Einführung
· Briefly about the organization, working group
WMN (WMN.hu), founded in 2015 by Kriszta D. Tóth, is an online women’s and lifestyle magazine that primarily focuses on public affairs and cultural topics, as well as human stories. WMN is independent, humane, and free, and from the very beginning has been characterized by a distinctive, personal voice that is sensitive to the perspectives of women, children, and vulnerable groups. This tone is consistently present in the magazine’s articles, podcast programs, and events.
· Was there a specific event or need that prompted you to start looking for a solution?
The Alrite speech recognition program primarily helps us produce our written interviews more quickly, with less time and energy investment. We wanted to support our authors in this process as well.
Herausforderungen
· What challenges or difficulties did you face before using Alrite?
Before we started using the Alrite speech recognition program, transcribing a one-hour interview could take up to a full working day. Every ten to fifteen seconds the recording had to be stopped, and it was cumbersome to rewind if something was not typed accurately. In addition to taking an enormous amount of time, this was an extremely boring and monotonous activity for most journalists. Many of us dreamed: if only we had an assistant or intern who could take over these tasks. Ten years ago, very few people in the profession foresaw that this would eventually be solved by a professional service like Alrite.
· Was there a situation where you especially felt that an automatic speech recognition solution would be useful?
When an interview turned out to be long, many journalists already dreaded in advance how they would cope with transcribing the material, and how much time and energy it would require. At the same time, we always considered it unprofessional to write an interview from memory or to put statements into the interviewee’s mouth that were not said, or not said in that exact way.
Entdeckung
· Where did you first encounter Alrite? (e.g. recommendation, conference, online search)
Almost all of us first heard about the Alrite speech recognition program from one of our colleagues. They praised how helpful it was, so we decided to try it ourselves, initially using the free version. We were also pleased to experience how much the service has improved over the years, so that today we have only minimal work correcting transcripts even if the recordings were made in noisier locations, such as cafés with background music or the occasional hum of a coffee machine.
· Based on which criteria did you ultimately choose Alrite? What was the “decisive argument”?
The decisive argument in favor of Alrite was precisely the improving quality: we experienced that the service keeps getting “smarter,” is particularly reliable in Hungarian, and the interface is user-friendly and easy to navigate.
Vorteile und Ergebnisse
Did you experience time savings or other measurable, quantifiable benefits from using Alrite?
Alrite means several hours of time savings in interview transcription. Although we did not conduct precise measurements, based on estimates (and typing speed), for a one-hour interview this can mean up to 6–7 hours of time saved. But the benefit is not only temporal: transcription used to be such a monotonous, mechanical task that one of our journalists put it this way: she “almost came to hate” the interview by the time the transcript was finished, even though that should have been the point when she actually began working with it in terms of content.
· Which features became real “favorites” of the team, and why?
In addition to transcribing audio files, one of our authors highlighted a feature she finds very useful: being able to start playback of the audio file from wherever the cursor is placed, without having to search within the recording.
· In summary, what is your overall experience with Alrite, and how has speech recognition influenced your workflows?
Thanks to Alrite, there are weeks when one of our journalists can save almost an entire working day by not having to prepare transcripts manually. As a result, we can produce more interviews, opinion pieces, and reports, and we can focus more on the part of our profession that requires human creativity, empathy, and the ability to place things in context.
· Was there a specific project or task where Alrite produced particularly striking results?
For now, we use Alrite for transcribing interviews and producing written content, and in this area we have experienced the advantages described above.
· Was there any surprising or unexpected benefit that you did not anticipate from speech recognition?
So far there has not been, but if you suggest a feature that would be worth trying, we would be happy to consider it.
· What changes have you observed in workflows since using the application?
Although we still review and correct the transcripts produced by Alrite, this fortunately requires less and less intervention. On average, the process now takes about one and a half times as long as simply listening back to the recording. While we hope this time can be reduced further, we also use this phase of the work to refresh our memories of the conversation and already start thinking about how to structure the material and what to highlight.
Testimonial (Reference recommendation)
· On the Alrite website, in the “What our clients say” section, we would like to feature your recommendation. For this appearance, could you send us a 1–2 sentence recommendation describing what you use the application for and why you like it?
Previously, as journalists, transcribing interviews was a painful, monotonous, mechanical task; by the end we almost hated the material, even though the real content work with the text should have begun only afterward. With Alrite, however, we have gained a huge number of working hours and mental energy, allowing us to focus on tasks that require human creativity and emotional presence.
Development suggestions
· Do you see any development opportunities or features, or do you have any suggestions that would make the speech recognition system more efficient to use or specifically support your use case?
One of our authors noted that when she does not copy the transcript using Ctrl C – Ctrl V, but saves it instead, the text formatting in the document becomes confusing, sometimes even individual letters end up on a new line. Another journalist said that although she tries to choose interview locations carefully in terms of noise levels, it unfortunately still happens that she has to correct the text relatively extensively by ear. Apart from transcription, we have not yet used other functions of Alrite, but we will keep an eye on whether there are additional workflows where it would be worth applying the service.
And thank you very much for everything so far. 🙂