PILOT - The Wireless Translator Earbuds

[TheSkeptic]

So this is about the recently posted Indiegogo campaign about the wireless translator earbuds which at the time of writing this post has gotten $1,843,774 USD of funding in just 5 days - Wow deserved. With the early bird version costing around $180 USD and the final price being projected around $200 USD.

There are a couple of things about this project that just cry out crowdfunding marketing garbage on this one but since i'm kinda lazy to write an eloquent paragraph to express my arguments, a bullet list will have to do
Here is the link to the crowdfunding page: https://www.indiegogo.com/projects/meet-the-pilot-smart-earpiece-language-translator--2#/

Now not all the points i give will be about the actual validity of the project, it can also just be about anything i've found dodgy on this marketing campaign.

• At 0:00 - 0:05 of the video, just a hearty cringe
• At 1:33 of the video, is that a Arduino Due with an Atmel SAM3X8E CPU? And a random assortment of Chinese modules on a breadboard which seem unconnected to anything meaningful?
• At 0:45 of the video, ahh the Ol' sound overlay trick gets em everytime
• My opinion but, the way that this campaign follows the crowdfunding marketing formula so perfectly with all the logos and pictures just reminds me of projects like the Fontus (its all about marketing!)
• The way that they don't talk about overcoming limitations such as background noise, and how they leave any negative aspect of the project all the way at the bottom of the page
• I'm sure a random stranger would love having someone who they cant understand demand they to put a mysterious piece of tech in their ear - PILOT better ship with some wipes 

Also in my attachment is a jpg of a picture that is on the crowdfunding page, and again we see the Ardunio Due (this time with the Adafruit logo) along side some other pieces of mystery tech. This makes me wonder that some of that tech is probably 'marketing tech' as in 'zomfg look at teh chips n tech mus bee expnsive' i mean, probably all the processing is done on the phone using their app (because it connects to Babblefish to retrieve the translations) and i hardly think a 84Mhz CPU would add much to the compute power of a modern phone. - yes sorry there is just something about crowdfunding marketing that makes my blood boil.

I have to admit though i still do believe they will come out with this project, its just that i think people will learn the limitations and problems of this device when they get it (alittle to late for my tastes)

[edy]

Before even looking at this campaign... when you said $180 USD now for crowd-funding for a $200 USD retail cost I immediately thought, "Why be on the burning edge to save $20?". Knowing how risky IndieGogo campaigns tend to be, I'd hold back my money and gladly pay $20 more when I have something tangible, reviewed and with a money-back guarantee from a retailer. Just not enough of an advantage to back this, in my books, even if it does pan out to be an awesome real product.

EDIT: Ok, I looked at the campaign page, it's 33% off of what is now $199 as the base price since all Early Birds have sold out. Doing the math.... so retail $300, so 33% off means a third off... brings it to $200. Is that right? Worth $100 savings to jump on this device which is ALREADY way over-funded (so it is happening, no need to feel guilty that not giving your money caused the campaign to fail). At this point, any further money is only people "buying" the product because they want a discount. If you bought before they reached goal, you could argue your money is seeding the project and needed to help it achieve goal. But now, it's just because people want to save $100.... assuming it is made into a final product that is available, you should be able to eventually pick one up.

Let's see.....

1. Input sound (from phone or earpiece?)... then Siri or Google translates speech to text
2. Text translation (google translate or other API's) language 1 to language 2
3. Text to speech synthesizer (Google API) to send sound to earbud via BlueTooth

All bundled into a nice package. Truth is, the *ENTIRE* thing can be done with your phone.... the earpiece is a misdirected fluff..... It is all APP-BASED. The earpiece would simply be a blutooth device with possibly integrated microphone (although that would be a bad place to put it). Why do you need it? Just make an APP that does all of it.

As far as pictures of Arduinos and other boards next to the earpiece... It's all B.S., The entire system is based on your phone and cloud-based connection, using existing speech to text, translation and text to speech services. They just chained the inputs and outputs. 3 separate steps, done in their software, one after the other, sequentially, using cloud-based services or a local software-based speech synthesizer API.

I SEE WHAT IS GOING ON HERE:

With over $1,000,000 they could make the the software.... But they can't SELL you software for $200-300 a piece. Instead, they "bundle" in the earpieces so you feel there is something more tangible you are buying. And all that "fluff" showing PCB's and circuits.... That's all marketing B.S. because they even say in the video it is all done in the app and the earpieces are likely simply bluetooth devices which can be obtained from a Chinese OEM under their design specs to look different. As far as the latency... I'd be more concerned about that, but even if it was fast, I'd like to know how people are willing to pay $200-300 for an app that does this.

Questions:

Why do they need so much processing in the earpiece? If everything is being done by the phone? Earpiece 1 and Earpiece 2 would simply be BlueTooth linked to the phone. How would the app or earpieces track who is saying what, to channel the output translated speech to the other earpiece? Would it compare microphone loudness levels to determine who is doing the speaking, then push the translated output to the other earpiece, and vice versa? So if microphone 1 on earpiece 1 is louder signal than microphone 2 on earpiece 2, then person 1 is doing the talking, and the output would be sent by the app to earpiece 2. If microphone 2 is louder than microphone 1, app assumes it is person 2 doing the talking and speech gets sent to earpiece 1 after translation. Or, the app most likely allows you to set the languages of the 2 speakers and who wears what earpiece (1 or 2) and that information alone is enough to do it.... NO need to "compare" audio levels between mics. So what is all the extra hardware on the earpieces for, and the ARM processor, and everything else?

[edavid]

Backers also get unlimited language addons, where retail buyers supposedly have to pay per language.

They claim to be doing noise cancellation, so maybe they use the ARM for that.
Or, maybe they just don't have the production volume for a dedicated earpiece IC, so they need the ARM to drive the Bluetooth chip.

[edy]

I think it's a cool idea, but if it is relying mostly on the phone for the processing, and people are purchasing language add-ons, I am assuming it will be locally processed. It does not sound like they will use the cloud or data plan. Otherwise why would you need to "purchase" a download library for a language? Wouldn't translation and dictionaries all be included in the cloud?

If they want to optimize it, perhaps instead of sound being transmitted to the phone from the earpiece, they will process the sound in the earpiece with the ARM chip to convert to phonetics/text and pass that on to the phone. The phone would then do the conversion, send the text back, and the earpiece may use the ARM chip to do speech synthesis.

If translating say Spanish to Russian and vice-versa, they may use an intermediate language like English as the center of the hub. For example, they would give you "Spanish <--> English" and "English <--> Russian" dictionaries and then translate through English as an intermediary language.

It'll be interesting to see how they pull this off.

[edavid]

I think it's a cool idea, but if it is relying mostly on the phone for the processing, and people are purchasing language add-ons, I am assuming it will be locally processed. It does not sound like they will use the cloud or data plan. Otherwise why would you need to "purchase" a download library for a language? Wouldn't translation and dictionaries all be included in the cloud?

They say they will do a cloud version first, local version later.  That doesn't mean they can't charge to activate extra languages.

Anyway, it's hard to believe they have their own translation engine.  The great thing about running a scam is that you can promise whatever you want, without worrying about making it work.

[Kalidor]

edy is right, it's basically this https://igg.me/at/phazon/x/10280792 plus software, interestingly it has the same price $180 now, $200 retail.
IMO it's Sigmo on steroids https://igg.me/at/sigmo/x/10280792 (scam from Russia with fake location San Francisco)

And as IGG no longer shows the bank account country scamming on IGG got again a bit easier.