Playing recording in the program

[JohnnyP]

I'm a newbie and trying to figure this out. I want to record a voiceover, and then listen to it in the program, without having to use my computer's media player to do so. Is there a Preview or Play feature within the program somewhere? Help!

 

[koala]

OBS is the wrong tool for this. It's no interactive media player and no interactive audio recorder. Creating a voiceover for an existing video is a task for postprocessing software. Many video editors allow recording voice within the editor, while playing the video preview, so you can directly comment the video as it will appear how you cut it. Usually, they also offer the functionality to cut this audio into snippets and place it on the timeline like video snippets, mixing it with the video snippet audio.

You can also record audio with some audio recording software such as Audacity, which also allows you to edit/mix the voice audio with the video audio, however that's much manual work, and you need to play back the video in a media player, which you explicitly want to avoid.

 

[AaronD]

I think you want a video editor for all of this, which OBS is not...unless you want to do everything live all in one shot. "Live-to-tape", as we call it. That can be a valid way to do it, and when you've finished recording a take, it simply is what it is. Not much opportunity to clean up, which is also somewhat freeing for those that obsess over that sort of thing.

If you want to tweak and make it perfect, then you'll want to use a non-live editor. The way that *I* do it, with *my* workflow - your workflow has a good chance of being different, for a wide variety of reasons - is to:

1. Record the voiceover first, raw, with no processing whatsoever, using several mics including the camera mic. (I use Audacity for this, with two different XLR mics feeding a 2-channel 24-bit USB interface, and I export raw from Audacity into two 32-bit floating-point mono WAV files.) Recording raw allows you to undo any or all processing as you decide you don't like it, provided that you keep that raw version sacred and don't overwrite or delete it. If you only record a processed version, then you're stuck with whatever processing that is, like it or not.
   • Having made a few videos now, with the same processing for each of the later ones, I feel more comfortable recording the processed version to save a little bit of time in post, but if you don't have that experience yet, DON'T DO THAT!!!
   • If you're going to have some silence, without another soundtrack to replace it, don't just let it be silent, but "record the room", run it through the same processing, and use *that* as silence. It's amazing what a difference that makes. If you have another soundtrack, then yes, you can take the room out completely while that plays.
2. De-noise and audio process the separate mic recording(s) to your liking, but don't do any time-based editing yet. Keep the mistakes; fixing those comes later. This overall audio processing may or may not take a while, depending on your prior audio experience. I've done Front-of-House for a handful of coffee-shop concerts, youth bands, and older-adult church services, so it's not all that hard for me. My audio chain for the studio, live or post, is:
   1. DE-NOISE THE ROOM!!! Turn off the fridge, ceiling fans, HVAC, phone ringer, everything. The better you can record, the better the final result will be. There *is* a Noise Suppressor later on, but it's not magic.
   2. Highpass filter, 2nd order (12dB/oct), around 150-200Hz or so. I don't sing bass (you're welcome), so anything below that is noise anyway, and there tends to be a lot down there by default. Set the cutoff frequency as high as you can get away with, without actually hurting your voice.
   3. Noise Suppressor, with settings to not be noticeable if all you hear is the result. It *does* take away some of the original sound, so be careful with it. The better you can do with de-noising the room, the better the result will be here.
   4. Compressor, with a really soft knee and high ratio, to effectively make it an automatic ratio. So when I'm quiet, it's less noticeable, and when I'm loud, it really clamps down. I normally ride partway up the knee. This does most of the work of making a consistent volume with well-tamed peaks, that can then be set to use all of the available headroom without ever going over. It also amplifies any noise that got through to this point, so you really need to be good about having minimal noise to start with.
   5. Limiter. This is the final "safety net", that is only supposed to do something *on occasion*. The output of the compressor should be just brushing against where the limiter starts to work, but not normally hitting it. So if I do get loud, the notion of being "transparent" goes out the window and I'm NOT going to clip! And if I'm that loud, I kinda want a different sound anyway.
3. Once you have some good-sounding audio, load it into the video editor, and sync it to the camera's audio. Then mute the camera; it was only to sync the "real" sound, and with that job done, it's just a picture from here on out. Get everything on the timeline, all in sync, and save the project.
4. Still working with only the camera and the voiceover audio, clean up the flubbed takes, making sure that the picture and sound stay in sync. So you'll probably need to do the same edits to both at the same time, just to keep them lined up. The picture and sound transitions don't necessarily have to be at the *exact* same time or rate, just as long as the result doesn't look goofy. You might have to be creative with some of the transitions, taking advantage of some accidental similarity somewhere to make a quick crossfade become seamless.
5. Once you have a good voiceover-with-camera all the way through, save it, trim off the excess leader and trailer, and save again.
6. NOW, add the visual aids. Probably as images or video, exported or captured from PowerPoint, or however you get them in image or video format. Use additional editing tracks to overlay the visual aids on top of the camera.
7. When you like it, save again, and export the finished video file. Before you send it anywhere, play it with an independent player to see if you missed anything or need to compensate somehow.

As you can see, I spend a lot of effort to get the audio right, both the soundtrack itself and removing mistakes, and the picture just comes along for the ride during that part, to make sure it stays in sync. Once that's done, adding the visual aids barely takes anything by comparison.

---

If you do all of that live, first of all, you need to be pretty comfortable with how your audio needs to be processed, because you *are* recording the processed version only, and not the raw. If you're okay with that, then you can spend about as much effort in *pre*-production as you would in post, to load everything into OBS that you're going to need live, set up and test whatever automation you're going to need to pull it off, and run a few takes with *that*.

Record all of them, just in case you get a "one-take wonder", but each take is what it is. Not much chance to fix anything after-the-fact. If you don't like something, *keep going*, so you can find problems with the rest of it too, but you'll still have to do the whole thing over again.