Questions tagged with Amazon Polly

I recently tried the Polly console set to Neural and SSML and keyed in some text and tried several voices with the Whisper action and it works for the Mathew voice but does not work for any other one. If its set to Normal instead of Neural the whisper does work for other voices but does not sound good. Is there a reason the whisper only works for that one voice using the Neural setting?

Hi All, I have integrated Polly into my WordPress blog and so far so good. The only issue I have is that it reads everything and I’d like to leave out certain things like twitter embeds etc. I’ve seen that you can insert SSML tags in to the content to do this but this is where I’m completely clueless. Am I inserting this into WordPress? If so, how and where? Or is it something I need to do in the Polly Dashboard (where the file lives in the S3 bucket perhaps)? Thanks in advance for any pointers.

Hi, I'm trying to put a lexicon in Amazon Polly console in order to censor some stopwords. My .xml file is accepted, but there is no result with the spoken input. This is my script <?xml version="1.0" encoding="UTF-8"?> <lexicon version="1.0" xmlns="http://www.w3.org/2005/01/pronunciation-lexicon" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:schemaLocation="http://www.w3.org/2005/01/pronunciation-lexicon http://www.w3.org/TR/2007/CR-pronunciation-lexicon-20071212/pls.xsd" alphabet="ipa" xml:lang="it-IT"> <lexeme> <grapheme>parola uno</grapheme> <alias></alias> </lexeme> <lexeme> <grapheme>parola due</grapheme> <alias></alias> </lexeme> <lexeme> <grapheme>parola tre</grapheme> <alias></alias> </lexeme> <lexeme> <grapheme>parola quattro</grapheme> <alias></alias> </lexeme> </lexicon> ******* I can't understand the error, Thank you very much if you can help me!

We are trying to setup AWS Polly to use inside a Unity project. Currently we are getting this error: "Invalid identity pool configuration. Check assigned IAM roles for this pool" We have added Polly permissions to the IAM (find screenshots below): https://ibb.co/ThL6mcc https://ibb.co/kqcSTP3 And set up an identity, with unauthorised access, pointing to the IAM containing Polly permissions (find screenshot below). https://ibb.co/TthT7xs Before we linked Identity to the IAM, the error we were getting was pointing towards a lack of permissions for Polly, so it appears that code on the Unity side is working as expected. What are the steps we need to take in order to set up the identity pool configuration to work with unauthorised access? The project we are working on is a site based VR experience, and we are managing all of the machines ourselves, so we have no need for user authentication.

Amazon Polly question. I have a sentence, such as "it will be OK." When I change the script to "It will be <say-as interpret-as="characters">OK</say-as>. " or to "It will be O K. " The output sounds muffled for that sentence only, then goes back to how it sounded in the rest of the script. I'm using: * Neural engine * English, US language * Matthew, Male voice * SSML turned on The same thing happens when I use any 'interpret as=' tags Is anyone else running into this issue? It's very distracting and I can't figure out why it is happening. Thanks for any help you can give me.

Hello we upgraded our php sdk and now Polly stopped working. I have searched around fourms if it supports php latest version and could not find any answer. Can you clarify if polly works with php version 8.0.17? PHP SDK 3

I find a weird issue with the second sentence while using Mathew in Neural voice. The entire Tone of Mathew changes when starting the second sentence starting "In the code... SSML Text below: <speak><amazon:domain name="conversational"><prosody rate="85%"><break time="1s" /> <p>Process all gives the advantage in avoiding the mistakes that may occur due to accidental missing of any signal listing, renaming or any addition of new signal.</p> <p>In the code on the slide, process all implies process of A, B, SEL as A , B and SEL are the inputs red by the process MUX.</p> <break time="1s" /></prosody></amazon:domain></speak>

I am trying to set up the proper policy json for IAM programmatic access, but the following runs into the below error ... I have tried it also without sid and with Resource as an array. ```{ "Version": "2012-10-17", "Statement": { "Effect": "Allow", "Sid": "AllowAllPollyActions", "Action": [ "polly:*" ], "Resource": "*" }} ``` `is not authorized to perform: polly:SynthesizeSpeech with an explicit deny in an identity-based policy`

The Hebrew version of Polly is all I need for my app. will be an update for that soon...?

If so, is there a roadmap? I see that Alexa supports <emotion> and was curious if Polly would implement that in the future.

Hi, We're using Amazon Polly ( boto3 Python client ) for production TTS. We have a "fixed" test text that we're using to monitor our system. Usually this text takes a few seconds to be completed. I've noticed that sometimes it takes more than 1 minute (!!) to complete successfully. As we're using it as part of an IVR system where user is waiting on a phone, I would like to know if there is a way to speed it up or to prioritize such operations as well as to know how to audit/log/troubleshoot such operations (via cloudwatch or any other method). thanks in advanced !!

Hello, I am trying to resolve the dreaded SSL Certificate error: ....[SSL: CERTIFICATE_VERIFY_FAILED] certificate verify failed: unable to get local issuer certificate when trying to validate a Polly endpoint, such as: https://polly.us-east-1.amazonaws.com/v1/speech I am doing this in the CLI The errors are: -- https://polly.us-east-1.amazonaws.com/v1/lexicons/awsLexicon [SSL: CERTIFICATE_VERIFY_FAILED] certificate verify failed: unable to get local issuer certificate (_ssl.c:997) --https://polly.us-east-1.amazonaws.com/v1/lexicons/awsLexicon [SSL: CERTIFICATE_VERIFY_FAILED] certificate verify failed: unable to get local issuer certificate (_ssl.c:1125) Among my steps tried so far was to update the certificate: @ CLI -> pip install certifi I verified in site-packages/certifi that 'cacert' is present I'm sure I am missing some step(s). Can anyone see what I should do next?

Hi Team, As we have built one contact flow where we are capturing some digit number from end user like FA1203040 but when we are using the Play prompt Block to playback (SSML type) the user input for confirmation. AWS is prompting the Digit 0 as "oh" or "O" Can anyone help how we can resolve this issue.

Hi team, We have built one custom bot for one of the use case and we are using voice feature of Lex but the Amazon Lex is not recognizing the correct user input every time. Can someone please help me on this Thanks in advance

Hi, I work as a copywriter and I'd like to create some voice posts with TTS. I'd like to know if Polly's synthetic voices have a license that can be extended to the employer (a web site owner in this case), and if Polly's voices can be used to create audio files for third parties. Thank you so much.

Hello. I'd like to turn a text (dialog) into an audio file. Amazon Polly offers a bunch of different languages and voices, but I was not able to find a way to use two different voices (languages) to speak text one after the other to simulate a conversation between two people. Is there any way to do it with Polly?

Here is the SSML in question: <speak><prosody rate="100%" pitch="-21%">India&apos;s massive COVID surge puzzles scientists</prosody></speak> Not that if I remove the pitch attribute it does not fail. Also, the same SSML works fine for Standard voices. Is there a bug? The same SSML worked fine previously for Neural voices but stopped working at some point recently. Any help is appreciated. Edited by: bretwilliam on May 2, 2021 10:26 AM Edited by: bretwilliam on May 2, 2021 10:27 AM

This line of text works in a batch line for CLI --text "<speak><prosody rate='slow'> but this one doesn't --text "<speak><prosody rate='85%'> is there a way to use prosody rate with a % for CLI in Windows? Thanks Edited by: jcwdad on Feb 18, 2021 4:36 PM

When i tried the exercise at https://docs.aws.amazon.com/polly/latest/dg/get-started-cli-exercise.html the process fails with Unknown output type: mp3 I'm using windows syntax: --output-format mp3 ^ Can anyone provide a clue as to why this isn't working? Thanks P.S. I tried mp3 as string "mp3" as well, same error full transcript below, admin command prompt, win 10 C:\WINDOWS\system32>aws polly synthesize-speech ^ More? --output-format mp3 ^ More? --voice-id Joanna ^ More? --text "Hello" ^ More? hello.mp3 Unknown output type: mp3 Edited by: jcwdad on Feb 12, 2021 3:41 PM

Amazon Polly Matthew Standard pronounces "live" incorrectly when in the sentence "You did not change the way that you live" none of the role attributes make any difference.

Hi, I'd need some help, please. If I put this in my console <speak> <amazon:domain name="conversational"> <prosody pitch="+10%">Hi, my name is Federico and I'm a student.</prosody> </amazon:domain> </speak> the answer is that it is impossible to load. It doesn't tell me the syntax is wrong. My goal is to raise the tone of voice. I'm using Matthew's neural voice. Thank you very much.

Good morning, I don't have programming compentencies, but I'm very keen in tts. I use Polly with SSML. I'm considering to start a web developer course and would like to ask you if php7 can be used for Amazon Polly with good results. Moreover, what is the most useful programming language I could study to get the best out of text to speech? Thank you so much.

Hi, could you please suggest me how to change the speech speed with ssml from Polly Console? Thank you.

I am using the neural mode in Polly. The neural mode does not allow for change in pitch. The only alternative I see is to increase the rate. However, I have tried this, and it sounds very unnatural. In order to emphasize words, I decrease the rate slightly, and increase the volume slightly. A slightly higher change in pitch would certainly help.

Hi Guys, I am new to Polly and am trying to use it to create Newscaster voice for a fun video. I am in North East and Amazon marks me as in Ohio region. I selected Mathew as the voice and copy pasted the script for the Newscaster style from Amazon polly's resource page in the SSML box. I got the error: "The input text contains invalid SSML syntax. Fix the SSML syntax, and then try again". I realized that it works only for the Neural voice. However, I do not see the option to select the Engine "Neural" on my end. The code that I used was exactly the same as in the resource page. (pasting it below). I was wondering if anyone can help me. I am a free account holder as I am doing this only for a colleague's retirement video and don't have too much insight into this. The code below was copy pasted from the developer guide <speak> <amazon:domain name="news"> From the Tuesday, April 16th, 1912 edition of The Guardian newspaper: The maiden voyage of the White Star liner Titanic, the largest ship ever launched, has ended in disaster. The Titanic started her trip from Southampton for New York on Wednesday. Late on Sunday night she struck an iceberg off the Grand Banks of Newfoundland. By wireless telegraphy she sent out signals of distress, and several liners were near enough to catch and respond to the call. </amazon:domain> </speak>

In a sentence like this one: The / father / and / the / mother. I would like Polly to read it without pausing. Right now it makes a pause after each slash symbol. Is there a command (e.g. skip symbol when pronouncing) so that Polly reads it as: The father and the mother. *This does not happen in Spanish, French and German. Edited by: maicart on Aug 4, 2020 2:20 AM

Hi. I'm trying to generate TTS through the console here: https://console.aws.amazon.com/polly/home/SynthesizeSpeech?region=us-east-1 I am using the ssml tab with the following text: <speak> <break time="5s"/> <emphasis level="strong">Hi!</emphasis> My name is Kendra. I will be one of your guides today. Welcome to the Hall of Heroes! </speak> I plays fine without the emphasis tag, but as soon as I add it I get "The input text contains invalid SSML syntax. Fix the SSML syntax, and then try again." I've even tried copy/paste with the default text from here: https://docs.aws.amazon.com/polly/latest/dg/supportedtags.html#emphasis-tag Is this a bug, or user error?

Hi, I'm using polly for a low number of SynthesizeSpeech requests. For a possible bug that could somehow send a very high amount of requests in a loop, there is a way to block this service following an out of the ordinary use of resources, maybe with some policies (I can't find a target function that's right for me to work with a cloudwatch alarm), or maybe like other aws services it is supervised and freezes following unusual activities? Thank you in advance.

I just began testing with the StartSpeechSynthesisTask api, and I noticed that every task I started is displayed in the list returned by ListSpeechSynthesisTasks. Am I supposed to delete the tasks when they are no longer needed, and if so how? I am deleting the audio files written to S3, but I don't know how to manage the tasks. Thanks!

I'm not sure if I'm doing things incorrectly, but sometimes it takes almost precisely a minute (60-61 secs) to get a response back. Most of the time it is < 1 sec, but sometimes it gets to 60.x secs - never longer, never shorter, always around that 1 minute mark. Some background: I'm using it in a Django application, and in a helper file, towards the top I initiate things: ``` session = Session( region_name="us-east-1", aws_access_key_id=... aws_secret_access_key=... ) polly = session.client("polly") ``` Then down in a function (in the helper file): ``` def some_function(): time1 = time.time() response = polly.synthesize_speech( Text=<....>, OutputFormat="mp3", VoiceId="Justin", TextType="ssml" ) time2 = time.time() ``` Then in a views file (since this is used in Django), it will call on "some_function" when requests arrive. As mentioned, time2 - time1 is usually < 1 sec, but from time to time, it gets to 60-61 secs... I initially thought it was some sort of rate limiting, but I looked up the limits for AWS polly and it seems like I am far away from any throttling limits. I'm out of ideas, so any inspiration is much welcome! Thanks in advance! Also, not sure if relevant, but the Django app is hosted on Azure - would that cause any random in occurrence, but consistent in duration delays?

It seems that there is some inconsistencies in how the Amazon Polly Console (and files processed through the CLI or API) is handling non-coded pauses when using <break> or where the </speak> is placed in the file. Here are some examples to illustrate the problem - it doesn't matter if a neural voice or standard voice is used. Example 1: ``` <speak>Article 1 Video provides a powerful way to help you prove your point. When you click Online Video, you can paste in the embed code for the video you want to add. You can also type a keyword to search online for the video that best fits your document. </speak >``` A pause is inserted after Article 1 since there is a paragraph break which is the expected result. Example 2: ``` <speak>Article 1 Video provides a powerful way to help you prove your point. When you click Online Video <break time = ".3s"/> you can paste in the embed code for the video you want to add. You can also type a keyword to search online for the video that best fits your document. </speak >``` A <break> is inserted in place of the comma. When this occurs, the paragraph break after Article 1 is ignored so there is no pause between Article 1 and Video. This is unexpected. Example 3: ``` <speak>Article 1 Video provides a powerful way to help you prove your point. When you click Online Video, you can paste in the embed code for the video you want to add. You can also type a keyword to search online for the video that best fits your document.</speak >``` Similar to Example 1, a pause **should ** occur after Article 1 since there is a paragraph break. This does not occur. Notice that there is no line break before </speak>. Any clarification would be most helpful as results are inconsistent now and it would be time consuming to have to insert `<p></p>` throughout to get expected results. Edited by: vabtm on Jul 3, 2020 1:57 PM

I was trying to add voice for my game and I found Amazon Polly, and I'd like to use the voice that the service outputs. However, after reading documents such as ToS, I cannot find a definitive answer on what you can do with the recordings. Suppose that I generated a voice recording using Amazon Polly and downloaded it as an audio file. What I want to do with this recording is: 1. to add audio effects to the voice recording and 2. to merge Polly recordings with other audio to reduce file count. Is it allowed to make 'direct' derivative works (I'm not sure if that is the right word) from the recording I made with Polly like this? Or do we always have to use the recording exactly as what Polly generated? (Note: I have all the required licenses/copyright to use the text itself to generate the recording and the audio I want to merge with the recording.)

I'm trying to make Polly say 'live' (as in 'Live from New York!') using the role tag. It works fine when live is at the beginning of the sentence. But if the word is buried in the middle of the sentence, the pronunciation changes. I'm using the Windows CIL interface and Neural voices. My region is set to us-east-1. I've also tried various workarounds using the phoneme tag, but I can't get anything to work. I think it's because I need to escape characters in the ph= option (like stress marks in ipa, and <? in x-samp), but \ doesn't seem to be working as an escape in CIL. AWS keeps returning "Invalid SSML Request" when i attempt to escape anything. Here are the role commands I'm sending: ``` aws polly synthesize-speech --output-format mp3 --engine neural --voice-id Joanna --text "<speak><amazon:domain name='conversational'><prosody rate='97%'><w role='amazon:SENSE_0'>Live</w> from New York, it's Saturday Night!<break time='.1s'/></prosody></amazon:domain></speak>" --text-type ssml good.mp3 aws polly synthesize-speech --output-format mp3 --engine neural --voice-id Joanna --text "<speak><amazon:domain name='conversational'><prosody rate='97%'>When sharing compatible files, select Request control. <break time='.35s'/>This allows you to <w role='amazon:SENSE_0'>live</w> edit the open documents during your video session.<break time='.1s'/></prosody></amazon:domain></speak>" --text-type ssml bad.mp3 ``` And here are the various phonem tags I've tried: ``` aws polly synthesize-speech --output-format mp3 --engine neural --voice-id Joanna --text "<speak><amazon:domain name='conversational'><prosody rate='97%'>When sharing compatible files, select Request control. <break time='.35s'/>This allows you to <phoneme alphabet='ipa' ph='līv'>live</phoneme> edit the open documents during your video session.<break time='.1s'/></prosody></amazon:domain></speak>" --text-type ssml live1.mp3 aws polly synthesize-speech --output-format mp3 --engine neural --voice-id Joanna --text "<speak><amazon:domain name='conversational'><prosody rate='97%'>When sharing compatible files, select Request control. <break time='.35s'/>This allows you to <phoneme alphabet='ipa' ph='laɪv'>live</phoneme> edit the open documents during your video session.<break time='.1s'/></prosody></amazon:domain></speak>" --text-type ssml live2.mp3 aws polly synthesize-speech --output-format mp3 --engine neural --voice-id Joanna --text "<speak><amazon:domain name='conversational'><prosody rate='97%'>When sharing compatible files, select Request control. <break time='.35s'/>This allows you to <phoneme alphabet='ipa' ph='lɪv'>live</phoneme> edit the open documents during your video session.<break time='.1s'/></prosody></amazon:domain></speak>" --text-type ssml live3.mp3 aws polly synthesize-speech --output-format mp3 --engine neural --voice-id Joanna --text "<speak><amazon:domain name='conversational'><prosody rate='97%'>When sharing compatible files, select Request control. <break time='.35s'/>This allows you to <phoneme alphabet='ipa' ph='\ˈlɪv'>live</phoneme> edit the open documents during your video session.<break time='.1s'/></prosody></amazon:domain></speak>" --text-type ssml live4.mp3 aws polly synthesize-speech --output-format mp3 --engine neural --voice-id Joanna --text "<speak><amazon:domain name='conversational'><prosody rate='97%'>When sharing compatible files, select Request control. <break time='.35s'/>This allows you to <phoneme alphabet='ipa' ph='"lɪv'>live</phoneme> edit the open documents during your video session.<break time='.1s'/></prosody></amazon:domain></speak>" --text-type ssml live5.mp3 aws polly synthesize-speech --output-format mp3 --engine neural --voice-id Joanna --text "<speak><amazon:domain name='conversational'><prosody rate='97%'>When sharing compatible files, select Request control. <break time='.35s'/>This allows you to <phoneme alphabet='x-sampa' ph='"lIv'>live</phoneme> edit the open documents during your video session.<break time='.1s'/></prosody></amazon:domain></speak>" --text-type ssml live6.mp3 aws polly synthesize-speech --output-format mp3 --engine neural --voice-id Joanna --text "<speak><amazon:domain name='conversational'><prosody rate='97%'>When sharing compatible files, select Request control. <break time='.35s'/>This allows you to <phoneme alphabet='x-sampa' ph='l<? ī ?>v'>live</phoneme> edit the open documents during your video session.<break time='.1s'/></prosody></amazon:domain></speak>" --text-type ssml live7.mp3 ``` Edited by: sharonhuston on Apr 7, 2020 10:11 AM More in this issue: I'm also getting inconsistent results between the web console and the CIL. Look at these two identical entries. The console says "live" correctly, the command line does not. link:<https://learning.realpage.com/downloads/sharon/cil.png> link:<https://learning.realpage.com/downloads/sharon/console.png> Edited by: sharonhuston on Apr 7, 2020 1:29 PM

Please add support for a Hebrew Polly voice. It would be much appreciated. Thank you.

I am trying to get the Amazon Polly Newscaster Voice to work. My text has 13,000 characters, so as per the AWS console instruction, my "task must be saved to an S3 bucket", as it's longer than 3,000 characters. I have created an s3 bucket and have configured my s3 task settings to output to the s3 bucket. But when I click on "synthesize to s3", I receive this message - "The input text contains invalid SSML syntax. Fix the SSML syntax, and then try again." I know that my SSML syntax is correct as I have done two checks. 1) I reduced my text to below 3,000 characters and clicked "Listen to speech". My text was successfully read out in Newscaster style. 2) I reduced my text to below 3,000 characters and clicked "Synthesize to S3". The MP3 audio file which perfectly read out in Newscaster style was saved in my S3 bucket. This is my SSML syntax <speak>  <amazon:domain name="news">  Lorem ipsum dolor sit amet, consectetur adipiscing elit, sed do eiusmod tempor incididunt ut labore et dolore magna aliqua. </amazon:domain>  </speak> Am I doing something wrong? Or is this a bug? Thanks in advance for any assistance.

Two weeks ago I used Zeina voice to generate the following sentence: ثَلَاَثَةٌ وَعِشْرُوُنَ طِفْلاً مِنَ الصَّفِّ ا انْضَمُّوا إِلَى ثَلَاَثِيِنَ طِفْلاً مِنَ الصَّفِّ ب.‎ and everything went well. Today when I generate Zeina voice for the same sentence I found obvious stress in the first letter in the words "طِفْلاً". I am just mentioning an example but the issue exists in multiple words for sure. Is there any update in the engines or so that caused that? and is there any possibility to use the old engine if so? I the two audio samples for reference but I had an error when I tried to attach them to this message. Thanks in advance.

On this preview page: https://us-east-1.console.aws.amazon.com/polly/home/SynthesizeSpeech I tried the below SSML, and got this warning ---- The input text contains invalid SSML syntax. Fix the SSML syntax, and then try again. --- <speak> <amazon:domain name="news"> Earlier today we rolled out Amazon S3, our reliable, highly scalable, and low-latency data storage service. Using SOAP and REST interfaces, developers can easily store any number of blocks of data in S3. Each block can be up to 5 GB in length, and is associated with a user-defined key and additional key-value metadata pairs. Further, each block is protected by an ACL (Access Control List) allowing the developer to keep the data private, share it for reading, or share it for reading and writing, as desired. The system was designed to provide a data availability factor of 99.99%; all data is transparently stored in multiple locations. </amazon:domain> </speak> Edited by: Gottleib on Aug 9, 2019 5:32 AM

I am trying to follow the documentation to try the neural newscaster style. The docs say that 1. Open the Amazon Polly console at https://console.aws.amazon.com/polly/ 2. On the Text-to-Speech page, for Engine, choose Neural. But as the screenshot I have attached shows there is no place to select "Engine". Any thoughts? ==== Update ====== Figured out the problem I was using us-east-2 instead of us-east-1. When I specify the us-east-1 console then I can see the expected frontend Edited by: cranperry on Aug 3, 2019 9:46 PM

One frustrating thing about using Polly is finding that small changes to a sentence have a dramatic impact on the way words in the sentence are pronounced. For example, consider the following sentences: - Your watch will now be activated. - Your car will now be activated. If you have Polly say this using English - Indian - Raveena, the word "activated" will be pronounced very differently between the two sentences. In fact, the "will now" will also be pronounced a bit differently. Is there anything I can do, aside from fully manually defining a phoneme for the word "activated", to get Polly to use one of the pronounciations instead of the other, regardless of the structure of the sentence? I realize there's a lot of smarts involved in getting a sentence to be spoken, but it's frustrating to know she can say "activated" better than she is in my particular sentence. Thanks.

Hello, We use polly for IVR and in our brazilian calls since 3/19 number 1 is being synthesized as OM and it should be UM. Data, 4, and valor are also being synthesized wrong I also notice that for some regions it gets it well. I test with ricardo pt-br the following text "digite 1 4 valor data" good (oregon) https://s3.us-east-2.amazonaws.com/ipcom-1234/speech_20190321221526364.mp3 bad (ohio) https://s3.us-east-2.amazonaws.com/ipcom-1234/speech_20190321221901729.mp3 sometimes ohio and n virginia produces good results sometimes bad. Please advise. Edited by: panitaxx on Mar 21, 2019 3:26 PM Edited by: panitaxx on Mar 21, 2019 3:33 PM

I'm trying out the Mandarin Chinese voice of Polly, but can't seem to find a way to tell it which Chinese tones to use. There are several words in Chinese that can be pronounced in multiple ways (i.e. 当天 → dàng tiān/dāng tiān), so it would be convenient if Polly could support pinyin input and not just hanzi. I tried using SSML and IPA tone letters, but they seem to be ignored by Polly. As it is now, Polly will sometimes use the wrong tones and I haven't been able to find a workaround. If there is some workaround for this, or plans for supporting tone information, please let me know.