Streamlining the Mortified production process

A participant in a Mortified stage show. Source: Flickr (CC BY-NC 2.0)

Teen angst. Embarrassing moments. Cringe-worthy choices. These experiences are the bread and butter of Mortified, the stage show and podcast that encourages ordinary people to “share the shame” of their childhood writing.

Dave Nadelberg founded Mortified in 2002, and today runs it alongside his producing partner, Neil Katcher. Together they help produce stage shows in 20 cities across the United States and the world — and transform a curated selection of those stories into three monthly podcasts. They use Pop Up Archive to automatically transcribe and tag tape in the editing process. Continue reading

Using Pop Up Archive to bring broadcast stories to digital communities

“Pop Up has been a way to make our process more efficient and allow ourselves to put more online, which is more sustainable for our staff and better for Vermont Public Radio as an organization.” —Angela Evancie, VPR Digital Editor for News

Angela Evancie is the digital editor for news at Vermont Public Radio. She facilitates online stories for VPR’s website, repurposing broadcast news for the web and overseeing special web-only projects such as data visualizations, interactive features, and videos (like this nifty explainer about the Iowa Caucus from 2016).

Most of any public radio station’s content is conceived of as an on-air (audio) product; the digital editor’s job is to make sure that the station is also serving online audiences, who are reading or interacting with the story in ways other than listening. Angela and her colleagues at VPR use a Pop Up Archive team account to transform on-air pieces into written content for the VPR website.

Source: Vermont Public Radio

VPR has three daily news magazines — Morning Edition, All Things Considered, and Vermont Edition — and short interviews from each of these shows often find a second life as online articles.

A typical process starts with the radio broadcast being uploaded to Pop Up Archive. Sometimes, VPR editors take advantage of the fact that they’re publishing online as an opportunity to post a longer version of the original piece.  “After we get the transcript from Pop Up Archive, then it then gets edited. If we’re keeping the online piece in an interview format, we’ll just edit it for accuracy. Other times, we’ll rework the transcript into an article format,” Angela says. Next, the article is posted to VPR’s content management system, given art and final polish, and published to the web. A few recent examples that went through Angela’s team to be republished as content for online readers include a story on a children’s book about coping with death, an interview with author Chris Bohjalian, and an interview with Vermont’s new attorney general.

VPR also leveraged Pop Up Archive in the weeks leading up to the general election. The station conducted interviews with all of the major and third party candidates running for statewide office: 24 interviews in total. The interviews occurred over several weeks and aired in different shows and segments. In order for the web content to be valuable to readers and voters, VPR wanted to package the interviews in such a way that their audience could do side-by-side comparisons of the candidates on specific issues. They used Pop Up Archive to transcribe the interviews and repackaged the interview highlights into a comprehensive voters’ guide. “The content was among our most successful in 2016, and it wouldn’t have been possible without Pop Up,” Angela said.

The process of transforming broadcast content into web content is one that occurs at VPR almost daily. But when Angela started at VPR three years ago, its web operations team was much less robust, and the process of translating on-air content into web content happened much less frequently. Each time they wanted to share a written version of a piece online, Angela or another producer would have to do their own transcription by hand.

Pop Up Archive has been an important tool for the station as its appetite to do more online, while being realistic about resource and time constraints, has grown. “Pop Up has been a way to make our process more efficient and allow ourselves to put more online, which is more sustainable for our staff and better for Vermont Public Radio as an organization,” Angela says.

See you in the archive,
The Pop Up Archive team

Analyzing audio with free speech-to-text models from Pop Up Archive

Source: Wikimedia Commons.

An emerging archival speech technology project is underway at the University of Texas at Austin, and its inspiration comes from a piece of software called ARLO that was developed to analyze an entirely different form of aural communication: bird song.

Called HiPSTAS (High Performance Sound Technologies for Access and Scholarship), the project is a collection of software tools and communication channels to help researchers access and analyze archival spoken word collections — for example, by analyzing audio spectrograms to identify particular traits such as pitch, tone, and speed. HiPSTAS was founded by Tanya Clement, an associate professor at the UT-Austin School of Information, along with collaborators at the University of Illinois Urbana-Champaign. The group has received several grants since its founding in 2012, including two from the National Endowment for the Humanities.

The ability to search archival spoken word recordings along multiple parameters has powerful implications for historical research and literary scholarship. While digital sound recordings are increasingly available to scholars, searching through the files for discernible patterns is time-intensive and cost-prohibitive.

Stephen McLaughlin is a research assistant and PhD student in Information Studies at UT-Austin and a contributor to HiPSTAS. Steve is currently working with WGBH and Pop Up Archive on the American Archive of Public Broadcasting, part of which entails a massive effort to use machine learning to identify notable speakers’ voices (for example, Martin Luther King, Jr.) from within AAPB’s 70,000 digitized audio and video recordings. As part of the AAPB project, Pop Up Archive transcribed the entirety of the AAPB and released free models for the open source speech-to-text software Kaldi, with the intention that the models could be used to enable other audio collections to be transcribed and thus searchable.

“The language models Pop Up Archive has assembled are more accurate, and do a way better job of identifying proper nouns, than other tools I’ve seen.”

While testing potentially useful tools to combine with HiPSTAS, Steve downloaded Pop Up Archive’s free models. “The language models Pop Up Archive has assembled are more accurate, and do a way better job of identifying proper nouns, than other tools I’ve seen,” said Steve. “The other exciting thing is that those models can be extended. I can take the model, transcribe a recording or correct its machine transcription, and use that output to make it more accurate for my needs. The flexibility of this system is really exciting.”

Kaldi is free software, released under an MIT License, that can be run on institutional servers, making it a natural choice for libraries, archives, and institutions that have a long-term approach and technical resources to support their projects. This is something Steve has personal experience with: “I’ve seen this happen, where commercial API services are here and gone. So the ability to run Kaldi locally is really an important tool to have in your tool belt. Pop Up Archive is providing a really valuable service by contributing their resources to a project like the Kaldi models.”

As an exercise, Steve recently used Kaldi to run through 80 hours of recordings by the poet Robert Creeley from the PennSound archive, the topic of his undergraduate thesis. Creeley used certain phrases and expressions over and over, such as the word “company” — so Steve searched the Kaldi transcript for every instance of “company” and then generated a supercut, combining the different tones and intonations into a continuous string of audio.

HiPSTAS team members are very interested to share what they’ve learned from their collective experience, and in the coming months Steve plans to publish code demonstrations for other researchers working with archival audio. He’d also like to write about his work with the American Archive of Public Broadcasting to identify speakers in large audio collections. Finally, he intends to run a workshop covering his work with Kaldi at one or more conferences later this year.

Anyone currently working on archival sound projects can access HiPSTAS’ preliminary release of tools called the Audio Tagging Toolkit. You can also download Pop Up Archive’s Kaldi models on Github.

See you in the archive,
The Pop Up Archive team

Find new listens with Alexa’s “Magic Podcast”


Amazon Echo digital personal assistant. Source: Flickr.

Wouldn’t it be cool if you could just tell someone, “Find me a podcast about science,” and they’d do it immediately? With the Amazon Echo digital personal assistant Alexa, you can.

Erik McBeth is the developer of the Alexa skill called Magic Podcast, which uses the Audiosear.ch API to find and play podcasts based on title or subject matter. This means you can simply say, “Play Fresh Air” — or whatever your podcast of choice is — and the newest episode of that podcast will begin. Even cooler, Magic Podcast can also be used as a discovery tool. Discover podcasts by subject area by asking Alexa to “Find a podcast about food,” or “Find a podcast about business.” Magic, indeed.

Erik got the idea for Magic Podcast after developing another Alexa skill called Magic Jukebox that allows users to stream SoundCloud music on their Amazon Echo. The most complicated part of the skill — responding to requests to pause, start, and skip tracks — was written during the Magic Jukebox development cycle. All Erik needed to adapt it for podcasts was to find an API that could plug into his existing infrastructure.

“So far so awesome. Don’t need to subscribe to anything. Say the name of the podcast and it plays the most recent episode. Simple.” — an Alexa Magic Podcast skill user

He looked at a few different options, including Feed Wrangler and Digital Podcast, as well as others that had seemingly gone defunct. None were a fit, since Erik needed links to actual mp3s served up over HTTPS (Alexa will not play audio from untrusted sources). He also liked the fact that by using Audiosearch’s API he could allow his users the option to play a specific show or look for podcasts on a certain subject.

Users have responded very positively to the skill, commenting on the novelty (“This skill just upped the cool factor of Alexa! Made her more functional for me.”) and simplicity (“So far so awesome. Don’t need to subscribe to anything. Say the name of the podcast and it plays the most recent episode. Simple.”) Erik has many plans for adding to the skill in the future, such as allowing users to save favorites or screen out podcasts they prefer not to listen to.

For now, users are simply enjoying the ability to add podcast functionality to their Echo experience. Have an Echo? Enable the “Magic Podcast” skill here. Or take the Audiosear.ch API for a spin yourself!
See you in the archive,
The Pop Up Archive team

Resurfacing treasures from California archives

screen-shot-2017-01-04-at-7-28-39-pmCalisphere is a growing aggregation of digital collections from 180 institutions across California. It currently contains more than 670,000 images, texts, and audiovisual materials on a sweeping range of topics that reflect the collecting areas of its participating institutions.

Sherri Berger is a project manager at the California Digital Library who managed the re-release of the Calisphere collection in 2015. We talked to Sherri about this massive project, including goals for the redesign, the response they’ve had from the public, and some of her favorite audio available in the collection.

What was the goal of Calisphere? Who is its audience?

Calisphere was developed and is maintained by the California Digital Library, in partnership with the ten campus libraries of the University of California. One of the biggest and oldest regional digital library sites of its kind, Calisphere is more than ten years old and still going strong!

Calisphere was originally launched in 2006 with the goal of making academic research collections accessible to non-academic audiences, with a deliberate focus on K-12 teachers. Today, the site maintains this public mission, but we think of the audience as even broader, including K-12 students, undergraduates, scholars, family historians, creative professionals, and the “casually curious.”

You recently did a re-design. What challenges did you encounter?

I think one of the hardest things about a “reinvention” of an existing site is balancing the new vision with the needs of legacy objects and metadata. This issue comes up in any redesign, but the scale at which we work and the heterogeneous nature of the collections can make it all the more complex. At the same time, you don’t want to always do the safe thing — you want to take risks that will make at least a majority of the content more easily discoverable and interesting for users.

Between the project team at the California Digital Library, additional development staff at UC Berkeley, and librarians at all ten campuses, I’d estimate about 30-40 people played some role in the implementation. It was a true collaboration.

Sound is part of your collection. What types of audio are included?

Calisphere only recently has started to take in sound objects in a systematic way, which is exciting. The audio really runs the gamut.

There are a ton of oral histories, featuring all sorts of people, whether they have had a renowned career (for example, jazz composer Dave Brubeck from University of the Pacific) or have important personal or community stories to tell (for example, the Vietnamese Oral History Project from UC-Irvine).

The single largest collection of audio is UC-Santa Barbara’s Cylinder Audio Archive, which includes recordings made from the late 1800s to early 1900s, including popular songs, vaudeville acts, comedic monologues, and more. (I am particularly fond of this recording of the “Maple Leaf Rag” hand-played by Scott Joplin.)

And many of the sound objects on the site were identified as essential for the preservation of California’s historical record. These were digitized by the California Audiovisual Preservation Project from a variety of institutions.

How did you decide how to organize the audio to make it searchable?

Admittedly, Calisphere isn’t currently super optimized for searching audio for a couple reasons. For one, the model for Calisphere is that it is first and foremost an aggregator: for audio objects, it brings the metadata in, but it currently doesn’t stream the files right on the site (except in some special cases).

Meanwhile, from a user experience perspective, the fact is that Calisphere was traditionally mostly an image and text repository, so we leaned heavily on the visual component. But now that we are getting more audio, we are revisiting some of our design decisions to make sure that content is just as findable and accessible.

Why is Calisphere important?

California contains such an incredible group of collecting institutions, from public libraries to universities to community historical societies. But many of those institutions lack the technical resources to really get their collections out there and visible to their users. For instance, maybe they’ve got some digital objects in a library catalog, but they are hard to find on the open web and not presented in a modern design.

In Calisphere, those collections really shine: they get the full user experience, they get indexed in Google, they get put in exhibitions, they get sent to the Digital Public Library of America. In short: they get discovered and used!

Thanks, Sherri!

See you in the archive,
The Pop Up Archive team