Hi all–a wonderful contribution: both the program and community support! Many thanks.
MY INTENT:
I’m trying to do a short lecture with my talking head in a pip on top of a program where I show what I’m talking about–my “slides” if anyone’s old enough to remember that term… ; )
MY PROCESS:
I’m recording the pip and slides as separate streams (with OBS). The catch is that my talking time is much shorter than the time to bring in all those slides–so I want to speed up the boring bits in the slide stream; the opposite of bullet time. And even before that speed-up, I need to edit out all the useless lags and mis-steps, so that stream is 50+ clips.
MY BLOCKING ISSUE:
In an ideal world I’d just highlight a bunch of clips (or better: a segment of time on that track), grab the right-hand edge, and squish it down to where the next slide matches what I’m talking about. The scores of little clips make it a serious pain to just increase the clip speed in Properties, and the chunks I want to speed up don’t align with clip boundaries. (I know I could split more…)
How close can we get to my ideal? Really: just the fewest steps–I don’t expect that simplicity. I’ll need to do this whole process dozens of times for a class I’m putting together.
WHAT I’VE TRIED:
I thought as a workaround I could just merge all the clips & re-split; I think there’s no clip merge? I’m in the middle of hacking the XML of my MLT save files because I did the edits in separate ones–hope to bring them together in one. (Also got caught by the can’t-add-bottom-track non-feature; you can see my empty “top” and “bottom” coping mechanism tracks in the image below. Bump on adding that feature–or just open with a top & bottom track to help out us newbies!)
ANOTHER POSSIBILITY:
Then I thought I could write them all out as one video stream, but I get terrible compression, even dead clear in this lowest-quality JPG:
Look at the Pool logo bottom right, and copyright text along the right edge.
(Also: why the black boundaries? I noticed the codecs added 8x6 pixels to my desired 1400x1050–a standard display resolution. Is this a video newbie issue that it’s impossible to get 1400x1050? Or could someone please tell me how?)
So someone could rescue me here by just telling me how to dump video at the same quality as the original–because I’m in the middle of the edit process and can’t afford to lose this much quality before the final encoding! (I’ve done hours of research here and know this is a desired but controversial subject.)
My mug benefits from the soft focus–but the text is cripplingly illegible…
Thought I’d done that (have been reading & trying more than a dozen lossless combinations by hand) and what I posted was the result.
Pardon the request for step-by-step instructions, but could you tell me what I should do besides clicking Ut Video or HuffYUV and export (which gives me 640x480)? Or doing that plus trying to duplicate my input Properties in the “advanced” tab (where I probably shouldn’t be but have been since 7:00 am today, and hours yesterday).
If the Stock Ut Video and HuffYUV give you this unexpectedly, then you composed your entire project at the wrong resolution and did not set Video Mode correctly. These presets do not impose a resolution. Only the presets with a “(video mode)” in their name or your custom presets impose a resolution. This is also the reason the image is not clear after export. It was rendered to 640x480 and then upscaled to 1400x1050. You should create a custom video mode, save it, reopen, and then review all filters that use position and resolution as they are not adjusted automatically. Alternatively, you can simply edit the <profile> element in the XML (profile = video mode).
Yes, unfortunately, this and many things in Shotcut and video editing in general are tedious and laborious - especially when you are new to the complex tools and still learning.
Perhaps explore more tools within OBS to help eliminate the need to edit so much. I’ve seen “Image Slide Show” in the sources for a long time, but until today never really messed with it. All of which can be controlled with OBS hotkeys, and/or a Streamdeck. There is even streamdeck type apps now available on Android and I believe IOS that integrate with OBS.
Thank you. I was underselling what I’m doing a little; a normal slideshow won’t really work. It’s not slides underneath, but a full running demo. (If you like, see http://54.144.78.101/research/#thePool for a previous demo; though what I’m preparing now is a class that overlays my face–I’m told that’s necessary–so I can teach people how to build such interfaces).
Some things I do while operating the demo are just too slow to match the lecture, so I have to speed them up. And occasionally too fast to match my talking. And I make mistakes I don’t want to waste peoples’ time with.
Likewise with my talking: I typically cut 3/4 or more out of an ex tempore topic discussion–I’m terrible writing/reading scripts and can only keep it flowing if I pretend I’m talking to someone. Then, like with the demo, I cut out the non-essential stuff so as not to waste a viewer’s time–and to focus my point.
So most of that source pre-editing is due to my process, not the tech. Then I need to do the stretching & shrinking of each to match the other.
–
But on a conceptual level your point is exactly right: I need to improve my process so it needs less editing. Need to get faster in the demo and have fewer pauses/tangents/re-dos as a talking head.
Very appreciated: I was just inching toward that realization, and this note helps. And I will poke more into OBS’s other capabilities; was only using it to do screen capture. That plus Shotcut make a remarkably powerful workbench–especially for the price!
