, they will keep up with bandwidth needed for 100MB/s sustained. Failed raid configurations can lose more data than Raid 1 which simple mirrors media. A raid 1 drive can be pulled and installed stand alone into another machine and be retrieved, should the controller fail.

I cannot speak to the Ninja Assassin’s metadata tagging.

[Aaron Star]

There is this video on enabling ray tracing in AE for AMD: https://www.youtube.com/watch?v=C4bBEvUswMU

It does look like the video above may be dated per this post:
https://blogs.adobe.com/aftereffects/2012/05/gpu-cuda-opengl-features-in-after-effects-cs6.html

I could not find the menus in the driver I am using on my AMD card however.

You could just do your own testing, get one, and return it if AE performance sucks.

[Aaron Star]

What is your primary camera, and resolution you plan on achieving for display/distribution?

720p will upres to 1080P just fine. The footage will have softer look to it. You could your friend shoot a few short test shots and upload it to dropbox, and test how the footage will compare. You could also do the News/CNN thing of creating graphic surround and inserting the 720p PiP scaled to 100% or smaller.

A d3200 at Target or Best Buy is only like $500 out the door, and the D5100 on ebay is <$300. The firmware for the D5100 and possibly the D3200 is free. That way you can get a rough 35mm look in a 1080-30P file. The firmware allows bit rate up to 36-54Mbs, but you will need a good SDcard like a SANDISK Extreme Plus, so it can keep up with the camera.

[Aaron Star]

I like that scripting for batch reasons. I had not really considered the audio half of the problem. Good Call. That is similar to converting AC-3 in MTS. I will have to look up a few of the other parameters you put in that command.

You should post that over on the Blackmagic Resolve forum, under the MTS thread. Those people are crying like babies over not having MTS support in Resolve, and most are being directed to a $50 GUI for FFMPEG. Peoples hatred of CLI is amazing.