Google Cloud Speech-to-Text Reviews

I can answer questions about my experience with SQL Server as we are trying to capture reviews for SQL Server. We don't use the reporting services within SQL Server; we're using this for heavy-duty engineering inside an AI engine.

We don't use SQL Server Data Warehouse or SQL Server Management. We just use SQL Server extensively. We use Office, but in terms of engineering products, the major use right now is SQL Server.

We use TensorFlow, which is different from TensorLeap. We use Google Cloud extensively; we use their AI, STT, and TTS capabilities.

We use these two products to be able to talk to our users, and then we have an AI meaning engine that back ends that. Once we get the speech, we can tell what it means.

Accuracy with Text-to-Speech is important. If you want to talk to your computer, using Google is the way to do it, but it's not without problems. Natural intonation and pacing in communications have gotten hugely better over the last two or three years.

We're following the research at MIT of Rosalind Picard about these kinds of issues, and we plan to implement it. This applies to both Text-to-Speech and Speech-to-Text because it's a conversation.

When we tested Google Cloud Speech-to-Text a few years ago, it was better than the equivalent IBM products because it was more accurate. However, it's not entirely accurate, so we have to correct for those errors in our AI software.

Google Cloud Speech-to-Text is 100 out of 100 when it works, and when it doesn't work, which is fairly often, it gets a zero. It doesn't fail gracefully; it fails in an unexpected way.

Google Cloud Speech-to-Text is not entirely accurate, so we have to correct for those errors in our AI software. It uses neural networks, and that stochastic processing is 70% to 75% accurate. It gets it wrong too often, and since I personally work with this, I don't appreciate that. However, they seem to be the best option currently.

We have to write our own improvements because their tools to improve transcription accuracy in our domain aren't very powerful.

The timestamp technology for recognized words is inadequate, so we don't use it. We understand words based on their meaning, and we have a whole AI engine that does that, which is one of our differentiators from a product standpoint.

We didn't use the custom voice creation feature; we just use their voices, which are fine for our purposes.

I have dealt with Google Cloud Speech-to-Text for five years.

On development machines, it works great out of the box, which is terrific and one of the reasons we chose them. However, in the cloud, we encounter numerous errors.

The support from Google Cloud Speech-to-Text is inadequate. We are dealing with them on our development talks, and there's significant finger-pointing regarding problem ownership. Moving our systems to Google Cloud and maintaining the same performance as on development machines is problematic.

I cannot provide a score for their support because when issues arise and developers report unintelligible error messages, there's no effective support from Google. Comparing the support we receive from Google versus Amazon, Google's support is substantially inferior.

Without considering support, the product rating would be closer to ten, but including support changes everything significantly.

Neutral

We evaluated IBM and Nuance as alternative options. The main differences between these solutions come down to accuracy. It doesn't matter what you do with the utterance if you don't recognize it. These other products didn't recognize speech as effectively as Google Cloud Speech-to-Text does.

We don't prioritize pricing considerations; they were competitive, but our primary concern is functionality, and currently, the competitors don't work well enough for us to consider them.

Our experience with pricing and licensing for Google Cloud Speech-to-Text is that we didn't have any other viable choices, so we cannot effectively evaluate if it's well-priced or badly priced.

We evaluated IBM and Nuance as alternative options.

Complex conversations require more work, and measuring the effectiveness of automated customer interactions involves abandonment rates and internal metrics we've developed about understanding conversational utterances and user interaction quality. Currently, the performance is disappointingly low.

We haven't implemented multi-language capability yet, though we're planning to try it as it's a significant advantage for our products.

We use Google Cloud rather than AWS marketplace for these products. Google Cloud Speech-to-Text sounds incredibly natural, which is impressive. The improvement in naturalness brings satisfaction, though the support needs significant enhancement.

This product is certainly suitable for enterprise environments.

The overall rating for this solution is 8 out of 10.

The main use cases involve clients handling various calls day-to-day who have a quality analyzer or auditor wanting to verify what representatives spoke with specific clients. This piece of technology comes into play because the auditor cannot go and listen to long audios for call recordings that span 10 to 20 or more hours. They won't be checking individual calls, but using Google Cloud Speech-to-Text, we can easily transcribe the call with respect to who has spoken what, with specific speaker diarization.

We can ask any open-source AI, or even paid AIs such as ChatGPT AI, to provide the transcription and the context of representative conversations with clients. From that, we will get a complete overview of the call in a few seconds.

We can transcribe multiple calls, and if we want to check our representative's productivity per day, we can easily transcribe all the calls and get an overall understanding of what has occurred in the calls. This is the broader scope of the Google Cloud Speech-to-Text solution I developed for my client.

The specific feature of Google Cloud Speech-to-Text that I find most beneficial for transcription needs is the speaker diarization, which Google Cloud provides. Only a few online cloud service providers offer that, and Google Cloud Speech-to-Text is one of them.

ChatGPT's Whisper model does not provide speaker diarization, and we need to implement that in-house. Google Cloud provides that on their end, so it's a paid service that we can enable through specific APIs and our library. That's a major part where we can identify which specific speaker is speaking in the calls.

The major challenge with Google Cloud Speech-to-Text is that not every call is clear. Our representative may be in a silent environment, but the client can be anywhere. We need to manage background noise for all calls, so handling audio file clarity is a challenge.

Sometimes, speaker diarization is affected, leading to incorrect speaker identification. For instance, if someone is speaking, it might attribute that incorrectly to another person. These challenges require an AI to analyze the complete output given by Google Cloud Speech-to-Text and correct it properly.

The output might have incorrect grammar, especially since many of our clients speak Spanish. Sometimes, the transcription will be in English when it should be in Spanish, and the AI can help correct that as long as we provide the right context.

A crucial update would be to autocorrect the output from Google Cloud Speech-to-Text because irrelevant words occasionally appear. In our integration, we handle the output with an AI to correct grammar mistakes and issues. If that feature were integrated into Google Cloud Speech-to-Text, it would remove the need for a third-party process. We could simply call Google Cloud Speech-to-Text without having to format or correct the output separately.

I have been working with it for one year, from the process of initial R&D to deployment, and we are still providing support.

We haven't initiated that specific functionality for us yet.

The stability and performance of Google Cloud Speech-to-Text are decent; however, even though it provides speaker diarization, the response is slower than that of OpenAI's Whisper model. OpenAI offers it very fast, but we require speaker diarization, which is why we still use Google Cloud Speech-to-Text.

It is easily scalable.

I have worked with OpenAI's Whisper. It does not provide speaker diarization, which is the main reason we switched to Google Cloud Speech-to-Text. We initially went with OpenAI and then made the switch to Google Cloud Speech-to-Text.

The initial setup process is similar to all services Google provides. We need to access the developer console, create service keys, and provide the required IAM roles. That is consistent with all the operations across the apps available on Google Cloud.

The pricing of Google Cloud Speech-to-Text is indeed higher than that of OpenAI, but the end goal was speaker diarization, so we held on to that.

My advice is that in the final process of the pipeline, it's best to summarize the output with AI. You cannot show the direct output from Google Cloud Speech-to-Text unless it is crystal clear. Even with clear audio, it may have grammar issues, so using an AI service to get a perfect response to show to end users is better.

I would suggest Google Cloud Speech-to-Text to others, primarily for the speaker diarization feature. If the requirement is to get the specific speaker from audio files with multiple speakers involved, then this is a great option.

We utilize real-time speech recognition in Google Cloud Speech-to-Text. We have integrated the service with 3CX and SugarCRM, where we get the call recordings.

Once everything is set up, it does not require maintenance unless new features are requested by the client. Maintenance is minimal unless issues arise with the audio files, in which case we need to format the code. For instance, we may need to manage background noise, which we have to handle in-house or custom.

I rate this solution 7 out of 10.

I have used Google Cloud Speech-to-Text in many hackathons during college and university level.

We use Google Cloud Speech-to-Text when making chatbots that take text as input, but require a feature that converts speech into text before uploading to the chatbot.

In one project, I created a speech quality evaluator similar to those used in international exams such as IELTS. I was tasked to create software that records voice in real-time, converts it into text, and checks the quality of speech. It analyzes how frequently someone hesitates, checks pronunciation accuracy, and other factors.

In another project, I built an inventory management system controlled entirely by audio. When given a voice command, it understands the instruction, processes it, and makes corresponding changes in the inventory.

The main advantage of Google Cloud Speech-to-Text is its accuracy. Even with background noise, it perfectly translates and converts speech into text.

Google Cloud Speech-to-Text handles different accents and dialects effectively. We have tested it with various accents, and it performs excellently.

It provides accurate real-time translation and transcription, delivering results that fully satisfied my requirements.

Given the numerous accents and dialects in India, Google Cloud Speech-to-Text could improve its handling of Indian accents. When someone speaking imperfect English uses the software, it would be beneficial if the translation could autocorrect missing words or elements that the AI identifies as necessary to perfect the English translation.

I have limited experience with Google Cloud Speech-to-Text and am only familiar with its translation capabilities.

There is no need for maintenance. Google Cloud Speech-to-Text functions perfectly. Even with updates from the service provider, we don't need to make any changes. It works perfectly fine in 90% of cases.

Google Cloud Speech-to-Text is very stable.

I haven't attempted to scale Google Cloud Speech-to-Text, but I believe it would maintain its performance level even when scaled up.

I haven't had much interaction with customer service, but for my project purposes, Google Cloud Speech-to-Text works perfectly fine.

Positive

The setup process for Google Cloud Speech-to-Text was very easy and straightforward to connect.

When scaling Google Cloud Speech-to-Text for public use, it may incur some charges, which is reasonable for a service. It would be beneficial if a free version were available for students who want to create projects.

The pricing is reasonable and not excessive.

I have worked with an open-source solution, though I cannot recall its name. I tried open-source alternatives because Google Cloud Speech-to-Text charges for long-term use. However, the open-source solution didn't provide the accurate results I needed, particularly with certain accents. This led me to continue using Google Cloud Speech-to-Text.

I would truly recommend trying Google Cloud Speech-to-Text and experiencing it firsthand.

The implementation is simple, and the outputs are very accurate and crisp. First-time users should definitely choose Google Cloud Speech-to-Text.

On a scale of 1-10, I rate this solution a 9.

In my internship over the last few months, I used a few of the Google Cloud tools, including Google Cloud Speech-to-Text, Text-to-Speech, and Dialogflow to create a chatbot.

While creating a chatbot, which was basically a voice bot, I used Google Cloud Speech-to-Text in it, but I did not use it separately.

Google Cloud Speech-to-Text is used to translate when a user provides input in the form of a voice or audio, converting that audio into text. Similarly, Google Text-to-Speech is used to respond and create a conversational chatbot or flow.

During the time I used Google Cloud Speech-to-Text, it was very impactful to the organization as it made our tasks much easier to perform.

My reasoning for the six or seven rating is that I would ask Google to enhance their platform's vulnerability and features to reduce the need for users to authenticate themselves repeatedly to perform specific tasks. They should also enhance the capabilities of Google Cloud and Cloud platforms to provide dynamic APIs for performing tasks.

I used Google Cloud Speech-to-Text for a week.

Deploying Google Cloud Speech-to-Text was easy for me because I just had to install or enable the API and use it in my function. However, we need credentials to use it, and if we do not have valid credentials, it is very difficult to get them again and to use the service.

Initially, I faced some issues with creating the service accounts or generating the API keys, but as I became more familiar with the Google Cloud platform APIs, it was easy for me to continue with others as well.

I did not face any drawbacks with Google Cloud Speech-to-Text. What I observed is that it can be used on various platforms, but since it is a paid service, it is very difficult to access if a user does not have the credentials. Also, we have to create the API keys and secret keys repeatedly to maintain authentication and privacy.

The scalability of the Google Cloud Speech-to-Text API is great; it is very flexible to use, and it can even understand words spoken by people who are not literate, which illustrates its great scalability and flexibility.

I have not contacted their support because I was working under a team that managed the membership and support, so I was not the one handling those aspects.

Neutral

We were customers and took the membership of the Google Cloud platform for our usage. I do not have information about the pricing or whether it is expensive or cheap.

I would rate Google Cloud Speech-to-Text six or seven out of ten.

For development purposes, my company uses Python in the back end, like the FastAPI framework, and then we utilize the clients of Google Cloud Platform.

The tool includes some models, such as chirp and telephony models. The chirp model is the most accurate, and Google Cloud Speech-to-Text chirp model is very fast, giving responses much faster than other models.

The tool's telephony model does not produce accurate results. With the telephony model, whenever a phone call occurs, we want to transcribe it live or something, and this is not possible in Module Google since it does not support Cloud Translation V2 API. If we are having, like, pre-recorded calls, it is giving you the transcript, but in some cases, where if the call is on hold for five to ten minutes or fifteen minutes or something, whatever we spoke about is not getting recorded, which is an issue.

I have been using Google Cloud Speech-to-Text for a year. I am just a user of the tool.

It is a stable solution.

Around ten people in my company use the tool.

I have not contacted technical support for the solution.

The product's initial setup phase is very easy. The tool includes documentation that explains deployment and setup.

The solution is not deployed in my company, but some of our clients use it for transcribing-related features.

The tool's cost is also very low. The tool is cheaply priced. It charges around 0.13 INR per call with a duration of five minutes.

Compared to the other solutions on the market, my company decided to go with Google since we had some use cases associated with it in our organization related to Google Voice Bot. My company decided to go with the tech stack from Google itself. My company has some other clients who require Google and are on Google Cloud Platform. They are not willing to leave Google Cloud Platform.

Speaking about the tool's multi-language support, I can say that Google supports more languages than any other cloud provider.

I have not experienced any difficulties or challenges integrating Google Cloud Speech-to-Text into our company's workflow.

I would suggest others choose the model correctly. For example, you must use a telephony model whenever it is a phone call or something that has been recorded. You can just go to the console and create it first, and then you'll have the entire code on the right side so that you can directly use it in your workflow.

The tool is easy to learn.

Considering that the tool is not accurate when it comes to native language, especially if you are going for some regional languages in India where there are more than 100 languages, I feel that the tool doesn't support regional languages, but it supports the most widely spoken languages, so only certain areas are accurate. If the call has been placed on hold, there are some deviations.

I rate the tool a seven out of ten.

Google Cloud Speech-to-Text can conduct quick and fast searches and does quick and fast dialogues with things like ChatGPT. It is also good for dictating documents sometimes. You could dictate a bunch of stuff, and then you can get ChatGPT or something to clean it up.

The one thing that I find is when I often use specialized terms, and the solution doesn't know them. I'm not sure if there's an easy way out of this, so I usually have to delete the text to correct something. That would be good if I could just tell it to change the spelling or something like that, and it'd be smart enough to figure that out.

I have experience with Google Cloud Speech-to-Text.

I have never contacted Google's technical support team. Usually, I find the information myself. I don't really use their support since they already offer a lot of articles. I use the articles. I find the articles pretty good. For example, if I'm searching for something and have to know how to set something up or access it so I can Google. After that, the instructions are clearly available on Google. But I've never actually had to talk to a person from Google. Maybe I have spoken to them once or twice, long ago.

I didn't really switch over to Google. I just use Google Cloud Speech-to-Text in whatever window I'm working on, so if it's in Chrome, I use it there. If it's in ChatGPT, I use it there, or if it's in Word, I use it there. I use Google Cloud Speech-to-Text on my Android phone, which always works well.

Speaking about the setup, I just need to click the microphone button on my computer to use it.

Cost-wise, I would say it is all-inclusive in the payment made to Google.

I wouldn't say it's an excellent tool, and neither would I say it's convenient a lot of the time.

I rate the overall product a seven out of ten.

We need IT corporate chatbots to help new people with some things. When you are new in a company, you need a lot of things, such as access. We want to make a chat for auto consulting. You can say, "Oh, hello, I'm new, I need an account in GitHub, please, and GitLab." And we want to do the chatbot with integration with all the systems for may access another common task for users automatically in the chatbot. 

We want to make a chatbot to resolve common problems for people. In that way, we will save time as it will allow support to help with other problems, harder problems. You can resolve common problems with a bot and non-common problems with humans.

Creating bots helps our IT team save time. It allows them to work on harder issues.

They make onboarding easier.

The solution is stable.

We've found the solution scales well.

The language is good in all translation services. Auto-detect in the language is a good feature.

The multilanguage support for the chatbot needs to be better.

If you incorporate the translation natively in a chatbot, maybe you can do a chatbot in English and automatically have the same chatbot, however, the same chatbot needs to be in all other languages as the translation services of Google is good. That translation is not bad for technical people. If they can get a multilanguage chatbot, it would be ideal.

I've been using the solution for two to three years. 

The stability of the solution is good. There are no bugs or glitches. It doesn't crash or freeze. It's reliable in general. 

The product is scalable. If a company needs to expand it, it can do so.

I don't have information about the billing. It's not my interest. I only work on implementations and tasks. The cost of the cloud is not a problem, according to the higher-ups, however. We do analyze the cost for the implementation of new things. However, for other things, we are not interested in seeing the billing. It's not my area, the billing.

I'd rate the solution a seven out of ten.

Google Cloud Speech-to-Text helps to keep my team more productive.

Google Cloud Speech-to-Text's price could be improved. Google Cloud Speech-to-Text's trial experience could be improved by adding some extra minutes in the trial version.

I have been using Google Cloud Speech-to-Text for some few months.

Google Cloud Speech-to-Text is a stable solution, and I don't have problems with connectivity.

Overall, I rate Google Cloud Speech-to-Text a ten out of ten.