Over many years I've found great advice here, but never posted. I'm having some trouble determine a tools/workflow combination I'm happy with for a specific purpose so I thought I'd start a thread. I have some decent experience with various encoders, decoders and media tools over a couple of decades. This isn't, however, a key hobby or profession of mine and I can easily become overwhelmed with choices or be unfamiliar with recent developments or options.
My church is looking to have a Youtube channel with an archive of sermons (the key weekly message) available to view. We have a decent consumer camcorder capturing the video every week.
1/ The source video is 1080p H.264 60fps around 35Mbps with stereo audio fed from our sound console.
2/ The camera is not manned, but left with fixed aim and records all through the service. Someone must trim out the central message in post. There are copyright and privacy issues with publishing the entire service.
3/ The camera writes in 4GB chunks, and 3-5 files must be concatenated before the main message (~20mins) is snipped from the middle
4/ With the particular (consumer grade) camera and typical lighting conditions, there's a fair bit of video noise I wish to remove prior to re-encoding and publishing
5/ Another member has produced a 30 second "intro" video clip, and a 20s "credits" video clip I can add to each week's file.
Prior to having the intro/credits from #5 available to me, I had fooled around with a few tools and settled on AVIDEMUX. My typical post processing now:
1/ Drag and drop the 3-5 files from the camera's SD card into AVIDEMUX
2/ Visually set A/B markers to clip out the sermon message
3/ Use the Mplayer denoise 3D HQ filter with 2,2,10,15 settings (more temporal noise reduction than spatial)
4/ Sometimes use the Mplayer eq2 filter if there's a gamma or brightness problem
5/ Save the clip using x264: slower, high level, IDC auto, CRF 20 AAC-LC: 96kbps MP4v2 muxer into MP4 container
While not amazing, they looked much better and compressed far better than out of the camera. I used a semi-graphical tool to do everything I needed to do (concat, clip, denoise, recompress and container). This has left me with a folder of clipped, denoised MP4 files but lacking intro or credits.
The catch is the intro/credits video segments which have now been delivered to me. Since this is a weekly task, I'd like something efficient. Avidemux fades where you have to type in the in/out times is not efficient. I tested and I am unable to simply select the three files and tell it to copy the streams without recompressing - the output is unplayable (I'm guessing because the H.264 parameters don't match).
Suggestions on either a new workflow? new tool I haven't considered? way to not have to reencode everything from scratch? way to glue intro and credits on to existing files with a simple fade and, maybe, re-encode only the fades themselves?
Everything I look at has some kind of "fatal flaw" compared to AVIDemux. Most of the tools are way to complex for anyone other than me to do the weekly work unfortunately, but that would be nice. I feel like I need to be a little OCD about this because of the video noise and not just use any old "came with the OS" NLE that doesn't denoise or let you tweak encoder output settings.
Thanks in advance,
Our website is made possible by displaying online advertisements to our visitors. Consider supporting us by disable your adblocker or try DVDFab DRM and remove iTunes movie & music protection! :)
+ Reply to Thread
Results 1 to 3 of 3
Add more lights.
Fix the problem, don't waste hours every week patching it.
Agree. Also, you can do the multiclip concat easily and repetitatively with ffmpeg script (maybe using a watch folder).
Then, manually do the sermon-length trim and export (using I-frame boundaries so lossless stage). Avidemux is good for that.
Then, get your friend to change their setup so they encode their intro, outro clips to the EXACT specifications matching what the camera puts out.
Then, use ffmpeg again to do the final concat.
This will keep things all 1st generation, with the exception of if you add x-fade transitions (and then you should do extra trim stages so you only re-encode the transitions, concatenating 5 instead of 3 clips).
As someone who spent years doing, among other things, TENS OF THOUSANDS of customized nationwide State Farm intro & outro bumpers & commercials (both radio & tv), I can say with some authority that you need to make the computer do what it does best and automate ALL but the important parts which require your intervention (trim choices, intro/outro credits data & artwork). Make the system work for you, not you the system.
Last edited by Cornucopia; 15th Apr 2018 at 23:36.