Upgrading Podcast Player and Audio Notes App in My Personal Knowledge Management System (PKMS)
In this article, I describe one piece of my Personal Knowledge Management System (PKMS) Fall 2022 upgrade — potential changes to my podcast player and audio notes applications. For my use case, these applications are intertwined and so I’ll cover both in this article.
Background
Previously, I used Apple Podcasts and QuickVoice applications. Refer to my August 2022 Medium article for further details of my PKM System configuration prior to beginning this upgrade cycle.
Upgrade Goals
My goal is to address two shortcomings in my prior system:
- Having no automated way to capture podcast highlights, and
- Using my own first brain to manually transcribe and organize my audio notes.
Motivation
I like to tersely summarize and Tweet my takeaways from podcast episodes that I find impactful. This helps me retain learning and encourages others to also learn and perhaps listen to the same episode. Having podcast excerpts in my Obsidian Vault is a bonus so I can use them for linked references or rediscovery via search months later.
I take audio notes whenever I am away from my desk. Then, at a minimum, during my evening shutdown and morning startup, I listen to any new voice recordings and type the takeaways into notes or my task list. Wouldn’t it be cool if I instead dragged and dropped a text transcript into the appropriate place?
Context
I am a long-distance trail runner and I almost exclusively listen to podcasts while I run. The podcast content and the physical act of running combine to predictably set off an explosion of thoughts. I use audio notes to capture these thoughts because I know that they will be long forgotten if I don’t capture them on the go. The audio note recording interrupts listening to the podcast episode, which I then return to once I finish recording.
Desired User Experience While Running
Recording while trail running drives some specific desires:
- For any required taps, the solution should use large icons and buttons so I can easily and safely interact with the apps while maintaining a running pace on often rough terrain.
- The solution should minimize the total number of interactions with the phone. For example, the podcast should restart on its own when I stop recording the audio note.
The ideal is either voice commands or to be so simple that I can record audio notes using only one hand and just my thumb on the phone screen. Scrolling and selecting from text menus are not realistic options.
To keep this next PKMS iteration as simple as practical, while hopefully largely still achieving the goals, I decided to for now forgo the alternative of using voice commands.
Podcast Player Selection
I obviously needed to change my podcast player to an app that supported recording highlights (excerpts) as audio and text. I considered Airr, Momento, and Snipd, which all have highlight functionality.
I eliminated Airr because of a lack of company Twitter traffic, and negative feedback on Twitter; for example, here, here, and here. [After first publishing this article, I also discovered what Snipd calls out as their competitive advantages over Airr.]
Momento looked promising; however, it isn’t receiving the Twitter buzz that Snipd sees, and it didn’t advertise an Obsidian integration, whereas Snipd does. Therefore, Snipd for the win in this PKMS iteration.
Audio Notes App Selection
This May 2021 Sweet Setup article along with internet and Apple App Store searches informed my longer list of audio notes apps to try.
Must-Have Functionality
To achieve my upgrade goals the audio app must have:
- Reasonably accurate transcription (voice-to-text)
- A workable workflow for importing text into Obsidian
Both of these requirements exist on a continuum. They are not simple ‘yes’ or ‘no’ answers.
I’m admittedly asking a lot for my transcription functionality. Tough conditions. Noise from my shoes impacting the ground. Frequently windy. The phone is moving with my arm swings. And, shortness of breath.
The ideal workflow would include, for example:
- The ability to bulk export. I may take ten or more audio notes during a single run and exporting each one individually is a nuisance.
- Default actions, as in a script. For example, all transcriptions append to a Daily Note unless specified otherwise.
- Retention of the audio version as a backup for any poor transcription.
- Automatic metadata in the transcription. For example, the date and time of audio note creation.
Narrowing the List
I considered, but immediately rejected:
- Voice Memo Recorder & Changer, which had too many screens and is not that popular.
- QuickVoice2, which is supposed to be the Pro version with transcription; however, after paying the US$2.99 and downloading, I was left with a non-Pro version without transcription.
For now, I narrowed my shortlist to three known to be solid apps. Listed in order of perceived popularity for voice-to-text: Otter.ai, Drafts, and Just Press Record.
I evaluated each of these apps for my desired user experience and must-have functionality. The highlights:
Otter
The Otter screen flow for exporting an audio note transcription is straightforward, albeit a bit long as seen in the next two Figures:
I understand Otter’s business model; however, I still find in-app ads and reminders of how many minutes I have remaining in the 600-minute monthly cap to be annoying eye clutter. I’d gladly pay for the Pro version if my desired user experience and must-haves were fully achieved with that plan. However, they will not be. For example, the simple requirement of bulk export isn’t available except with the Business plan that costs even more than Pro, at US$20 per month.
When running, aside from podcasts, I occasionally listen to Pandora Music or iHeart radio news. With Otter, when I interrupt either of these apps to take an audio note, the sound from Pandora or iHeart continues. In contrast, the Drafts and Just Press Record apps correctly handle the interruption by stopping the app I’m listening to.
A typical Otter transcription:
what would I need to have before I would pay money? Of course they are. Transcripts can be good, but also the integration with that city got to be a lot better than what I’ve seen before. And if I had those things for my audio notes, have you willing to pay at least a couple of bucks a month
Versus the actual spoken words, where the boldface indicates where there are shortcomings in the transcript:
what would I need to have before I would pay money? For say Otter. Transcripts can be good, but also the integration with Obsidian got to be a lot better than what I’ve seen before. And if I had those things for my audio notes, I’d be willing to pay at least a couple of bucks a month
Otter is the only app of the three that automatically assigns tags to the audio notes. That said, I’ve not found them to be particularly helpful so far. The tags might be useful if I was to retain a large collection of recordings; however, for my use case, I delete the audio recordings after capturing in Obsidian.
Drafts
Unlike the other apps, Drafts requires some unworkable steps to begin an audio note while trail running. See the next two Figures for two alternatives. This shortcoming is at least partially explained by knowing that audio is not Drafts’ core functionality. It’s text. The good news is that, with some further setup, the steps in the figure below can be replaced with a Home Screen Widget (reference).
As seen in the following figure, the export to Obsidian follows a similar flow as in Otter; however, with only four instead of six screens:
My biggest disappointment with Drafts is that the audio isn’t saved. After stopping a recording, only the text transcription is available. This presents a problem when the transcription is so inaccurate that I can’t follow the point I was trying to convey. Given this shortcoming, I’ll not try to do a transcription accuracy first look as I did with Otter and Just Press Record.
Drafts handles interrupting Snipd as desired. Snipd stops whenever a Drafts recording starts and automatically restarts when completing the audio note. No need to reopen Snipd and tap Play as required with Otter and Just Press Record.
Just Press Record
The Just Press Record app costs US$4.99 for lifetime access.
The screen flow for exporting to Obsidian is virtually identical to Drafts:
A unique feature of Just Press Record is that it automatically stores audio clips in iCloud, as seen in the next screenshot. Good to know; however, text files from transcriptions would be more appealing to me. The following support answer from five years ago is the most detail I’ve been able to find on this: “Just Press Record stores transcripts inside its audio files. It does not use separate text files.” The answer then goes on to suggest using Export to obtain text files (as per the Figure above). I have not figured out how to extract transcripts from the mp4 audio files.
A typical Just Press Record transcript:
No I think about holding with one hand and doing the foam I’m having really watching me rhyme I don’t know is do that it’s nice to yeah he’s not Ashley
Versus the actual spoken words, where the boldface indicates where there are shortcomings in the transcription:
I’m thinking about the holding with one hand and doing with the thumb. Now that I’m really watching me run, I don’t always do that. It’s a nice to have, not a must have.
Installation and Configuration
Capturing podcast highlights and voice-to-text pushed me to install Obsidian on my iPhone, and iCloud on my PC. For anyone in the same starting place, the next figure has some quick tips to get you started:
For now, I’m using a separate Vault named Mobile to isolate any potential challenges with fully syncing my production Vault. For now, I’m fine with manually copying Snipd podcast highlights into my production Vault.
If I decide to merge Vaults, I will more seriously consider paying for Obsidian Sync. In my current lifestyle, I’m not feeling a strong pull towards having my full Obsidian Vault available on mobile. If this changes, I’ll happily pay for Obsidian Sync. OneDrive currently handles my production Vault safety backup.
There is nothing special about the installation or configuration of any of the audio notes apps. All are standard Apple App Store workflow.
I have Snipd and whichever audio note app I’m currently testing in The Dock (the fixed lower row) portion of my iPhone. The feature photo at the top of this article shows this. The placement makes the icons as easy as possible to find and use while I am running.
Initial Impressions
Snipd
As I write this, it is now day three with Snipd and it’s going well. The screenshot below illustrates my first Obsidian note populated by Snipd. I created this by following the instructions in How To Export Your Podcast Highlights To Obsidian, by the folks at Snipd in November 2021 here on Medium.
I have some remaining work to do to polish my workflow and increase my competency with the app; however, I have no concerns or worries at this point. This app is a “keeper”.
1 October 2022 Update: See Using Snipd With Obsidian To Maximize Learning From Podcasts for details of the workflow I settled on.
Audio Notes Apps
Based on an informal review of all the recordings I’ve made, Otter has the most accurate transcription. Just Press Record has the best user interface; by a lot. And, all of the apps require more work than I had in mind to get to the end state of a day’s worth of notes embedded into a Daily Note in Obsidian.
So far, I have a fairly small dataset of transcription samples from my real-world trail running conditions. I wonder about orchestrating a more rigorous test where I would somehow feed all the candidate audio notes apps the same audio clips and see if Otter still remains the winner.
To substantiate my user interface claim, compare the following screenshots for Just Press Record and Otter interaction to Pause or End an audio note:
Even with Otter, the transcripts are far enough off the mark that by the time I edit them to the desired and do all the extra work of export/import, I’ve used much more time than my current system of manually transcribing directly into an Obsidian note. Continuing to explore voice-to-text is in jeopardy. I may just let this goal go by the wayside.
As of now, my audio notes tend to be one to three sentences, maximum. My conclusion for time savings might be different if I start doing longer notes as in crafting paragraphs for an upcoming article.
If I was to push harder for an audio notes solution that meets my requirements, a few other apps that might be worth considering include Day One (described in The Sweet Setup article), Dragon (pricy for my needs at US$15/month), and Voiceliner.
Follow me on Medium and on Twitter to learn more about Personal Knowledge Management, Tools for Thought, and Second Brain.
