That's my fault. I didn't notice on my hands on and our specification sheets didn't specify. We didn't get the promo photos in advance. I've updated the text.
It's difficult to have a reliable IP67 or better rating with a headphone jack. Sony learned this with the Xperia Z3, and hasn't released anything claiming to be "waterproof" since.
Courage isn't the angle Apple should have taken. They should have simply said fluid dynamics. It has already been proven a headphone jack can be added to an iPhone 7, that there is physical space inside, contrary to what Apple claims.
And before someone trolls this. I'm aware there are phones that have reliable waterproofing with a 3.5mm jack (the Kyocera Hydro for instance) but it's important to note there are patents on those headphone jack designs and, like manual transmissions, there are only a few ways to encase the internals without being patent trolled, and it appears the licensing fees from companies that hold those patents are prohibitive, otherwise we'd see more headphone jacks in IP67 phones. This is why there are only a handful of companies in the world that make manual transmissions (ZF, Getrag, VAG, Suzuki, Ford, Holden, etc)
Sony does have an IP68 rating with a headphone jack and Samsung does too - and have had that for a while now. Sony didn't have an uncovered USB port until Z5.
What's changed is that they don't claim devices to be waterproof as there are no guarantees. Water resistant is a more accurate term anyway.
No, IP ratings are defined by the IEC and tested by independent labs. Like many certification sticker stuck to the box or the phone itself, it's a standard set by an outside agency. If it has an ip67 rating then it's not as good at water resistance as a device certified IP68. IPx7 equals 30min at up to 1m of water, IPx8 is to exceed IPx7 standards at a depth greater than 1m of water as agreed upon with the manufacturer, and not necessarily 1.5m although that is typical. So your statement "Testing has shown that the iPhone is at least as water resistant than the Samsung which is rated at 68. Apple is just more conservative." is straight up false. At the same time, if either fall into a tub, both should be ok. Neither should be taken swimming.
Actually, I believe you make a claim - I'm not sure any official agency checks up on that claim.
Nothing says that you have to make a claim for the harshest conditions device can stand up to, and I don't think any manufacturer will warrant you for water damage no matter what the conditions were claimed to have caused the damage.
I've seen dunk tests of Samsung and Apple devices under pretty harsh circumstances, but these devices are new and lord knows what their water resistance will be like nearer to the end of their service lives - and I believe neither do the manufacturers of these devices, especially after years of heat from CPUs, GPUs, and battery charging. Seals and adhesives dry out, shrink, and grow brittle and the only certainty is that while water resistant now, actual usage conditions will probably determine how wanter resistant these devices are in two or three years.
Given that, Apple's propensity is to be conservative and ignore the completion and specs, whereas Android manufacturer's - with the dire need to differentiate with wolves snapping at their heels - is to claim more.
LOL. Another fandroid losing mind about display resolution. Something like 5 years ago phone display resolutions have pushed the DPI into area that the human eye can't discern, yet today people still complain that a phone with QHD display is not good enough.
It isn't PenTile, cannot possibly be, as it's an IPS LCD display. Still has no idea what do they mean by "RGBW". The Amoled display on the Pro model is probably a PenTile, which is a shame as it would only have an effective resolution of around 720p.
Sorry I was wrong. RGBW is PenTile. Not diamond arrangement but different type of PenTile That works with LCD displays. So the Mate 10 has an effective pixel density of about 332ppi, and the Pro's effective pixel density is about 268ppi, which is a shame. On this website the OnePlus got criticized for not being sharp enough using 1080p PenTile on a 5.5 inch display. This one is 6 inches, so it's going to be much worse. The contrast ratio on the Pro is very weak compared to other OLED displays- 1:70000 vs 1:1000000 on the iPhone X. So display is a Major compromise with the Mate 10 "Pro". The iPhone X unfortunately also uses PenTile display with an effective pixel density of about 305ppi.
> On this website the OnePlus got criticized for not being sharp enough using 1080p PenTile on a 5.5 inch display. This one is 6 inches, so it's going to be much worse.
No, better, as it's resolution is 40% higher while the screen size barely differs (It isn't 16:9).
Moreover the contrast thing is nonsense. Whatever the claim, 1:7000 or 1;1000000 - they are oled screens so the numbers are made up: pixel off = 0 light, and wherever fhe brightness divided by 0 is infinite so contrast on both is infinite.
It's not better. Both phones have official pixel density of 401ppi, ehich translates into an effective 268ppi, which is not good. As for the contrast ratio, I know in theory you're right but I'm wondering what would make Huawei advertise such a relatively low number for their display. They haven't publicized the max brightness level of the Pro, it might be that the display doesn't get as bright to create strong contrast, or that the dynamic range is lacking like we've seen with the LG V30 in low brightness.
But if the number is contrast "ratio" then it could be possible that the display is relatively dim and reflects outside light. In that case, Lumia oled displays were outstanding.
I have googled RGBW pixel arrangement and found both, pentile and stripe matrices. So unless confirmed by some of the reviewers, I would not claim it. It seems however that pentile is more prevalent and thus more probable - unfortunately.
Unless there have been some changes in the layout of pentile displays, it is generally noted that pentile display at less than 1440 on ~5.5inch displays can lead to artefacts due to the arrangement of the green emitters. Therefore, if you're using a pentile display on a phone, you want a resolution of 1440 - higher is pointless. Conversely, IPS screens at the same display size see no benefit above 1080. I don't know about 6"+ devices, but the person above is correct about the oneplus device being criticised for choosing a 1080 pentile display and it did lead to artefacts in certain situations. Is it that important? Only for a flagship where the difference between manufacturers can be very small, prices are similar and the tiniest chink in your armour can result in a massive drop in market share.
It's important to note that pentile matrices only lower the chromatic (as opposed to luminatic) resolution of the display, which the human eye is highly insensitive to, to the point of making it irrelevant to anyone that doesn't make an obsessive effort squinting at the screen at unnaturally short distances to detect the pixels. It's a non issue at the ppi's we're dealing with these days. We should be criticizing OEMs for not exploiting the battery life free lunch that pentile martices offer.
The only thing interesting about the Kirin 970 is the AI processor, and they can't provide even a single interesting app that uses it. I'm sure there will eventually be cool uses for it, but probably not before Huawei's next phone comes out. That's not much of a selling feature.
I might sound like an old fart here (29 going on 70?), but I just don't get all this AI hype. If the neural network isn't trained through use on the phone, how is this different from any application doing the same job through conventional means? The bar for calling something AI seems to be precariously low. Is it AI if the app/algorithm in question was written by an AI, and it never changes or adapts? I would say no, not unless the output itself is reactive and flexible and adapts over time. After all, any algorithm adapts to new input - that's what they do. If it doesn't learn over time, there's nothing making it even remotely "intelligent". Now, I'm probably way over my head in terms of both programming and hardware understanding, but isn't this just a fancy new marketing term for highly specialized parallel workloads? I.e. something a GPU can do with minor (if any) modifications? If so: what's intelligent about it?
If all AI can do for consumers is a slightly better job at applying preset filters to your photos than the current algorithms do, I'm glad to pass on that. Thanks, but no thanks.
I wonder what the AI is for. My usual day isn't very usual other than the coffee, weather check and commute, and the phone-related changes are all taken care of by Tasker on Android. I don't need an AI to preempt my actions by doing the opposite of what I want.
A static non learning AI neural network can do things a normal core can do (e.g. face recognition and object following, translating, voice recognition or generation etc) far more efficient. Just like there is no need to include a GPU unless ypubcare about efficiency... the learning aspect is not core to AI as most networks get trained on super computers and then just used.
So in other words, the term AI is utterly meaningless. If the algorithm doing face recognition on my phone is generated by training a neural network rather than being written by a programmer, how does this make it "intelligent" in any way? An algorithm is an algorithm, no matter its origins. Sure, it might be a /better/ algorithm, but calling that "intelligence" is nothing more than marketing fluff. I suppose we could call it "AI-originated" or some such, but again: what value does that bring to the table?
I'm totally with you on this. The AI thing just seems to be a load of marketing bollocks. It may be useful for huge datasets where you want to determine patterns but most of the stuff we see as "AI" is just a load of IF, AND and OR statements. Alexa is a prime example - you have to get the phrasing just right for your question and even then it often struggles. It's the voice assistant equivalent of typing commands in DOS - get one word wrong and it won't work. To me this isn't intelligent, it's voice recognition attached to a list of "IFTTT" statements. Maybe I'm just old and stupid but I made stuff like this as a kid using BASIC which identified key words or phrases being typed and responded appropriately.... or inappropriately given I was a teenager and enjoyed giving out abuse using "AI".
I think the same. AI really sounds like a hype in mobile industry. Its possible apple or some other companies research on AI. But that does not mean they can make any notable changes to the apps or architecture that can actually "learn" unless predefined in the program itself, which makes it just another program with minor changes. Apple's AI claims are laughable btw
My guess is a little bird told Huawei that Apple was including a neural processor capable of 600 billion ops/second, and Huawei decided to put in one capable of a trillion ops/second to give them bragging rights (if anyone remembers the snark with which they announced the 970 as the iPhone 8 killer [the iPhone X name at the time in the rumor mills]).
They had to have compared the 970's CPU and GPU against the A10 and they certainly don't warrant that level of snark - or any snark at all really.
Unfortunately for them, they had no idea what was coming from the A11 as this is one aspect of Apple's development Apple could actually keep under wraps (since it's entirely in-house).
RGBW LCD? There goes any hope that Huawei should start caring about colour accuracy, then. Not that my Oneplus 3T needs replacement any time soon, but this is disappointing (and what kept me from getting a Huawei phone last time around - that, and the UX).
Oh, and that chip is gigantic for a phone SoC. Wow.
While it might just be a poor implementation, look up any review of the Lenovo Yoga 2 Pro - one of the most prominent RGBW LCD implementations. The white subpixel kills contrast, makes for poor grayscale accuracy, and in general hurts color saturation.
Any word on the camera sensors? Is it the same sensors they've used for a while now?
Also, interesting that they flipped the fingerprint sensors on the 10 series compared to the 9. The Pro had it on the front then and regular on the back. I wonder what the thought process is.
AI adjusting the camera settings is a big deal as users are terrible at it.The sensors are pretty great but the users can't use them at full potential and fixing that is a huge deal- not saying that Huawei is there yet as i got no clue how well it works right now. Other small things like battery life, keyboard, maybe accidental touch detecting can matter quite a bit.
You do take a weird position here on AI, you practically ask what is AI good for and than somewhat dismiss it. If you look at the GPU with the same mentality, it's even less of an asset. And it's not about only new experiences , it's also about better experiences and using less power for a given task.
At the end of the day, it is true that the functionality of a smartphone doesn't have much room to evolve, we'll still do a few core things and not much else.That's why it's all about design and display now, that's the low hanging fruit. First smaller bezels and then foldable.
I understand the article's approach to be more of "what's the point of an AI applying pre-defined, dumb filters" rather than a clear dismissal - a question I find key to this whole thing. Also: what's the difference between this "AI" implementation (that isn't really anything more than a fancy algorithm generated by training a neural network) and the algorithms that generally do this job? Okay, so they can adapt to the subject of your photos, rather than measurements like light and focus distance. But it won't adapt to the user's input over time - like, say, how I prefer noisier flash-free indoor photos to flash-lit ones - which makes this just as dumb and non-intelligent as any other algorithm.
If it deserves being called AI it needs to adapt to the user patterns and experience. To do this, systematic user feedback is necessary. Until that point, its a placebo. We don't know the difference. We can only assume there could be some difference. To successfully implement AI the language and logic should change. The binary logic of dumb, should be replaced with multi-valent logic like ternary logic and the semiconductor blocks should be replaced with appropriate methods. Companies are in advertisement-profit path. This should be changed.
No headphone jack, No expandable storage, smaller, lower resolution PenTile display. Why is that called a Pro? It's worse than the standard model! And the cat. 18 LTE is only on the Pro, so you've got it wrong.
You forgot that only the non-pro RGBW display goes to brught 730 nits. So , the 'pro' display is also dimmer. Huawei tried to handwave through this too, but it is clearly there in thr presentation snd the specs.
The only thing missing id LPDDR/emmc/UFS lottery, which I am sure Huawei can pull off.
When I think about AI on the phone, I imagine like an assistant who can remind me when something is due or tell me I got a sms from my wife.
The problem with AI right now is that it is at its infancy. Who really cares about my phone can find cats in my photos?
Until someone develop an AI where it can help you do things (instead of you checking on the phone), AI is still more of a marketing fad than anything.
I remembered ten years ago AR (augmented reality) is all the hype. People were predicting that one day we will have AR everywhere. Ten years later and the only thing AR made a big dent is the mobile Pokémon game.
Until AI can make our lives easier, I won't be getting hyped having it on my phone.
Think of AI as a pattern recognition engine. What does imply? Well for one thing, the engine is only going to see patterns in what it is FED! So what is it being fed? Obvious possibilities are images (and we know how that's working out) and audio (so speech recognition+translation, and again we know how that's working out). A similar obvious possibility could be stylus input and so writing recognition, but no-one seems to care much about that these days.
Now consider the "smart assistant" sort of idea. For that to work, there needs to be a way to stream all the "activities" of the phone, and their associated data, through the NPU in such a way that patterns can be detected. I trust that, at least the programmers reading this, start to see what sort of a challenge that is. What are the relevant data structures to represent this stream of activities? What does it mean to find some pattern/clustering in these activities --- how is that actionable?
Now Apple, for a few years now, has been pushing the idea that every time a program interacts with the user, it wraps up that interaction in a data structure that describes everything that's being done. The initial reason for this was, I think, for the on-phone search engine, but soon the most compelling reason for this (and easiest to understand the idea) was Continuity --- by wrapping up an "activity" in a self-describing data structure, Apple can transmit that data structure from, say, your phone to your mac or your watch, and so continue the activity between devices.
Reason I bring this up is that it obviously provides at least a starting point for Apple to go down this path. But only a starting point. Unlike images, each phone does not generate MILLIONS of these activities, so you have a very limited data set within which to find patterns. Can you get anything useful out of that? Who knows?
Android also has something called Activities, but as far as I can tell they are rather different from the Apple version, and not useful for the sort of issue I described. As far as I know Android has no such equivalent today. Presumably MS Will have to define such an equivalent as part of their copy of (as subset of) Continuity that's coming out with Fall Creators, and perhaps they have the same sort of AI ambitions that they hope to layer upon it?
The thing is, the implementation here is a "pattern recognition engine" /without any long-term memory/. In other words: it can't learn/adapt/improve over time. As such, it's as dumb as a bag of rocks. I wholeheartedly agree with not caring if the phone can recognize cats/faces/landscapes in my photos (which, besides, a regular non-AI-generated algorithm can do too, although probably not as well). How about learning the user's preferences in terms of aesthetics, subject matter, gallery culling? That would be useful, especially the last one: understanding what the fifteen photos you just took were focusing on, and then picking the best one in terms of focus, sharpness, background separation, colour, composition, and so on. Sure, also a task an algorithm could do (and do, in some apps), but sufficiently complex that it's likely that an AI that learns over time would do a far better job. Not to mention that an adaptive AI in that situation could regularly present the user with prompts like "I selected this as the best shots. These were the runner-ups. Which do you prefer?" which would give valuable feedback to adjust the process.
I do plenty of manual editing and cataloging of photos shot on the phone, in a mirror of the processes I use for a DSLR and a laptop. I don't think an AI will know if I want to do a black and white series in Snapseed from color photos, it won't know which ones to keep or delete, and it won't know about the folder organization I use.
It COULD improve over time if they pushed out upgraded NN. Off device training with occasional updates to supported devices is going to be the best option for awhile.
Disappointed they aren't using the AI accelerator for more advanced computational photography, like say better identifying a person's face and body and then doing the boken around that, or improving dynamic range by knowing exactly which parts to expose more, and so on.
Auto-switching to a "mode" is really something that other phone makers have had for a few years now in their "Auto" mode.
This was one of my comments to Huawei. I was told that the auto modes do sub-frame enhancements for clarity, though I'm under the assumption those are the same as previous tools and algorithms and not AI driven. Part of the issue here is AI can be a hammer, but there needs to be nails.
With the supposed high performance of this neural chip when compared to Apple’s right now, I’m a bit confused.
Since we don’t know exactly how this works, and we know even less about Apple’s, how can we even begin to compare performance between them?
Hopefully, you will soon have a real review of the 8/8+. As well as the deep dive of then SoC, something which was promised for last year’s model, but never materialized.
A comparison between these two new SoCs will be interesting.
Apple gave an "ops per sec" number, Huawei gave a FLOPS number. One was bigger than the other. That's all we have.
There are a million issues with this. Are both talking about 32-bit flops? Or 16-bit? Maybe Apple meant 32-bit FLOPs and Huawei 16-bit? And is FLOP actually a useful metric? Maybe in real situations these devices are really limited by their cache or memory subsystems?
To be fair, no-one is (yet) making abig deal about this precisely because anyone who know anything understands just how meaningless both the numbers are. It will take a year or three before we have enough experience with what the units do that we CARE about, and so know what to bother benchmarking or how to compare them.
Baidu have, for example, a supposed NPU benchmark suite, but it kinda sucks. All it tests is the speed of some convolutions at a range of different sizes. More problematically, at least as it exists today, it's basically C code. So you can look up the performance number for an iPhone, but it's meaningless because it doesn't even give you the GPU performance, let alone the NPU performance. We need to learn what sort of low-level performance primitives we care about testing, then we need to write up comparable cross-device code that uses the optimal per-device APIs on each device. This will take time.
I assumed Apple's numbers were something in the inferences/second ballpark as this was a neural processor (and MLKit seems to be processing data produced by standard machine learning models). We know the Apple neural processor is used by Face ID, first as a gatekeeper (detect attempts to fake it out, and then attempt to see if the face in view is the same one store stored in the secure enclave.
Flops seems to imply floating point operations/second.
I hate to ask this but have been wondering, will there be a review on the iPhone 8 or 8s at all or are you waiting for the X to come out? I don't trust a lot of the other review sites so was looking forward to an Anandtech review on the phone.
Heh, both are now “old” SoCs, even though Google is using the 835 in their new phones. I guess we’ll have to wait for the u45, in several months, as well as Samsung’s equivelant. The only real competitor to Apple’s new A11 is the 970.
It's pretty cool that the phone tosses out a kitty icon on a photo where a kitty is present. Cats are really awesome and having the photos on a device auto-sorted based on their contents (cats or no cats) would be useful. With that said, I don't see very much value in that since I can look at a photo with a cat and identify it myself. It isn't strictly necessary for a phone to do that for me.
Think of all the time you would save from manually sorting cat from non-Catholic photos. This phone will pay for itself in just a few short days of use for most consumers!
I think its way too late to escape religious profiling performed via AI. I'm pretty sure Facebook, Alphabet, Twitter, and everyone else that offers "free" services has captured and profiled relgion as an sellable attribute against which to target adverts.
I'm not sure if you meant that seriously or in jest. AI could be useful for identifying subjects in photos and keeping it in an offline database (for privacy reasons), so you can search for recent photos of baby Peter or Auntie May.
The article title states it's a Hands-On review, yet there is almost no hands on content. What about the audio and audio apps? How does the over-the-top 384kHz / 32-bit audio sound and what is the look/feel/sound of the special ear phones? What audio player is included and how did it work?
How about the AI development environment? What is it like?
Was under the impression we were correct based on the information we had. If we second-guessed everything, we'd post next week. The article was updated very soon after posting.
> The details were (annoyingly) light but it related to memory fragmentation (shouldn’t be an issue with DRAM),
Memory fragmentation is a real performance issue. It's not that the DRAM itself cares about being fragmented (it doesn't). However, the memory allocation routines in the kernel and C libraries do care because they have to work harder to satisfy an allocation request. The more CPU cycles allocation consumes, the more power the allocation takes (reduced battery life) and the longer it takes (increased application latency).
Besides real-time translation, what other ready for market development in AI is NOT about transferring and converting more personal identity into properties owned by for-profit corporations?
I thank you for this post and for mentioning automatic audio transcription. As far as I know there is no app that can do that even making use of Google Cloud. What I'd like is the possibility to search through audio/video recordings the way we do now with text in files and obviously being able to do that offline and on the fly.
There is an iPhone app that does this called Just Press Record. It will record audio when you press the button (as per the name) and transcribe it LOCALLY (no cloud). It runs on pre-iPhone 8 so doesn't HAVE to use the NPU, but it may be using Apple's generic neural APIs, which use whatever hardware is available and optimal for a particular task. The accuracy seems pretty good in my experience.
I didn't know about this app, and being confined to Apple is out of reach for me to try it out but from what I've read it's more about dictation while I was more about being able to search for words from a general audio/video recording that implies in my opinion an higher grade of difficulties as there could be different voices mixed with other sounds.
The article mentions that the Pro version has a larger screen than the regular version. This is sloppy thinking that falls short of the standards of this site. A 6" 18:9 has a lower screen area than a 5.9" 16:9 screen. The regular mate 10 is the larger screened of the two. In combination with the fact that non-cropped 16:9 media are displayed at about 5.5" equivalent on the pro (another consequence of aspect ratio voodoo), I consider the non-pro version a much, much better buy. I'm avoiding all >16:9 ratio phones for this reason and am very glad Huawei finally produced a small bezeled 16:9 one.
I am not sure what is great here! This was done and dusted by Microsoft long time back with 900 series phones! It was a better implementation with Windows as OS for desktop as well which is proper Desktop OS. Nothing interesting! Huawei should stick to its original *cheap*affordable* tags rather than AI and what not!
I have used P9, the P 10 Plus and the Mate 9 Pro and I have always been underwhelmed by Huawei's offerings. The only phone that I did like was the 6P but that wasn't truly theirs. Oh well. I don't think I'll dip my hands in Huawei's phone well in the near future.
We’ve updated our terms. By continuing to use the site and/or by logging into your account, you agree to the Site’s updated Terms of Use and Privacy Policy.
Facebook
Twitter
RSS
103 Comments
Back to Article
polygon_21 - Monday, October 16, 2017 - link
At least the 3.5mm socket is still therePotato Cat - Monday, October 16, 2017 - link
true (poking at apple again)Dr. Swag - Monday, October 16, 2017 - link
And GoogleSetiroN - Monday, October 16, 2017 - link
Actually, not on the pro, unfortunately.Aberamati - Monday, October 16, 2017 - link
The "Pro" model doesn't have a headphone jack.protomech - Monday, October 16, 2017 - link
From the article: "All three devices have a headphone jack"Maybe the article is wrong.
Aberamati - Monday, October 16, 2017 - link
Yes the article is wrong. Look at the pictures. All other websites noted that the Pro doesn't have headphone jack.Ian Cutress - Monday, October 16, 2017 - link
That's my fault. I didn't notice on my hands on and our specification sheets didn't specify. We didn't get the promo photos in advance. I've updated the text.Samus - Monday, October 16, 2017 - link
It's difficult to have a reliable IP67 or better rating with a headphone jack. Sony learned this with the Xperia Z3, and hasn't released anything claiming to be "waterproof" since.Courage isn't the angle Apple should have taken. They should have simply said fluid dynamics. It has already been proven a headphone jack can be added to an iPhone 7, that there is physical space inside, contrary to what Apple claims.
And before someone trolls this. I'm aware there are phones that have reliable waterproofing with a 3.5mm jack (the Kyocera Hydro for instance) but it's important to note there are patents on those headphone jack designs and, like manual transmissions, there are only a few ways to encase the internals without being patent trolled, and it appears the licensing fees from companies that hold those patents are prohibitive, otherwise we'd see more headphone jacks in IP67 phones. This is why there are only a handful of companies in the world that make manual transmissions (ZF, Getrag, VAG, Suzuki, Ford, Holden, etc)
Trixanity - Monday, October 16, 2017 - link
Sony does have an IP68 rating with a headphone jack and Samsung does too - and have had that for a while now. Sony didn't have an uncovered USB port until Z5.What's changed is that they don't claim devices to be waterproof as there are no guarantees. Water resistant is a more accurate term anyway.
melgross - Monday, October 16, 2017 - link
No such thing as a waterproof device. The us government doesn’t even allow it to be used unless proof of something being waterproof is provided.lilmoe - Monday, October 16, 2017 - link
Uhhhhh. Apple isn't the trend setter when it comes to hardware. The trend setter has ip68, not 67, AND a headphone jack.But sure, keep it up bruh.
melgross - Monday, October 16, 2017 - link
Good luck. Testing has shown that the iPhone is at least as water resistant than the Samsung which is rated at 68. Apple is just more conservative.goatfajitas - Tuesday, October 17, 2017 - link
Melgross, you are batshizcrazy man. You cant even let the minor diff from an ip67 to an ip68 go without defending Apple. Get a grip.melgross - Wednesday, October 18, 2017 - link
Get a grip yourself child. Every test that’s been done in these phones shows that their water resistance is about the same.Manch - Thursday, October 19, 2017 - link
No, IP ratings are defined by the IEC and tested by independent labs. Like many certification sticker stuck to the box or the phone itself, it's a standard set by an outside agency. If it has an ip67 rating then it's not as good at water resistance as a device certified IP68. IPx7 equals 30min at up to 1m of water, IPx8 is to exceed IPx7 standards at a depth greater than 1m of water as agreed upon with the manufacturer, and not necessarily 1.5m although that is typical. So your statement "Testing has shown that the iPhone is at least as water resistant than the Samsung which is rated at 68. Apple is just more conservative." is straight up false. At the same time, if either fall into a tub, both should be ok. Neither should be taken swimming.lilmoe - Friday, October 20, 2017 - link
@ Manch,NO, testing and certification is done by youtubers and drop testers!
varase - Sunday, November 19, 2017 - link
Actually, I believe you make a claim - I'm not sure any official agency checks up on that claim.Nothing says that you have to make a claim for the harshest conditions device can stand up to, and I don't think any manufacturer will warrant you for water damage no matter what the conditions were claimed to have caused the damage.
I've seen dunk tests of Samsung and Apple devices under pretty harsh circumstances, but these devices are new and lord knows what their water resistance will be like nearer to the end of their service lives - and I believe neither do the manufacturers of these devices, especially after years of heat from CPUs, GPUs, and battery charging. Seals and adhesives dry out, shrink, and grow brittle and the only certainty is that while water resistant now, actual usage conditions will probably determine how wanter resistant these devices are in two or three years.
Given that, Apple's propensity is to be conservative and ignore the completion and specs, whereas Android manufacturer's - with the dire need to differentiate with wolves snapping at their heels - is to claim more.
Take that for what it's worth.
peevee - Thursday, October 19, 2017 - link
Patents for manual transmissions have expired 100 years ago.Allan_Hundeboll - Friday, October 20, 2017 - link
:SamusI dont really believe that only a handfull of campanies make manual transmissions?
Do you have any sources to backup your claim?
lmcd - Tuesday, October 24, 2017 - link
The table is still wrong -- lists USB-C and 3.5mm for all 3.jjj - Monday, October 16, 2017 - link
699 euros for the Mate 10 is a bit much given the budgety pseudo 1440p PenTile display.UtilityMax - Monday, October 16, 2017 - link
LOL. Another fandroid losing mind about display resolution. Something like 5 years ago phone display resolutions have pushed the DPI into area that the human eye can't discern, yet today people still complain that a phone with QHD display is not good enough.zoxo - Monday, October 16, 2017 - link
Actually, once you use the phone as a HMD, the resolution is suddenly lacking.K_Space - Monday, October 16, 2017 - link
+1risa2000 - Monday, October 16, 2017 - link
Where did you get the information that the LCD in Mate 10 is pentile?Aberamati - Monday, October 16, 2017 - link
It isn't PenTile, cannot possibly be, as it's an IPS LCD display. Still has no idea what do they mean by "RGBW".The Amoled display on the Pro model is probably a PenTile, which is a shame as it would only have an effective resolution of around 720p.
melgross - Monday, October 16, 2017 - link
Pen tile is RGGB.Aberamati - Monday, October 16, 2017 - link
Or RGBW. Look it up.melgross - Wednesday, October 18, 2017 - link
No, RGGB. They may be making other formulas for other phone makers, but that’s thecPentile configuration.melgross - Wednesday, October 18, 2017 - link
Ok, I checked. It’s RGGB for OLED, and RGBW for LCD.Aberamati - Monday, October 16, 2017 - link
Sorry I was wrong. RGBW is PenTile. Not diamond arrangement but different type of PenTile That works with LCD displays. So the Mate 10 has an effective pixel density of about 332ppi, and the Pro's effective pixel density is about 268ppi, which is a shame. On this website the OnePlus got criticized for not being sharp enough using 1080p PenTile on a 5.5 inch display. This one is 6 inches, so it's going to be much worse. The contrast ratio on the Pro is very weak compared to other OLED displays- 1:70000 vs 1:1000000 on the iPhone X. So display is a Major compromise with the Mate 10 "Pro". The iPhone X unfortunately also uses PenTile display with an effective pixel density of about 305ppi.jospoortvliet - Tuesday, October 17, 2017 - link
> On this website the OnePlus got criticized for not being sharp enough using 1080p PenTile on a 5.5 inch display. This one is 6 inches, so it's going to be much worse.No, better, as it's resolution is 40% higher while the screen size barely differs (It isn't 16:9).
Moreover the contrast thing is nonsense. Whatever the claim, 1:7000 or 1;1000000 - they are oled screens so the numbers are made up: pixel off = 0 light, and wherever fhe brightness divided by 0 is infinite so contrast on both is infinite.
Aberamati - Tuesday, October 17, 2017 - link
It's not better. Both phones have official pixel density of 401ppi, ehich translates into an effective 268ppi, which is not good. As for the contrast ratio, I know in theory you're right but I'm wondering what would make Huawei advertise such a relatively low number for their display. They haven't publicized the max brightness level of the Pro, it might be that the display doesn't get as bright to create strong contrast, or that the dynamic range is lacking like we've seen with the LG V30 in low brightness.FreidoNumeroUno - Saturday, October 21, 2017 - link
But if the number is contrast "ratio" then it could be possible that the display is relatively dim and reflects outside light. In that case, Lumia oled displays were outstanding.peevee - Thursday, October 19, 2017 - link
Except there are such things as reflectance and light spillage from adjacent pixels.risa2000 - Tuesday, October 17, 2017 - link
I have googled RGBW pixel arrangement and found both, pentile and stripe matrices. So unless confirmed by some of the reviewers, I would not claim it. It seems however that pentile is more prevalent and thus more probable - unfortunately.philehidiot - Thursday, October 19, 2017 - link
Unless there have been some changes in the layout of pentile displays, it is generally noted that pentile display at less than 1440 on ~5.5inch displays can lead to artefacts due to the arrangement of the green emitters. Therefore, if you're using a pentile display on a phone, you want a resolution of 1440 - higher is pointless. Conversely, IPS screens at the same display size see no benefit above 1080. I don't know about 6"+ devices, but the person above is correct about the oneplus device being criticised for choosing a 1080 pentile display and it did lead to artefacts in certain situations. Is it that important? Only for a flagship where the difference between manufacturers can be very small, prices are similar and the tiniest chink in your armour can result in a massive drop in market share.Rdmkr - Saturday, October 21, 2017 - link
It's important to note that pentile matrices only lower the chromatic (as opposed to luminatic) resolution of the display, which the human eye is highly insensitive to, to the point of making it irrelevant to anyone that doesn't make an obsessive effort squinting at the screen at unnaturally short distances to detect the pixels. It's a non issue at the ppi's we're dealing with these days. We should be criticizing OEMs for not exploiting the battery life free lunch that pentile martices offer.Look here if you need convincing that the difference between chrominance and luminance matters: https://en.m.wikipedia.org/wiki/Chrominance
peevee - Thursday, October 19, 2017 - link
1440 is in 10 (not Pro), and it is LCD.cfenton - Monday, October 16, 2017 - link
The only thing interesting about the Kirin 970 is the AI processor, and they can't provide even a single interesting app that uses it. I'm sure there will eventually be cool uses for it, but probably not before Huawei's next phone comes out. That's not much of a selling feature.Valantar - Monday, October 16, 2017 - link
I might sound like an old fart here (29 going on 70?), but I just don't get all this AI hype. If the neural network isn't trained through use on the phone, how is this different from any application doing the same job through conventional means? The bar for calling something AI seems to be precariously low. Is it AI if the app/algorithm in question was written by an AI, and it never changes or adapts? I would say no, not unless the output itself is reactive and flexible and adapts over time. After all, any algorithm adapts to new input - that's what they do. If it doesn't learn over time, there's nothing making it even remotely "intelligent". Now, I'm probably way over my head in terms of both programming and hardware understanding, but isn't this just a fancy new marketing term for highly specialized parallel workloads? I.e. something a GPU can do with minor (if any) modifications? If so: what's intelligent about it?If all AI can do for consumers is a slightly better job at applying preset filters to your photos than the current algorithms do, I'm glad to pass on that. Thanks, but no thanks.
serendip - Monday, October 16, 2017 - link
I wonder what the AI is for. My usual day isn't very usual other than the coffee, weather check and commute, and the phone-related changes are all taken care of by Tasker on Android. I don't need an AI to preempt my actions by doing the opposite of what I want.jospoortvliet - Tuesday, October 17, 2017 - link
A static non learning AI neural network can do things a normal core can do (e.g. face recognition and object following, translating, voice recognition or generation etc) far more efficient. Just like there is no need to include a GPU unless ypubcare about efficiency... the learning aspect is not core to AI as most networks get trained on super computers and then just used.Valantar - Tuesday, October 17, 2017 - link
So in other words, the term AI is utterly meaningless. If the algorithm doing face recognition on my phone is generated by training a neural network rather than being written by a programmer, how does this make it "intelligent" in any way? An algorithm is an algorithm, no matter its origins. Sure, it might be a /better/ algorithm, but calling that "intelligence" is nothing more than marketing fluff. I suppose we could call it "AI-originated" or some such, but again: what value does that bring to the table?philehidiot - Thursday, October 19, 2017 - link
I'm totally with you on this. The AI thing just seems to be a load of marketing bollocks. It may be useful for huge datasets where you want to determine patterns but most of the stuff we see as "AI" is just a load of IF, AND and OR statements. Alexa is a prime example - you have to get the phrasing just right for your question and even then it often struggles. It's the voice assistant equivalent of typing commands in DOS - get one word wrong and it won't work. To me this isn't intelligent, it's voice recognition attached to a list of "IFTTT" statements. Maybe I'm just old and stupid but I made stuff like this as a kid using BASIC which identified key words or phrases being typed and responded appropriately.... or inappropriately given I was a teenager and enjoyed giving out abuse using "AI".FreidoNumeroUno - Saturday, October 21, 2017 - link
I think the same. AI really sounds like a hype in mobile industry. Its possible apple or some other companies research on AI. But that does not mean they can make any notable changes to the apps or architecture that can actually "learn" unless predefined in the program itself, which makes it just another program with minor changes.Apple's AI claims are laughable btw
varase - Sunday, November 19, 2017 - link
My guess is a little bird told Huawei that Apple was including a neural processor capable of 600 billion ops/second, and Huawei decided to put in one capable of a trillion ops/second to give them bragging rights (if anyone remembers the snark with which they announced the 970 as the iPhone 8 killer [the iPhone X name at the time in the rumor mills]).They had to have compared the 970's CPU and GPU against the A10 and they certainly don't warrant that level of snark - or any snark at all really.
Unfortunately for them, they had no idea what was coming from the A11 as this is one aspect of Apple's development Apple could actually keep under wraps (since it's entirely in-house).
Valantar - Monday, October 16, 2017 - link
RGBW LCD? There goes any hope that Huawei should start caring about colour accuracy, then. Not that my Oneplus 3T needs replacement any time soon, but this is disappointing (and what kept me from getting a Huawei phone last time around - that, and the UX).Oh, and that chip is gigantic for a phone SoC. Wow.
Andrei Frumusanu - Monday, October 16, 2017 - link
RGBW has absolutely no effect on colour accuracy.melgross - Monday, October 16, 2017 - link
Saturation is affected.Aberamati - Monday, October 16, 2017 - link
RGBW does affect the grayscale accuracy of the display, since one white pixel is responsible for 4 pixelsValantar - Tuesday, October 17, 2017 - link
While it might just be a poor implementation, look up any review of the Lenovo Yoga 2 Pro - one of the most prominent RGBW LCD implementations. The white subpixel kills contrast, makes for poor grayscale accuracy, and in general hurts color saturation.Trixanity - Monday, October 16, 2017 - link
Any word on the camera sensors? Is it the same sensors they've used for a while now?Also, interesting that they flipped the fingerprint sensors on the 10 series compared to the 9. The Pro had it on the front then and regular on the back. I wonder what the thought process is.
jjj - Monday, October 16, 2017 - link
AI adjusting the camera settings is a big deal as users are terrible at it.The sensors are pretty great but the users can't use them at full potential and fixing that is a huge deal- not saying that Huawei is there yet as i got no clue how well it works right now.Other small things like battery life, keyboard, maybe accidental touch detecting can matter quite a bit.
You do take a weird position here on AI, you practically ask what is AI good for and than somewhat dismiss it. If you look at the GPU with the same mentality, it's even less of an asset.
And it's not about only new experiences , it's also about better experiences and using less power for a given task.
At the end of the day, it is true that the functionality of a smartphone doesn't have much room to evolve, we'll still do a few core things and not much else.That's why it's all about design and display now, that's the low hanging fruit. First smaller bezels and then foldable.
Valantar - Tuesday, October 17, 2017 - link
I understand the article's approach to be more of "what's the point of an AI applying pre-defined, dumb filters" rather than a clear dismissal - a question I find key to this whole thing. Also: what's the difference between this "AI" implementation (that isn't really anything more than a fancy algorithm generated by training a neural network) and the algorithms that generally do this job? Okay, so they can adapt to the subject of your photos, rather than measurements like light and focus distance. But it won't adapt to the user's input over time - like, say, how I prefer noisier flash-free indoor photos to flash-lit ones - which makes this just as dumb and non-intelligent as any other algorithm.FreidoNumeroUno - Saturday, October 21, 2017 - link
If it deserves being called AI it needs to adapt to the user patterns and experience. To do this, systematic user feedback is necessary.Until that point, its a placebo. We don't know the difference. We can only assume there could be some difference.
To successfully implement AI the language and logic should change. The binary logic of dumb, should be replaced with multi-valent logic like ternary logic and the semiconductor blocks should be replaced with appropriate methods.
Companies are in advertisement-profit path. This should be changed.
Aberamati - Monday, October 16, 2017 - link
No headphone jack, No expandable storage, smaller, lower resolution PenTile display. Why is that called a Pro? It's worse than the standard model!And the cat. 18 LTE is only on the Pro, so you've got it wrong.
halcyon - Tuesday, October 17, 2017 - link
You forgot that only the non-pro RGBW display goes to brught 730 nits. So , the 'pro' display is also dimmer. Huawei tried to handwave through this too, but it is clearly there in thr presentation snd the specs.The only thing missing id LPDDR/emmc/UFS lottery, which I am sure Huawei can pull off.
Cliff34 - Monday, October 16, 2017 - link
When I think about AI on the phone, I imagine like an assistant who can remind me when something is due or tell me I got a sms from my wife.The problem with AI right now is that it is at its infancy. Who really cares about my phone can find cats in my photos?
Until someone develop an AI where it can help you do things (instead of you checking on the phone), AI is still more of a marketing fad than anything.
I remembered ten years ago AR (augmented reality) is all the hype. People were predicting that one day we will have AR everywhere. Ten years later and the only thing AR made a big dent is the mobile Pokémon game.
Until AI can make our lives easier, I won't be getting hyped having it on my phone.
name99 - Monday, October 16, 2017 - link
Think of AI as a pattern recognition engine. What does imply?Well for one thing, the engine is only going to see patterns in what it is FED! So what is it being fed?
Obvious possibilities are images (and we know how that's working out) and audio (so speech recognition+translation, and again we know how that's working out). A similar obvious possibility could be stylus input and so writing recognition, but no-one seems to care much about that these days.
Now consider the "smart assistant" sort of idea. For that to work, there needs to be a way to stream all the "activities" of the phone, and their associated data, through the NPU in such a way that patterns can be detected. I trust that, at least the programmers reading this, start to see what sort of a challenge that is. What are the relevant data structures to represent this stream of activities? What does it mean to find some pattern/clustering in these activities --- how is that actionable?
Now Apple, for a few years now, has been pushing the idea that every time a program interacts with the user, it wraps up that interaction in a data structure that describes everything that's being done. The initial reason for this was, I think, for the on-phone search engine, but soon the most compelling reason for this (and easiest to understand the idea) was Continuity --- by wrapping up an "activity" in a self-describing data structure, Apple can transmit that data structure from, say, your phone to your mac or your watch, and so continue the activity between devices.
Reason I bring this up is that it obviously provides at least a starting point for Apple to go down this path. But only a starting point. Unlike images, each phone does not generate MILLIONS of these activities, so you have a very limited data set within which to find patterns. Can you get anything useful out of that? Who knows?
Android also has something called Activities, but as far as I can tell they are rather different from the Apple version, and not useful for the sort of issue I described. As far as I know Android has no such equivalent today. Presumably MS Will have to define such an equivalent as part of their copy of (as subset of) Continuity that's coming out with Fall Creators, and perhaps they have the same sort of AI ambitions that they hope to layer upon it?
Valantar - Tuesday, October 17, 2017 - link
The thing is, the implementation here is a "pattern recognition engine" /without any long-term memory/. In other words: it can't learn/adapt/improve over time. As such, it's as dumb as a bag of rocks. I wholeheartedly agree with not caring if the phone can recognize cats/faces/landscapes in my photos (which, besides, a regular non-AI-generated algorithm can do too, although probably not as well). How about learning the user's preferences in terms of aesthetics, subject matter, gallery culling? That would be useful, especially the last one: understanding what the fifteen photos you just took were focusing on, and then picking the best one in terms of focus, sharpness, background separation, colour, composition, and so on. Sure, also a task an algorithm could do (and do, in some apps), but sufficiently complex that it's likely that an AI that learns over time would do a far better job. Not to mention that an adaptive AI in that situation could regularly present the user with prompts like "I selected this as the best shots. These were the runner-ups. Which do you prefer?" which would give valuable feedback to adjust the process.serendip - Tuesday, October 17, 2017 - link
I do plenty of manual editing and cataloging of photos shot on the phone, in a mirror of the processes I use for a DSLR and a laptop. I don't think an AI will know if I want to do a black and white series in Snapseed from color photos, it won't know which ones to keep or delete, and it won't know about the folder organization I use.So what exactly is the AI for?
tuxRoller - Tuesday, October 17, 2017 - link
It COULD improve over time if they pushed out upgraded NN. Off device training with occasional updates to supported devices is going to be the best option for awhile.Krysto - Monday, October 16, 2017 - link
Disappointed they aren't using the AI accelerator for more advanced computational photography, like say better identifying a person's face and body and then doing the boken around that, or improving dynamic range by knowing exactly which parts to expose more, and so on.Auto-switching to a "mode" is really something that other phone makers have had for a few years now in their "Auto" mode.
Ian Cutress - Monday, October 16, 2017 - link
This was one of my comments to Huawei. I was told that the auto modes do sub-frame enhancements for clarity, though I'm under the assumption those are the same as previous tools and algorithms and not AI driven. Part of the issue here is AI can be a hammer, but there needs to be nails.melgross - Monday, October 16, 2017 - link
With the supposed high performance of this neural chip when compared to Apple’s right now, I’m a bit confused.Since we don’t know exactly how this works, and we know even less about Apple’s, how can we even begin to compare performance between them?
Hopefully, you will soon have a real review of the 8/8+. As well as the deep dive of then SoC, something which was promised for last year’s model, but never materialized.
A comparison between these two new SoCs will be interesting.
name99 - Monday, October 16, 2017 - link
Apple gave an "ops per sec" number, Huawei gave a FLOPS number. One was bigger than the other.That's all we have.
There are a million issues with this. Are both talking about 32-bit flops? Or 16-bit? Maybe Apple meant 32-bit FLOPs and Huawei 16-bit?
And is FLOP actually a useful metric? Maybe in real situations these devices are really limited by their cache or memory subsystems?
To be fair, no-one is (yet) making abig deal about this precisely because anyone who know anything understands just how meaningless both the numbers are. It will take a year or three before we have enough experience with what the units do that we CARE about, and so know what to bother benchmarking or how to compare them.
Baidu have, for example, a supposed NPU benchmark suite, but it kinda sucks. All it tests is the speed of some convolutions at a range of different sizes. More problematically, at least as it exists today, it's basically C code. So you can look up the performance number for an iPhone, but it's meaningless because it doesn't even give you the GPU performance, let alone the NPU performance.
We need to learn what sort of low-level performance primitives we care about testing, then we need to write up comparable cross-device code that uses the optimal per-device APIs on each device. This will take time.
melgross - Wednesday, October 18, 2017 - link
This is the problem I’m thinking about. We don’t have enough info to go by.varase - Tuesday, November 21, 2017 - link
I assumed Apple's numbers were something in the inferences/second ballpark as this was a neural processor (and MLKit seems to be processing data produced by standard machine learning models). We know the Apple neural processor is used by Face ID, first as a gatekeeper (detect attempts to fake it out, and then attempt to see if the face in view is the same one store stored in the secure enclave.Flops seems to imply floating point operations/second.
Color me confused.
Dennis Travis - Monday, October 16, 2017 - link
I hate to ask this but have been wondering, will there be a review on the iPhone 8 or 8s at all or are you waiting for the X to come out?I don't trust a lot of the other review sites so was looking forward to an Anandtech review on the phone.
name99 - Monday, October 16, 2017 - link
The previous iPhone deep dives came out in November!Be patient!!
zeeBomb - Monday, October 16, 2017 - link
how fast is the 970 in comparison to the 835 and the A10?Aberamati - Monday, October 16, 2017 - link
No benchmarks or testing exist currently, but it should be on par with the 835 or slightly faster.melgross - Monday, October 16, 2017 - link
Heh, both are now “old” SoCs, even though Google is using the 835 in their new phones. I guess we’ll have to wait for the u45, in several months, as well as Samsung’s equivelant. The only real competitor to Apple’s new A11 is the 970.lowphas - Monday, October 16, 2017 - link
Maybe the same as the kirin 960 for the cpu, and couple +% on the gpu. Seems to be essentially the same.Bondurant - Monday, October 16, 2017 - link
Stereo speakers ?Bondurant - Monday, October 16, 2017 - link
And why in the world did they skip the Bluetooth 5.0. ?BrokenCrayons - Monday, October 16, 2017 - link
It's pretty cool that the phone tosses out a kitty icon on a photo where a kitty is present. Cats are really awesome and having the photos on a device auto-sorted based on their contents (cats or no cats) would be useful. With that said, I don't see very much value in that since I can look at a photo with a cat and identify it myself. It isn't strictly necessary for a phone to do that for me.Rocket321 - Monday, October 16, 2017 - link
Think of all the time you would save from manually sorting cat from non-Catholic photos. This phone will pay for itself in just a few short days of use for most consumers!Rocket321 - Monday, October 16, 2017 - link
Hmm, auto correct took an interesting turn with that reply. In retrospect I don't want Ai to do any sort of religious profiling.BrokenCrayons - Tuesday, October 17, 2017 - link
I think its way too late to escape religious profiling performed via AI. I'm pretty sure Facebook, Alphabet, Twitter, and everyone else that offers "free" services has captured and profiled relgion as an sellable attribute against which to target adverts.serendip - Tuesday, October 17, 2017 - link
I'm not sure if you meant that seriously or in jest. AI could be useful for identifying subjects in photos and keeping it in an offline database (for privacy reasons), so you can search for recent photos of baby Peter or Auntie May.narmermenes - Monday, October 16, 2017 - link
The article title states it's a Hands-On review, yet there is almost no hands on content.What about the audio and audio apps?
How does the over-the-top 384kHz / 32-bit audio sound and what is the look/feel/sound of the special ear phones?
What audio player is included and how did it work?
How about the AI development environment? What is it like?
Ian Cutress - Tuesday, October 17, 2017 - link
Where does it say this was a review? We had some small amount of hands-on before launch.axi6ne8us - Monday, October 16, 2017 - link
I still don't get which of the two has the 32bit premium DAC audio?Ian Cutress - Tuesday, October 17, 2017 - link
They both have 32-bit / 384 kHz support.quiksilvr - Monday, October 16, 2017 - link
The Mate 10 Pro nor Porsche has a MicroSD slot nor 3.5mm headphone jack. Perhaps you shouldn't rush articles and proofread them a bit.quiksilvr - Monday, October 16, 2017 - link
Furthermore, the Type C connection is also Displayport 1.2 compliantIan Cutress - Tuesday, October 17, 2017 - link
Was under the impression we were correct based on the information we had. If we second-guessed everything, we'd post next week. The article was updated very soon after posting.mfenn - Monday, October 16, 2017 - link
> The details were (annoyingly) light but it related to memory fragmentation (shouldn’t be an issue with DRAM),Memory fragmentation is a real performance issue. It's not that the DRAM itself cares about being fragmented (it doesn't). However, the memory allocation routines in the kernel and C libraries do care because they have to work harder to satisfy an allocation request. The more CPU cycles allocation consumes, the more power the allocation takes (reduced battery life) and the longer it takes (increased application latency).
wr3zzz - Monday, October 16, 2017 - link
Besides real-time translation, what other ready for market development in AI is NOT about transferring and converting more personal identity into properties owned by for-profit corporations?x3kiwi - Tuesday, October 17, 2017 - link
For some reason this article is not up to the par of usual Anandtech's usual quality."Qualcomm are going to push the concept via the Windows on Snapdragon platform, using the Snapdragon 835. Qualcomm is working with Microsoft.." ?
anawim - Tuesday, October 17, 2017 - link
I thank you for this post and for mentioning automatic audio transcription. As far as I know there is no app that can do that even making use of Google Cloud. What I'd like is the possibility to search through audio/video recordings the way we do now with text in files and obviously being able to do that offline and on the fly.name99 - Tuesday, October 17, 2017 - link
There is an iPhone app that does this called Just Press Record. It will record audio when you press the button (as per the name) and transcribe it LOCALLY (no cloud). It runs on pre-iPhone 8 so doesn't HAVE to use the NPU, but it may be using Apple's generic neural APIs, which use whatever hardware is available and optimal for a particular task.The accuracy seems pretty good in my experience.
anawim - Wednesday, October 18, 2017 - link
I didn't know about this app, and being confined to Apple is out of reach for me to try it out but from what I've read it's more about dictation while I was more about being able to search for words from a general audio/video recording that implies in my opinion an higher grade of difficulties as there could be different voices mixed with other sounds.X-Nemesis - Friday, October 20, 2017 - link
How long until we get phones with IP69?Rdmkr - Saturday, October 21, 2017 - link
The article mentions that the Pro version has a larger screen than the regular version. This is sloppy thinking that falls short of the standards of this site. A 6" 18:9 has a lower screen area than a 5.9" 16:9 screen. The regular mate 10 is the larger screened of the two. In combination with the fact that non-cropped 16:9 media are displayed at about 5.5" equivalent on the pro (another consequence of aspect ratio voodoo), I consider the non-pro version a much, much better buy. I'm avoiding all >16:9 ratio phones for this reason and am very glad Huawei finally produced a small bezeled 16:9 one.sgmuser - Wednesday, October 25, 2017 - link
I am not sure what is great here! This was done and dusted by Microsoft long time back with 900 series phones! It was a better implementation with Windows as OS for desktop as well which is proper Desktop OS. Nothing interesting! Huawei should stick to its original *cheap*affordable* tags rather than AI and what not!JGaviola - Wednesday, October 25, 2017 - link
I have used P9, the P 10 Plus and the Mate 9 Pro and I have always been underwhelmed by Huawei's offerings. The only phone that I did like was the 6P but that wasn't truly theirs. Oh well. I don't think I'll dip my hands in Huawei's phone well in the near future.Ketzilla - Wednesday, October 25, 2017 - link
If you are going to buy then reason will be "Battery" https://www.linkcue.com/leasedline/Bondurant - Tuesday, October 31, 2017 - link
When will the full review be out ?oranos - Wednesday, November 22, 2017 - link
what artificial intelligence? dont make me laugh.