Deaf and hard of hearing people often find it difficult understanding speech, especially in group conversations. This is made even more difficult when the environment has lots of background noise or poor acoustics. Just think of noisy restaurants, social gatherings, meetings, etc. This can make it hard to join in with the conversation and can bring about feelings of social isolation.
Speaksee wants to help by giving people who are deaf or hard-of-hearing a way to join group conversations and support communication. The Speaksee Microphone Kit transcribes speech into text in real time and shows every speaker in a different colour, regardless of background noise.
There are several apps available for free that also transcribe speech to text. With all this choice, it is important to understand what differentiates Speaksee from other apps and why is our price point different?
What is the difference between Speaksee and other cheap (or free) apps?
Speaksee is the only solution that uses multiple external microphones, while cheaper apps just use the microphone of the mobile phone. This brings significant advantages to the user such as:
- The distance between the speaker and the microphone is small. The smaller the distance, the better the speech is captured, the better the quality of the transcription.
- Each person in the conversation is given their own microphone, ensuring top quality transcription, even when people are speaking over each other!
- Our microphones use cutting edge, beamforming technology to isolate speech and eliminate all background noises in any environment.
- Each person in the conversation has a different colour for their subtitles, this makes following the conversation a much more natural experience.
Cheaper apps do not have these advantages, resulting in less accurate transcription making it more difficult to fully participate in the conversation.
Furthermore, Speaksee protect your privacy! All conversations using Speaksee are entirely private (even from Speaksee) and all your data is encrypted giving you peace of mind regardless of what you are discussing. Other apps can hear everything you are saying and most even use your conversations to improve their services!
At the beginning of 2023, a comparative study took place at Radboud UMC in Nijmegen (The Netherlands) by clinical physicist audiologist Jan-Willem Wasmann. Now it is also objectively proven that Speaksee works better than alternatives!
A quote from Jan-Willem Wasmann from the conclusion of the comparative study: “In measurements where we mimic the usage situation, Speaksee outperforms the other systems. This indicates that with a single Dutch speaker, Speaksee is the best option of the systems tested. We were not yet able to simulate multiple speakers (with multiple speakers and microphones simultaneously), our expectation is that Speaksee will further distinguish itself from alternatives in this usage scenario.”
During the comparative study, multiple research methods were used with the Speaksee Microphone Kit and alternatives, for the purpose of assessing the quality of transcription. During the comparative study Speaksee was compared with NALScribe and Live Transcribe.
In 2020 the same tests were executed with other apps; AVA, Earfy and Speechy. The results of the other apps of 2020 also are also described in this article.
Speaksee has the best accuracy when transcribing speech
The most common metric used to measure automated speech recognition is the Word Error Rate (WER). WER is the ratio of errors in a transcript to the total number of words spoken. The lower the WER, the more accurate the transcript.
The results from the WER test show that Speaksee makes the least mistakes in transcription compared with other apps in a Dutch dialogue. While Speaksee’s transcription results in only 2 errors per 100 words, other apps resulted in around 20 errors per 100 words or higher.
Speaksee has the highest accuracy in Dutch transcription.
Speaksee works better than other apps in noise
To investigate accuracy in background noise, the Dutch PLOMP test was used. This involves playing sentences in noise. This measurement investigated how quiet speech can be made in noise while still transcribing speech to text properly.
A negative result means that speech can be made quieter than noise and still be transcribed to text well. The more negative the result, the better the performance in background noise.
The results of the PLOMP test show that when using Speaksee the speech level can be made much quieter than the noise level, and the sentences are still transcribed well. With the other apps, the speech level cannot be made quieter than the noise.
With this result, Speaksee matches the performance of a well-hearing person in background noise. This indicates that hard of hearing individuals would experience very significant benefits from Speaksee with background noise.
Noise was played as loud as street traffic. When sentences were played which had a quieter volume than background noise, Speaksee was still able to transcribe them properly to text. The speech level of Speaksee corresponded to a normal conversation level. In the case of other apps, sentences had to be played much louder than the background noise to still be properly transcribed. This speech level corresponded to loud shouting or a train passing by.
Applications for using Speaksee Kit
The study shows that Speaksee is the best transcription app in situations with background noise and also has the highest overall accuracy.
Because of this, the Speaksee Kit can be used in so many situations such as:
- In a restaurant or during a lively social gathering:
- In professional situations such as meetings, presentations or briefings:
- At a clinical visit:
- In the car or in a Taxi or Uber:
- During class:
Interested to learn more?
Did this blog spark your interest and would you like to receive more information about how to get Speaksee? Would you like to see how it works for yourself during a demo?
We would love to receive your e-mail at email@example.com.
At Speaksee, we want to enable deaf and hard of hearing individuals to fully participate in group conversations, even with background noise. We are dedicated to our mission to create accessible communication for people who are deaf or hard of hearing!