Creating audio/video demo/teaching files

[Folderol]

This seems to cover almost all SOS topics so i really don't know where to place it

For a long time I've been wanting to create Youtube vids. for Yoshimi, starting with the basics and working up from there. Only now do I have nearly everything required to achieve this.

The only hardware issue I have is that the M-Track 8 mic inputs don't have enough gain and are comparatively noisy. I'm guessing the best answer would be to put my hand in my pocket, get a decent mic preamp and plug it into one of the line-level sockets.

I have the mic on a stand so it's fairly close, but to one side to eliminate plosives, and at the same time give me a clear view of what I'm doing and of the keyboards. I'll be using a small cheap computer keyboard that has virtually no 'click'. Like that, the room characteristics are noticeable but not intrusive.

Making the actual presentation is the biggest issue. I've done a few live talks at LAC events, and tend to stumble when trying to talk and operate mouse/keyboard at the same time - even with a prompt sheet. I could really do with help sorting that out!

[James Perrett]

Looking at the specs of the M-Track 8 it should be pretty quiet. The EIN is a respectable -126dBu unweighted which is only a 2-3 dB higher than the quietest preamps. There are plenty of preamps with worse noise specs.

What mic are you using? Could it be the mic that is generating the noise?

[Wonks]

As for fumbling the words, can’t your actions in the video follow the prompt sheet and you add the words later? You could then position the mic nearer your mouth to hopefully get around any gain issues?

[muzines]

Making the actual presentation is the biggest issue. I've done a few live talks at LAC events, and tend to stumble when trying to talk and operate mouse/keyboard at the same time - even with a prompt sheet. I could really do with help sorting that out!

Editing is your friend, here!

Honestly, the best way imo to figure this out imo is to make one, start to end. Even if you don't ever publish it, go through the process, and figure out the preferred workflow and problems along the way, look at the results, and assess what worked, what didn't, what could be better etc.

Get a few under your belt, you'll have learned how to do it, and be in better shape to start showing them to the world...

[Sam Spoons]

Are you aiming to be on camera or an off camera narrator with the camera concentration on the hardware?

If the latter could you, having written the script, film the visuals first then dub the narration afterwards?

[ef37a]

Hi Will, I am also with the idea of not doing this in real time. My son watches a lot of YT stuff, mainly science vids and the odd guitar/amp related things. Most of them are painful to watch because the presenter waffles for 15mins before showing anything interesting (yes! But what does the fekker SOUND like!?)

I think your noise problem can be most easily fixed with a FetHead.

Love to see some.

Dave.

[MarkOne]

As has been said, the art is in the edit.

If you are narrating off screen, you can pretty much just read your script (Yes, write a script don't try and wing it unless you are really good at this stuff), but you do need to concentrate on making it sound natural.

If you are doing it to camera, you can get one of these to turn a tablet or iPad into a teleprompt, which avoids the unnatural looking off to one side or down to refer to your notes.

Either way, unless you are a real VO expert the chances are very high you won't get it all down in a take and I would say don't even try. If you stumble, flub etc, just pause, restart at a point just before the error and go again. You may well end up with twice as much footage as the actual final video, but you can loose all the bad lines on the virtual cutting room floor.

Filming with two cameras, and cutting between them or dropping in a graphic to mask the cut looks more natural than a weird jump cut.

Even the well experienced YouTubers do this, some often mention how long the episode took to film sometimes. I watch the astrophysics channel Dr Becky and she makes a feature out of her bloopers at the end of each video (usually including the line 'Science is Hard, Words are Harder' and a jaunty little song at the end)

[muzines]

Basically - when it looks natural, it either isn't, or you're probably missing the amount of skill the presenter has to make it look that way. (Especially on live TV!)

[MarkyC]

Hi there,

I do a lot of these for work and always record vocals separate to the actual computer screen.

For me I create a kind of story board, just on paper so I know what I’m doing and have focus. As said before people just faffing on is no one’s friend.

Once I have the area to be covered down, I then usually write some sort of script, (This actually takes a while, but is really important)

Then record the screen with the script next to me, as many takes as it needs to get a nice flow, then edit.

After that use the “script” to do the vocals.

My one bit of advice is to try and make them short, you would not believe how easy what you think is a 5 minute talk ends up being 15 mins that no one has the time to watch…….

[Wonks]

Why not start with making a 1 minute test video? You could try recording it directly and doing the sound and vision separately and compare the effort and end results and see what you like best.

I’d also consider breaking any longer video down into sections, especially if doing it all live, so your maximum take length is say 30 seconds. Far easier to get 20 good 30 second takes to edit together than one good 10 minute take.

And to make any progress, you’ll have to acknowledge that you’ll still be making a few ‘ers’ and ‘ums’, at least to start with. They generally don’t get noticed unless there are excessive numbers of them.

[Drew Stephenson]

Loads of good advice above, two additional points from me:
1)

And to make any progress, you’ll have to acknowledge that you’ll still be making a few ‘ers’ and ‘ums’, at least to start with. They generally don’t get noticed unless there are excessive numbers of them.

If you're speaking on camera we are really tolerant of umms and ahhs and hesitations etc when we can see the body language as well. As Wonks says, it doesn't get noticed.
If it's a voice-over though, with no visible speaker, we're much less tolerant of these things. So (as mentioned above) you'll need a script, you'll need to practice it, and you'll need to edit it. If you fluff something, pause, take a breath, and start again.

2) The only transitions you need are fade to black and cross-fade!

[Wonks]

No calm-fade or happy-fade?

[Folderol]

Thanks a lot everyone.
There seem to be a lot of common/similar suggestions (always a good thing).
Obvious as it now seems, I hadn't thought of creating a script - DOH!
Recording the video and audio separately would definitely help too, especially from the point of view of cleaning up the audio before overdubbing the video.
I'm not intending to be in shot, so that make synchronisation a lot less critical.

I'll try a short test piece.

[Drew Stephenson]

If you're going to completely replace the audio then it can be useful to talk through what you're doing on screen as you record it. This can significantly speed up navigating around the project when it comes to editing.

[Folderol]

At the beginning there will be no audio from Yoshimi, but that will change fairly quickly - although I might have a brief sample of it running full fat at the very start.
Besides I've now got a fairly clear idea of how to manage this.
I often took part in school plays (yes that long ago). I'll make out scripts in the same style - speech in normal text, actions in italics. If I use that for both the video and audio it should all fit together... in theory

[BWC]

If the latter could you, having written the script, film the visuals first then dub the narration afterwards?

Or, record the audio, then do the visuals along with audio playback to guide your hands?

[N i g e l]

Things I hate on technical utubes:

1. umms & errs [at least mute them on the audio.
2. infomatic corrections after an error (i said cut the "red wire", shoulda been "blue wire") oops !
3. rambling on without a plan/script/story board.
4. Intro is life history followed by history of item. I just want to learn how to put a nut on a bolt; not meet your family or learn about steel making from the industrial revolution onwards.

Bear in mind that uTube offers close captioning in any language [ I am asuming thats done on the fly because it does come up with some howlers, especially for technical words ]
Your video can be international and accessable to those that are hard of herring.

When I actually do a utube video [getting closer] it will be bop bop bop, thank you good night. There might be an intro summary which gives ALL the details & conclusion.
It wont be done as a live performance but using all the assistance audio & video editors have to offer .

I would be striving for perfection as I wouldnt be doing that many.
If your doing 1 a day, its a different kettle of fish.

Thats just one style though, others are available.

[Folderol]

If the latter could you, having written the script, film the visuals first then dub the narration afterwards?

Or, record the audio, then do the visuals along with audio playback to guide your hands?

That's sort-of what I had in mind. record the audio, then play it back stop-start while twiddling the controls then adjust the audio spacing so I have time to do the necessary twiddling without being in a rush.

Write the the control actions into the script lined up with the text.

Once the timing is right, record the video. if I'm confident enough I could overdub the spoken audio track at the same time while Yoshimi is also generating sound - with the volume levels already matched.

[MarkyC]

I know this may sounds silly but one of the things I’ve learnt that really helps is don’t move your mouse to much when recording a screen. OK this may sounds odd but it can really help.

Try and make the mouse movements as smooth as possible, then once you have clicked on the thing you want to show, just let the mouse site there for a few seconds, then move on to the next bit.

What this means is that when you are editing stuff together if your vocal goes over you can just extend the frozen screen. It also gives you options to zoom in/highlight stuff, or if you don't require it just cut it out!

[tea for two]

Just couple of things.

Already mentioned on a few posts : scripted and not real time.
Utube tech reviewer channels with 15million+ subscribers, 2million+ views per video : their stuff is mostly partially scripted, the filming edited.
Yoshimi should be the main focus not your dear self : makes it far easier to read off a script edit the video.

USP first : what's the Uniqueness (Unique Selling Point) of what you are showing getting go this straightaway.
These top utube reviewer channels don't get to the UPS first as they are monetising the video by drawing it ooot. They can get away with it as they have millions of subscribers.

::

Most of them are painful to watch because the presenter waffles for 15mins before showing anything interesting (yes! But what does the fekker SOUND like!?)

This drives me up the wall.
Get to the USP already !

[Matt Houghton]

If you do it in real time treat it as a performance. Storyboard it. Script it. Rehearse. Record. Don't be afraid to do retakes or drop in scenes.

Recording very short sections/scenes like this and stitching them together in post is a good bet.

Bear in mind you can talk over still images too... a few screengrabs of salient info can be really handy to splice in to mask clumsy transitions.

You'll get a slicker video doing it the other way around recording either script or video first, and then syncing the other to it. Doing it that way for say software tutorials, script first video second often works better. Because you keep your own voice and pace. But you need a tight script to do that well. OTOH, if you expect viewers to follow your actions in real time, consider video first, and do it all slowly and deliberately. Then the narration can make clear what you're doing.

If you've not done this before, I'd urge you to set aside a dah or two to figure out what works for you.

[Folderol]

Thanks again for all these suggestions. I've made a few short trial attempts (mostly to checkout the equipment) and started to work out a script. As I said, I won't be in shot for any of this, it will all be direct screen grabs.

I'm also working out how much of an introduction I need for just the first of (hopefully) a series - both of me, and of Yoshimi

It will be a while before I can get anything released. I'm still sorting out the 'music' room, and also working on the upcoming release of Yoshimi itself.

[Drew Stephenson]

I'm also working out how much of an introduction I need for just the first of (hopefully) a series - both of me, and of Yoshimi

If it's a series of videos then you can set them up as a playlist on Youtube, this means you can pull your introduction into a short, separate video so that people who don't need that can jump straight to the content they're after.

[Folderol]

I'm also working out how much of an introduction I need for just the first of (hopefully) a series - both of me, and of Yoshimi

If it's a series of videos then you can set them up as a playlist on Youtube, this means you can pull your introduction into a short, separate video so that people who don't need that can jump straight to the content they're after.

How do you set up a playlist there? Do you have to do it at the start or can you do it once you have one or two vids up there?