If I were Microsoft...

Additionally at least Google Voice and Siri draw from a near constant internet connection to leverage a larger pool of data for voice recognition. That is only available on Kinect when using Bing search while connected to the internet, at which point accuracy improves dramatically.

Regards,
SB

Bing search needs to parse any possible spoken word. What I describe, even though I wrote "complete library", can actually operate with a rather tiny subset. What really would need to happen is implementation of the other crucial parts of the nuance engine... editable and customizable library, training, learning. Many people already achieve nearly perfect results with kinect voice commands, and I have my doubts that MS has suddenly become the technological leader in speech recognition software. But I have no doubt there is serious unlocked potential and room for improvement, and I believe MS has in making Kinect work well developed a nice hardware solution for a housewide speech recognition solution.

Look, I use speech recognition all day for my job, and it isn't perfect in ideal situations and it never will be, and won't be close for a long time still.

Didn't stop "good enough" from being an absolute game changer in my field, and it shouldn't stop MS from vastly expanding its scope of vision, product abilities, and market appeal.
 
I use Dragon on the Mac and don't use a mic, it's using the basic mics built into the iMac and MacBook. I can pace around the room and dictate and it's incredibly accurate; distance seems to have very little impact on its accuracy in my many years of experience.
the signal will be clearer at a greater distance, too close you have breaths etc. I think ~30cm-1 meter is prolly ideal perhaps 3 meters then sounds starts getting infected by outside noises
 
So youre implying kinect does worse than the above due to it being handicapped, OK thats well and good but the above are crap
As someone who uses "one of the above" (a nuance product/dragon variant) many hours each day, I beg to differ. I've stated here at B3D previously that even what I consider to be the most capable software running on a powerful workstation in a quiet room with good microphone isn't perfect (but then, neither are specially trained transcriptionists with the vocabulary I require)... it's still a long way from crap.

I don't see my suggestion facing any serious technological hurdle. Its more a matter of whether the bean counters running the numbers find what I assume to be true, and whether management at MS has a similar vision, or any vision, or balls for that matter. Lol.
 
the signal will be clearer at a greater distance, too close you have breaths etc. I think ~30cm-1 meter is prolly ideal perhaps 3 meters then sounds starts getting infected by outside noises
No, dragon is definitely more reliable with a mic. I used to use Dragon with my workhorse Plantronics Bluetooth headset but the dual mic setups in modern Macs are very good and I no longer use a headset. Obviously, depending on environmental factors, YMMV.
 
I think the best ecosystem provides 99% of what an individual user doesn't need.
How many people do you really think buy a product where a mere 1% of it's functionality meets their actual needs? I don't think there is such a huge market of people who can be so indiscriminate with their money. A diverse ecosystem is essential because needs will vary customer to customer but I think your balance is way off.

It shouldn't be MS or Sony job to provide apps but the third party developers. All the platform provider has to do is provide a marketplace and decent developer tools.
Agreed but we're not talking about apps. I like to think my friend, but particularly my technical colleagues, are very enthusiastic about technology but all that stuff about home automation? I really don't know a single person who wants to do that.

XB1 is hardly a powerhouse, and you want to go reducing performance?? How's that going to sell to the core, when the PS4 is going to more powerful at the same sort of price?
I think the idea is to lose all the functionality that isn't game related so you keep the exact hardware resources needed to run non-Kinect games but dispense with the cores for the background OS, Kinect, drop the TV input, apps, snapping etc. Easier said than done given a ton of OS stuff is still necessary even if you remove the non-game stuff.

This is an experiment that is not in a lot of people's interests, such as the key Xbox One project team who made the system what it is. How would it look such a 'core' system sold wildly, possibly better than the full system? I'm not sure it would but I've no doubt it would find a niche demand. I would certainly consider such a system if it came in at £150 cheaper.
 
i agree with bigus' post as i have been waiting and dreaming for MS to release Xbox voice recognition to Windows.

I though with windows 8, they would release it. But noooooo. its all still exclusive for xbox.

i just want to have simple voice command like "NEXT SONG" that works however the room noise and where ever my posisition in room. i even dreamed to have a Kinect-house where i can voice command from everywhere on my house.

in the end, i come in peace using wifi/bluetooth to my cellphone. I always bring my cell with me around house, so its fine.
 
How many people do you really think buy a product where a mere 1% of it's functionality meets their actual needs? I don't think there is such a huge market of people who can be so indiscriminate with their money. A diverse ecosystem is essential because needs will vary customer to customer but I think your balance is way off.


Agreed but we're not talking about apps. I like to think my friend, but particularly my technical colleagues, are very enthusiastic about technology but all that stuff about home automation? I really don't know a single person who wants to do that.


I think the idea is to lose all the functionality that isn't game related so you keep the exact hardware resources needed to run non-Kinect games but dispense with the cores for the background OS, Kinect, drop the TV input, apps, snapping etc. Easier said than done given a ton of OS stuff is still necessary even if you remove the non-game stuff.

This is an experiment that is not in a lot of people's interests, such as the key Xbox One project team who made the system what it is. How would it look such a 'core' system sold wildly, possibly better than the full system? I'm not sure it would but I've no doubt it would find a niche demand. I would certainly consider such a system if it came in at £150 cheaper.


It's a terrible idea! Dropping two CU's would probably save almost nothing from SOC area. Going to a 192 bit bus would require a entire engineering of the SOC with a lot of expense there. 2GB's of DDR probably saves like a buck wholesale (okay I may exaggerate, but probably not too wildly).

I dont necessarily disagree with the idea of a core Xone, but there's no valid reason to cut the already apparantely underpowered hardware. But the way to a core Xone is consolidation/remove Kinect/maybe remove Blu Ray drive.

All the voice stuff seems to be the main selling point of Xone. I dont know they'll cut Kinect, then. Nothing saying you cant cut the price with Kinect included anyway. If Xone is 399 by fall of 2014 with Kinect, doesn't this idea lose a lot of steam already?

It might be an interesting thing then to discuss if they can do the voice stuff without Kinect (some type of cheaper mic-only array? a mic on the controller?). Still havent opened my One but from what I gather, voice is so integral the system is (perhaps intentionally) clunky to use without it.
 
So youre implying kinect does worse than the above due to it being handicapped, OK thats well and good but the above are crap
So its voice recognition is worse than crap .. um OK ditto silent buddah. Anyways the average user dont give a monkeys about how amazing the tech hurdles are they just want something that works

yes the first 2-3 million till christmas (though they are available in shops its not like the wii (sans nz) or ps4 ) but I think like the wii U (and not the wii) we will soon bes seeing them aplenty on shops shelves, dropping the price under the main competition will maintain momentum. Im Sure Sony will not be able to follow the pricecut. PRICE ADVANTAGE = xbox

edit - just done a quick availablity check, so sold out for the xbone is not correct, Im still picking it to sell more in NPD dec due to having less places to sell to
sales.jpg

I'm not sure who you are talking to, but the Kinect voice recognition isn't crap. In fact once you learn the nuances of speaking to a machine, the hit rate increases over time. These days I simply say in a low tone, "Xbox on", "Xbox turn off", "Xbox, one guide", "Xbox, watch tv" and it just works. From the couch. Even people on this site (those faceless ppl you are talking about) have remarked about just how high quality the recognition is.

Also far from it being some gimmick you don't need like siri or google now... Kinect is not only a time and effort saver versus traversing on screen menus, it also provides an additional input in addition to your ten fingers. Its your eleventh finger/third hand so to speak. You are almost totally wrong in your assessment and post.
 
Last edited by a moderator:
It's a terrible idea! Dropping two CU's would probably save almost nothing from SOC area. Going to a 192 bit bus would require a entire engineering of the SOC with a lot of expense there. 2GB's of DDR probably saves like a buck wholesale (okay I may exaggerate, but probably not too wildly).
Read what I wrote, which isn't advocating the changes originally proposed, but cutting the resources unnecessary if you design an Xbox One only intended to deliver non-Kinect games.

I dont necessarily disagree with the idea of a core Xone, but there's no valid reason to cut the already apparantely underpowered hardware. But the way to a core Xone is consolidation/remove Kinect/maybe remove Blu Ray drive.
No, the proposal is to remove any and all hardware now redundant having removed the non-gaming functionality. The hardware accessible to Xbox One games would need to be exactly the same, otherwise you've lost compatibility.

All the voice stuff seems to be the main selling point of Xone.

The "voice stuff" is a means to an end not a reason to be. The ability to voice control Xbox One is there to do something with the Xbox One. If, like me, you're not interested in TV, plugging something else in, and would rather use a laptop, phone or tablet to augment your TV experience, rather than trying to squeeze stuff in a narrow pane down the right-hand side of the TV display, then that leaves games. What is what I'm interested in. It's all I'm interested in. I'm a gamer.
 
So short thread summary? "If I were MS I'd make the Xbox One be an exact duplicate of Sony PS4".
 
Erm, I think that's accurate to what some are saying, but let's not confuse anyone as it sounds like you're saying that is a summary of what the thread was created to say! Lol.

I don't think MS should make all this stuff I've described integral to the core console. They couldn't, there's too much variability in potential devices to interface with. In fact, if MS were to focus on speech recognition and allow peripherals and apps from third parties more broad access to system resources in a PC like way, that would mostly open up these possibilities.
 
So short thread summary? "If I were MS I'd make the Xbox One be an exact duplicate of Sony PS4".
Yup. I want nothing more than the ability to play all the exclusive games without having to buy into someone's grand product vision and technology I have no interest in and will never use.

Unless.. unless, somebody actually does something interesting with Kinect in games.
 
Yup. I want nothing more than the ability to play all the exclusive games without having to buy into someone's grand product vision and technology I have no interest in and will never use.

Unless.. unless, somebody actually does something interesting with Kinect in games.

So whats the time frame for your last sentence?
 
As someone who uses "one of the above" (a nuance product/dragon variant) many hours each day, I beg to differ. I've stated here at B3D previously that even what I consider to be the most capable software running on a powerful workstation in a quiet room with good microphone isn't perfect (but then, neither are specially trained transcriptionists with the vocabulary I require)... it's still a long way from crap.
So for you it works OK, but like you say youre in an ideal situation good mic/quiet room etc. Im talking about the usual siri/google voice which uses the inbuilt mic on the phone/pad
 
So whats the time frame for your last sentence?

Sometime during the life of the console. Something needs to happen to make the Xbox One appeal to me. A great Kinect game would be ideal because that would convince me on the premise that Microsoft are selling. Maybe that game is in the works but has yet to be announced.

Alternatively, I have no doubt the Xbox One will drop in price over the years and at some point the price will be low enough and the number of exclusive games will hit a point where the appeal is enough for me to buy one despite Kinect.
 
I use Dragon on the Mac and don't use a mic, it's using the basic mics built into the iMac and MacBook. I can pace around the room and dictate and it's incredibly accurate; distance seems to have very little impact on its accuracy in my many years of experience.

Does it work well in a noisy room from a distance? For example what if video game music, explosions or a movie soundtrack were blasting from your home theater speakers, can you still talk to dragon from 10 feet away and have it recognize your voice as accurately as kinect does? That's the only way to do an apples to apples comparison because that's the environment they are trying to get kinect to work in. If it's just for voice dictation in a quiet room well that's another matter entirely.


So youre implying kinect does worse than the above due to it being handicapped, OK thats well and good but the above are crap
So its voice recognition is worse than crap .. um OK ditto silent buddah. Anyways the average user dont give a monkeys about how amazing the tech hurdles are they just want something that works

See my above reply, we have to compare apples to apples otherwise posting here is a waste of time. For example leave your phone next to you kinect module on your tv stand between your speakers, play a movie, go sit on your couch and talk to them. Would dragon, siri, google voice or whatever work just as well as kinect in said environment? That's the real world environment they are trying to get it all to work in.


yes the first 2-3 million till christmas (though they are available in shops its not like the wii (sans nz) or ps4 ) but I think like the wii U (and not the wii) we will soon bes seeing them aplenty on shops shelves, dropping the price under the main competition will maintain momentum. Im Sure Sony will not be able to follow the pricecut. PRICE ADVANTAGE = xbox

You're suggesting they drop the price now which makes no sense when they are selling everything they can make at a profit. Why not wait and see if the bottom falls out of the xb1 market as so many here seem to predict before chopping the price. If anything what the 360 taught them is that dropping the price is not the answer but rather offering better software bundles.
 
Last edited by a moderator:
Does it work well in a noisy room from a distance? For example what if video game music, explosions or a movie soundtrack were blasting from your home theater speakers, can you still talk to dragon from 10 feet away and have it recognize your voice as accurately as kinect does?
I've never tried it, but I'm doubtful. It'll ignore the TV with a news channel on - as long as it's not too loud - or a ringing phone.

It's very good at ignoring the noise coming from the Mac though, you can even play music (again, not too loud as to drown out your voice) on the same device and it still works well.

I'm guessing Xbox One's advantages are a) Kinect is generally in the vicinity of the primary speakers and b) it knows what the noise is being output to assist in noise cancellation. Isn't that why you have to calibrate it?
 
If I was MS, (since they still have not shown a single must have app for kinect) I'ld release a kinectless version of the xbone for US$350.
Xbox One is Kinect. Kinect is not a peripheral anymore. The console is built around it.

So this is never going to happen. More reasons why it won't happen here:

http://www.oxm.co.uk/58866/blog/why-microsoft-wont-sell-xbox-one-without-kinect/

If this ever happened I would never play the console anymore and leave videogaming on consoles altogether.
 
Last edited by a moderator:
Does it work well in a noisy room from a distance? For example what if video game music, explosions or a movie soundtrack were blasting from your home theater speakers, can you still talk to dragon from 10 feet away and have it recognize your voice as accurately as kinect does? That's the only way to do an apples to apples comparison because that's the environment they are trying to get kinect to work in. If it's just for voice dictation in a quiet room well that's another matter entirely.
well its relative
heres a MS guy whos practised with it and even he has about ~60% success and hes speaking into a microphone
now in a typical household room its going to be worse than this

joker youre implying OK kinect is worse than voice/siri because its handicapped by being in a room noisy etc, so its no wonder that it performs worse than siri/voice OK fair enuf. What Im saying though is siri/voice are in themselves crap (everyone Ive met who've used them thinks so, great at first but a PITA to really use since you have to correct stuff all the time) so even if its the most technically thing ever if it performs worse than crap, its not going to find favor with the public
 
Back
Top