Making Podcast Transcriptions with Pixel 7

Production on Chapter Select Season 5 – Resident Evil is in full swing. A new element you will find with each episode is a complete* transcription. I’ve wanted to provide this for some time. Now, with the Pixel 7, I am taking the plunge.

A major part of why I bought the phone was for the live transcription via the Recorder app. I wrote about my plans back when I got the Pixel 7:

I tested this feature out just this morning by playing a few minutes of Ben Thompson’s interview with Mark Zukerberg and Satya Nadella about partnering in the Metaverse. The transcription is almost realtime. With decent grammar. When Google releases the update to their recorder app to add speaker labels, I plan to use this to make passable transcripts of Chapter Select and The Max Frequency Podcast. It may not be that perfect transcription, but the out-of-the-box accuracy and ease of use makes offering this type of resource affordable to me. It is magic.

I had been holding out for the Speaker Labels update (which I also wrote about since I tried to sideload the update).1 With the update out in the world, I started figuring out the best way to turn the show into a transcript.

There’s no import feature in the Recorder app, so the bare bones solution is just to playback the episode out loud for the Pixel to hear and record. The obvious catch with this method is that no one can make noise while recording is happening. Not that the transcription is 100% accurate to begin with, but additional dialogue would throw the entire recording off. I needed a way to turn audio output in to a microphone input for the Pixel 7.

The first required piece of hardware was a USB-C to AUX dongle for the Pixel. Pretty sure I remember Google taking a dig at Apple for dropping the headphone jack, but I guess that doesn’t matter for Android users now that Google has their own wireless earbuds. I snagged this Anker USB-C to AUX braided cable. It’s a bit steep at $20, but I use a ton of Anker products and trust the build quality.

I have a few high quality wireless microphones at my disposal: I’ve got a Sennheiser AVX-MKE2 and a Rode Wireless Go. I tried taking those fusions of input and output into the Pixel, but made no progress. The only time a microphone was recognized as such on the Pixel was when I plugged in an XLR input adapter that came with the Sennheiser. The catch was I didn’t have a way to connect the sound to the input receiver.

It was a snake’s nest of cables strewn across my desk as I tried making a connection. By far my wackiest chain was using the Pixel’s included USB-C male to USB-A female dongle to a Lightning cable with the outer housing removed so it could be plugged into an Apple Pencil gen 1 adapter to the Apple AUX dongle to an AUX cable plugged into my iMac’s headphone jack. I assume the Apple Pencil adapter doesn’t transmit sound/data. 😅

I was over-engineering my solution. Turns out I had all I needed right in my PS5 capture set-up. Before God of War: Ragnarök came out, I bought Elgato’s Chat Link Pro. I wanted to record all of Ragnarök, but with a baby in the house, I also wanted to be able to use headphones. When you plug in a headset to the PS5, all sound is routed to the controller. This Chat Link Pro is a glorified AUX cable and headphone splitter. It worked as advertised and I could play with my headphones on, while capturing all the sound.

The other key feature of this cable is in the title—Chat. It can also capture your microphone and party chat. Remembering this, I took a guess that the Pixel 7 would recognize the cable as a microphone. I was right.

The only thing missing was another AUX cable to connect to the phone, since the Chat Link Pro’s mic-in side was being plugged into my iMac. If this beefy braided cable was meant to capture party chat, why not use the wired connector that came with my PS5 Pulse headset. That would be a mic too. Boom bada bing, I had an external mic input to the Recorder app that was transcribing my iMac’s audio output. All it took was one dongle and two cables. So simple.

When Season 5 kicks off here soon, you’ll see a link to grab a PDF transcript of the entire episode. Turns out that the speaker labels don’t migrate to the Google Doc, which is a huge bummer. Fix that Google. But whatever Google thinks we said is laid out for you to read.

I’m aware that this is a far cry from a human-proofed transcription, but I feel like this is better than nothing. And sort of like my attempts to complicate this capture solution, sometimes the simplest option can work out best. This method is free for me to use, which is vital at our current scope, and it gives listeners (and now readers) an option for assistance/enrichment. If these transcript help just one person better enjoy the show, then it will have been worth the time and effort.

1. Turns out that the sideloaded version stopped working, so I just had to wait for the official Play Store to update anyway. Thanks Google!

Leave a Reply