Blaugust 2019 Wrapup, Part 2 – Streaming

Writing a blog post every day is not new for me, although the exact process I used this year was new. But livestreaming every day for a month is definitely new to me.

I have livestreamed before, but most of my experiments were back in 2015 before Twitch Culture became, ahem, that thing it is now. I also never made any attempt to impose any discipline on myself.

I was curious how I would find the process of streaming every day. Honestly it was my main experimental focus of the month. The blogging thing was just a side effect. Here’s what I learned.

Streaming every day at a specific time is very much like writing every day. It’s a habit that’s hard to start, but once it’s ingrained, it’s relatively easy to maintain. It starts to feel like a normal part of the day. It takes me about two weeks to establish a habit like that.

YouTube livestreams get considerably more views within the first day than an uploaded video. I’ve heard this anecdotally but now I’ve seen it with my own eyes. I guess for some reason there are a lot of people out there randomly browsing livestreams, but less people searching for specific videos.

YouTube re-encoded my livestreamed videos at considerably smaller bitrates than what I sent to them. I streamed at 5500K and YouTube turned them into ~3500K videos. They looked godawful compared to my regular uploads.

The biggest hurdle that I struggled with every day was balancing the game audio against the microphone. I added a limiter in OBS to smash the heck out of the game sound if it was too loud, but more modern games tend to have far higher dynamic range and ended up being too quiet. Since I played a different game every day, the mix was different every single day. The only way I could tell how to balance the audio at stream time was to visually examine the OBS audio meters and try to judge how loud it looked by how far the meters were jumping, and listen to how loud the game sounded in my ear buds. I had no way to monitor the outgoing audio mix in realtime.

The YouTube livestreaming dashboard is a giant steaming pile of junk. It was extremely difficult to monitor my own streams, and most days the best I could tell was only whether it was “on” or “off.” I had to reload the page constantly because the dashboard froze up. There were constant authentication problems, because it kept getting confused about which Google account to use. For some weird reason, LastPass kept thinking it had to update account information and popped up a window all the time. Everything about the experience was absolutely awful. Twitch’s dashboard is a million times better.

I dislike Streamlabs OBS. It’s full of ads, it’s gimmicky, and most importantly, you can’t visually see when you’ve muted an audio track. I have no idea why people use that thing, except for the pre-packaged overlays, which I don’t particularly want.

As for streaming in my future, it’s fairly clear that at least occasional livestreams will be beneficial in putting my content in front of more eyeballs and building the YouTube channel. It’s also about a thousand times easier to add content to the channel with a livestream. You just press a button and go, and the content is there instantly. The quality though … that’s something I’ll need to spend time pondering. I can’t abide the thought of antiquated 30 fps game videos on my channel.

As far as I know, there isn’t anyone out there clamoring to “hang out” with me while I’m gaming, so I feel no compelling need to target Twitch or Mixer specifically. But Twitch is where the audience is, and it’s far easier to work with its dashboard. I have some ideas about building a local RTMP server which would, among other things, allow me to broadcast to YouTube, Twitch, and Mixer simultaneously. I’d still ignore the Twitch and Mixer chat though hehe.

I would need to devise a method to record a local, high-quality, multi-channel video at the same time I’m streaming. I don’t like surrendering control of my videos completely to the cloud. With a local copy, I can re-edit and re-mix and re-encode the videos as I see fit in post-processing should the need arise. I already made a working process for this which I may write a post explaining later. (I experimented with it all weekend, but I still have some logistical problems to work out.)

I have a lot of ideas that need to percolate in my brain for a while before I decide what, if anything, I’ll do with them. The two key things I need to remember is that, without making any money from it, expenditure of effort at game-time needs to remain as close to zero as possible. The other main goal is to ensure that I make something that I’d also want to watch (and/or listen to-I’m weirdly obsessed with making my videos so that watching them is not required).