Business earbuds interpreter Timekettle W4 Pro with AI tested: Star Trek's tricorder or the babel fish at last?
Smartphones have recently become quite versatile and powerful everyday computers. Thanks to Google Translate, DeepL and the like, they are also pretty effective translation devices.
Nevertheless, there are still stand-alone interpreting devices, in the case of the Timekettle W4 Pro you pay a hefty 450 Euros (~$476) for it. For that price, you already get a very good mid-range smartphone. Is it worth it? What can the interpreter earbuds do better than the smartphone? Thanks to artificial intelligence, do we have the long-awaited translation functionality of a tricorder from Star Trek? We find out in our test.
Technical features - 40 languages, 93 accents
To clear up one fact right away: The W4 Pro does not replace a smartphone for interpreting, the Timekettle app is an absolute prerequisite for use, this is where the AI functions are first available and even the 1-to-1 conversation is not possible without an app connection.
It is also possible to translate offline, but then "only" 13 languages are available. It is only online that the full capabilities of the interpreters come to light. The W4 Pro can translate into 40 languages and is said to master 93 accents.
The manufacturers pay particular attention to the almost seamless translation speed. The response time is said to be just 0.2 seconds.
Are you a techie who knows how to write? Then join our Team! Wanted:
- News Writer (Romania based)
Details here
The earbuds are important for this. Unlike the Timekettle X1, for example, which had in-ear earbuds, the W4 Pro has an open design, which makes total sense. After all, in-ear headphones can come across as unhygienic, especially in 1-to-1 conversations when you hand one of your two earbuds to the other person. Unlike the W4 Pro, the open design means you don't have to worry about the previous user's earwax sticking to them. This also makes it easier to hear your own voice, which also has advantages. There are three microphones built into each of the two earbuds, which are designed to ensure precise recording of what is being said and filter out unimportant sounds.
In addition to 1-to-1 calls, the earbuds also offer the "Listen and Play" and Media Translation functions in conjunction with the app. With the former, for example, a wearer can follow a meeting in a foreign language with both earbuds. If the wearer wants to speak, they tap one of their earbuds and then their speech is translated and passed on to everyone via the smartphone speakers.
Media translation refers to translations of apps that run on the smartphone. Here, for example, you can have a YouTube video running on your smartphone translated.
In contrast to Google, the manufacturer also advertises with "Privacy-First Translation Earbuds"; all conversations should remain private and not be used. Nevertheless, data can be passed on and the data protection conditions are quite extensive and very general.
Housing and ergonomics - Unusual charging case
The earbuds themselves are mostly made of plastic. They keep in place on the ear, but they do wobble quite a bit if the person wearing them moves their head a bit fast. A small, colored LED indicates when they are being charged or inserted into the charging case. There is no splash protection or other features. The W4 Pro are also quite large, even for over-ear headphones. For comparison, the Marshall Motif II A.N.C. in-ear buds.
The charging case is even larger and somewhat unusual. The outer frame is made of aluminum, the rest is plastic. It can be divided into two halves so that you don't have to hold out the bare earphones to a conversation partner, but can instead hand over the handset in the charging case. However, the size of the box could also lead to the opinion that the W4 Pro are clunky and significantly less mobile than other earbuds. There is a USB-C charging port on the back of the case. An LED indicates the charging status. The case including earbuds weighs 187 grams (~6.6 oz).
Sustainability
Sustainability plays no role in the marketing of the expensive interpreter earbuds; the manufacturer does not provide any information on the materials used.
The earbuds are packaged in a stylish black cardboard box, and a rigid foam is also used inside as insulation.
Operation and functions - AI app handles all translations
As mentioned, the Timekettle app is mandatory to use and you need to create an account the first time you start it. Unfortunately, you won't find the app in the stores, a QR code leads to the download page, which you have to trust. To pair the handsets, the location must also be activated each time the app is restarted.
The handsets themselves only have a single operating function: by tapping on the handsets, you can switch between listening and speaking in "Listen and Play" mode, which is practical as you then don't have to reach for your smartphone, which may be lying on the conference table, to translate your voice to all participants.
In addition to the three translation modes One-on-One, Listen & Play and Media Translation, offline mode can also be activated on the main screen of the app, but media translation is then grayed out. The status of the earbuds is also displayed at the top. Feedback can also be submitted directly via the home screen or customer service can be contacted.
There are also tabs at the bottom, under "Discover" there are a few instructions on the Timekettle products, in the case of the W4 Pro only one, namely how to pair them with the smartphone. The AI tab is certainly exciting. Here you can type in terms and have them translated and even spoken aloud. Memos of past conversations and recordings can also be found here, which can also be summarized. You can also train your own language skills, as there is an AI model with which you can speak in the practice language, the model then makes suggestions for improvement and asks questions to simulate a conversation. In contrast to the 40 languages in the translations, the AI bot only seems to be available in English, Spanish and Chinese.
Practical test - Where is the context?
Sound
Unfortunately, the sound quality of the buds is disappointing for the price. You can clearly tell that the manufacturer did not design headphones for universal use, but rather that they were designed specifically for the purpose of translations. Music does not sound appealing on the W4 Pro. Bass is clearly under-represented, treble and mids don't seem particularly differentiated. The overall result is a tinny, unattractive sound when playing music.
It is also noticeable that adjusting the volume (via the smartphone) affects the balance of the earphones, which should not actually happen. So if we change the volume, it is not lowered or raised equally on the left and right ear, but suddenly the left or right earbud seems louder/quieter compared to the other before.
The earbuds are therefore unlikely to be used for anything other than translating. In addition, you can hear a low background noise on the earbuds without sound, more like a kind of minimal rattling, which is no longer perceptible with sound. For the price, Timekettle could have given the earbuds significantly better audio quality.
The core business of the W4 Pro. If you want to have a bilingual conversation with the other person, you simply hand them one of the two earbuds in the charging cradle, set the two languages in the app and you're ready to go. Of course, we can't make any statements about the many languages, we concentrated on English and German in the test.
However, even here the quality is nowhere near as good as a human interpreter. Pauses in the speaker's speech in the middle of a sentence are a particular problem for the Timekettle system. The AI translator then simply inserts an inappropriate end to the sentence in the middle. If the speaker starts again and completes the sentence, the AI translator starts a new sentence. It's easy to imagine that this can sometimes significantly confuse the meaning of what has been said.
It is also noticeable that the system does not grasp the larger context of the conversation. Although I am testing with my son (he German, I English) and he addresses me as "Papa", the AI also interprets my address to him as "Du" and then again as "Sie" in the course of the conversation. In view of the business context, at least stick to "Sie" consistently.
The quality of other translations is also rather mediocre, which in most cases is due to not recognizing the context. Similar-sounding words such as "Meine" and "Deine" are often confused, which means that what is being said no longer makes sense; context recognition could correct this. Another example is lesser-known words that are then simply translated incorrectly and randomly. The AI was apparently also unable to deduce from the context that an exchange of words was about the trendy sport of slacklining. "Flecklein" is just as little a sport as "Select Lining", although the AI had correctly recognized "Slackline" a minute earlier, but this did not stick, so unfortunately the AI does not learn from the course of the conversation. Recognizing the context would at least not always lead to new word creations.
As the translation voice in the other person's ear is always very delayed and you don't want to talk over the translation, you should stick to certain conversation rules to ensure the best possible translation. These include speaking fluently, but not too quickly and clearly, without pauses for thought, but then pausing after about 2 sentences, clear sentence structures and as little colloquial language as possible. Some rules of conduct are more difficult to adhere to than others, even in a business environment.
Probably the most versatile mode. Here you listen to various sources and receive the translation on your ear and the transcript on your cell phone. It is advertised for use in meetings, but radio or TV broadcasts can also be translated in this way, and other situations are also conceivable. If you want to reverse the roles and speak yourself, all you have to do is tap on the right earpiece or in the app and your own cell phone will translate what the other meeting participants are saying.
On the other hand, this is perhaps also the most error-prone mode, as the system often has to operate in a larger, wider context here, so it may have to cope with multiple, alternating voices, more background noise and speakers who are further away. We also learned from Timekettle that the Listen part in Listen & Play is also the only mode in which the smartphone is responsible for audio recording instead of the earbuds. Recognition is therefore highly model-dependent.
For example, we played a YouTube video through our laptop speakers. While the volume was normal to high for us, the W4 Pro apparently couldn't receive well enough. Only when we raised the volume significantly could the earphones be motivated to translate. It also happens all the time that a translation starts and then stops in the middle of a sentence because the recognition apparently suffers via the smartphone recording (Pixel 7 Pro). Mind you, on my desk with a fairly loud YouTube video (70% laptop volume via internal speakers) without background noise. Even at 100%, the translation often simply does not continue, which cannot really be due to the volume. Overall, the mode is rather unreliable, at least when capturing media being played.
Capturing natural voices works better here, at least without background noise. However, the translation quality sometimes leaves something to be desired. In particular, if the speaker pauses a little during a sentence, this is quickly interpreted as the end of the sentence and the speaker starts again, which sometimes distorts the meaning of what was originally said. Speaking for yourself seems to work really well, what we said ourselves was translated without any problems and if you take breaks, the meeting participants can listen to the translations via the smartphone speaker.
Media Translation is specifically about translating from apps that run on the same smartphone as the Timekettle app.
So if we activate this mode in the Timekettle app and open YouTube (or Zoom etc.) at the same time, for example, we can have the voices in a video translated, and what is said is also transcribed and, if desired, displayed in the original language and translation or only as a translation or not at all. So far so good. The translations also work quite well and quickly because the app is able to capture the audio track directly from the smartphone. The overlay window with the transcriptions can be reduced or enlarged and also freely moved or closed - very good.
However, there are also one or more minor problems here.
- As there is only one volume rocker for all apps on the smartphone, the volume of the original, which unfortunately does not get quieter, and that of the translation overlap, which makes it very difficult to concentrate and listen to the translation channel in particular. Here, the original YouTube video should automatically become quieter.
- Timekettle claims that the translation is largely free of delay. However, this is true at best for the transcription, but not for the audio. The translation audio only starts several transcript lines later. This means that the audio does not coincide with the transcription, which makes it difficult to read along. What is worse, however, is that the translation audio does not match the original video. For example, while our good Alex is talking about a specific laptop model in the video (in English), the German translator's voice is just starting to translate his intro. This makes it much harder to follow a video. We don't hear the goal in the sports match summary in the audio translation until 7 - 10 seconds later.
- If we stop a video, the (audio) translation also stops about 2 seconds later. This makes sense, of course, but in conjunction with the long delay between the original and the audio translation, this can mean that after the end of a video (i.e. the automatic stopping of the video), we no longer receive an audio translation of certain end sections. If we stop a video/audio and then resume it, the audio starts again after the usual delay, but only from the new start; what was not yet translated to the end due to the delay and pause is no longer translated auditorily, but can only be read in the transcription.
Overall, however, the translation works quite well, all tested apps were recognized, including radio plays with the VLC player or the ARD media library. This is very practical and the translation quality is impressive. However, Timekettle still needs to work on the long delay in the translation voice.
Price model with fish
Even after purchasing the interpreter earbuds, additional fees may be incurred, which is due to the fact that manufacturers pay for AI integration, presumably also because of license fees. It is not known exactly which LLM Timekettle is relying on here. Every user gets 300 free minutes of translation per month, after which you have to buy additional minute packages. And how do you pay for them? With "Fish", of course.
Fish probably goes back to the Babelfish, which you can put in your ear in the Douglas Adams novel "The Hitchhiker's Guide to the Galaxy" and then understand all spoken languages. Anyone unfamiliar with this background may find the currency name a little dubious, especially in the business sector.
In any case, 500 additional minutes cost 5 Fish, 1,200 minutes 10 Fish, for 2,000 minutes you pay 15 Fish. And where do you get the "Fish" from? Fish Coins, i.e. the virtual coins, must be purchased by making a deposit via Google's in-app payments via the Google Play Store. Here, 5 Fish currently cost 10.99 Euros (~$11), which is exactly the price you pay for 500 additional minutes.
In addition to the extra minutes, you unfortunately also pay extra for offline packages. The first two are still free via vouchers, after which each offline package costs 5 Fish, i.e. also around 11 Euros (~$11). So if you want English - Spanish (German is only available offline in combination with English and Chinese), you have to pay 11 Euros, for English - French another 11 Euros and so on. You can do it that way, but it doesn't exactly arouse any sympathy.
Pros
Cons
Verdict - No baby fish yet
So how much Babelfish or Tricorder is in the Timekettle W4 Pro? It's cool to be able to communicate with the other person in a foreign language even without knowing the language and when the recognition works well, as in Media Translation mode, you get a great feeling for the vision behind it. But the overall package doesn't quite convince us yet.
The translations do not yet work reliably enough, too often the recognition does not work optimally or is inaccurate. In list mode, this can still be blamed on the smartphone, but the suboptimal recognition and the lack of context recognition are general weak points. The audio delay is also annoying, the earphones are of inferior audio quality and are not suitable for listening to music, but the package is still quite expensive. What's more, additional fees are charged for more intensive use even after purchase. The manufacturer charges extra for offline packages.
The Timekettle interpreter W4 Pro is not yet fully optimized in its current state, especially the recognition, but also speed, translation quality and audio quality show a lot of potential, but are not (yet) at the advertised level. In addition, the package is currently too expensive for what it offers.
Price and availability
The Timekettle W4 Pro is available both on the manufacturer's website as well as on Amazon for $449. However, as mentioned in the test, there are still various potential additional costs after purchase in the form of offline voice packages, additional minutes of use and more.
Transparency
The selection of devices to be reviewed is made by our editorial team. The test sample was given to the author by the manufacturer free of charge for the purposes of review. There was no third-party influence on this review, nor did the manufacturer receive a copy of this review before publication. There was no obligation to publish this review. We never accept compensation or payment in return for our reviews. As an independent media company, Notebookcheck is not subjected to the authority of manufacturers, retailers or publishers.