Google Cloud Speech-to-Text vs. Phrase Localization Platform

In our real time meetings or webinars where larger audience are expected we have enabled the captions options with Google Cloud Speech-to-Text tool this start transcribing the complete audio conversation in the neat text format. Also while performing the interview process as well we use this tool to make sure that we adhere to certain rules and are being checked by the superior management team to make sure the transcription has required questions being asked on for quality analysis. Also during the customer call we use this tool to make sure two way communication is transcribed and will be later reviewed when there is an escalation by the superior management

Incentivized

I like Memsource because when I am not home and I don't have my laptop, I can borrow a computer, log in to my Memsource account and I'm ready to begin translating. I can even download the source files and check the TM and glossary there. It's not necessary to download the software and lose time on that.

• An amazing tool which helps a lot in a meetings.
• It's an efficient tool for improving efficiency by saving a lot of time typing. It saves at least 40-50% of our time, thus increasing efficiency.
• Incredible accuracy with multiple accents & multiple language.
• It takes punctuation into consideration.

Incentivized

• The synching of the desktop app with the online app is great and very helpful.
• It is easy to look for all segments containing a specific term and correct them in a batch.
• Suggestions of other segments where a certain term already appears are very helpful.

Incentivized

• The software does occasionally get confused by confusing terminology.
• Its web-based interface can also feel a tad hard to use compared to more appealing desktop apps.
• I've experienced the occasional technical issue, though the provider's support team is quick to troubleshoot.

• In the case of EN to JA translation, we enter text and then convert it to get the correct final text (word or phrase) which uses correct Chinese character(s), as there are often multiple Chinese characters with the same reading but different meanings. When those conversion options are displayed, it is usually possible to change our selection among them by hitting the Tab key, but in Memsource, hitting the Tab key makes us leave the text conversion and move to the source segment, so we must make sure to use arrow keys to choose the right conversion option when using Memsource.
• When there are tag elements used in the source text, the tags must exist in the target text of course, but the tag order also must be the same. The tags cannot be moved around in the segment, which causes problems in the case of the Japanese language, because the word order differs between EN and JA.
• In the Japanese language, Italics are basically not used, so there must be no texts between the tags which specify Italic font. But then the segment cannot be confirmed and the job status cannot be changed to "Complete".

I use and manage an Academic edition (approx. 15 students) and the corporate license (10 users). In both cases, the features are easy to find, set and use. My students and my translators understand the dynamics of the platform easily and get used to them quickly.

The reasoning behind my 10 is that the UI is very intuitive; I didn't require any formal training to use it. Google's speech-to-text is not just a conversion tool; it helps automate mundane tasks, saves time, and has an almost human-like understanding.

Incentivized

It is designed to be a lean system, but that means that things are clustered in weird groups and a lot of trial and error is involved getting the settings right. If something goes wrong first level support is very quick and has very fast answers which are quite typical of first level support and essentially equate to asking you if you have switched it off and on yet. Anything more involved is challenging for support to address.

We rarely use support, but most questions were answered in a timely fashion, although we didn't exactly find them satisfactory. That's mostly the fault of the software and not the Support team because we asked for things that Memsource couldn't do.

Incentivized

It delivered high accuracy in accented and noisy environments. Regarding its language support, it offers a variety of languages and dialects. Its's Api's are well-documented and easily integrated with our GCP-based stack. Also, its deployment is fast, and it is cost-effective. And if we talk about its translation, it gives real-time and generic translation with great punctuation. Finally, its speaker diarization makes it a cool and yet powerful tool that helps people.

Incentivized

It is very expensive and more difficult to navigate. Compared to the other support is slow and ineffective. Dealing with anything more involved is all but impossible. It has become more prone to bugs since being bought by TMS. Support suggests things like "use incognito" mode to log in. Use a different browser to log-in. That is a concern given the customer data involved if things at that level are not being fixed. If that's what it's like in the showroom, what is the kitchen like?

• Automating the transcription process saved time and resources compared to manual transcription.
• Speech-to-text enabled us to make audio content accessible to a wider audience, including individuals with disabilities.
• We gained valuable insights into customer preferences, behaviors, and sentiment by analyzing voice data.

Incentivized

• Positive Impact due to saved time.
• Positive Impact due to not worrying about the format of the file (it is automatically taken care of).
• Positive Impact due to being able to follow the client's style guide clearly and without errors.
• Positive Impact due to avoiding common errors through QA feature.