@@siafok6960 i ended up getting a full refund and decided to just stick with my Tab Mini C which didn’t have any issues. Would have liked a 10in tablet but oh well, the Note Air 3 C wasn’t it for me and I can’t afford a Tab Ultra C Pro. So for now I’ll stick with my Tab Mini… its fine for my needs honestly
Thank you very much, Voja. I got 2,4 version and I like it a lot. I was thinking that for e-book maybe it's even better to have slightly lower clocks to have longer battery life. But after seeing this video I believe you are totally right that the cores will switch to higher clocks qiute rearely in some very specific situations and most of the time both versions will run at the same clocks, so very similar battery life. In this case synthetic tests to stress the battery don't make any sense - I doubt anyone plays stressful 3D games on their e-reader (maybe some specific math apps could be an exception).
Yeah, I have a 2.8 Ghz version, but was thinking exactly the same... for my usage it is more likely to be a tiny disadvantage because of battery usage. Good to have the reasoning spelled out for why that is likely not going to be an issue. Very informative video and comments here.
Boox should've been on top of this when they sent you the review sample. If the actual model they put in is the 685 but they sent you the 680 there's a good chance they would've asked for the unit to be sent back and retry with a 685 variant. By sending the 685 but actually shipping the 680 it does feel a bit dishonest, even more so as you had to effectively figure this out and inform them. That said, they did indeed market 2.4GHz and the performance difference is indeed marginal, but it, imo, does set a rather nasty precedent. What's next? Or is it really a one-off?
Nice findings. Even as both CPUs are in the 'minimal specifications' I think the answer from Onyx isn't very satisfying. I don't think it's a coincidence that all reviewers got the slightly faster CPU. I wonder if there is a big difference with the battery life of both CPU versions. Furthermore, I suspect the faster one runs a little shorter without recharging. That being said, for the daily usage it probably doesn't make a huge difference.
When will the video with in-depth analysis be released? Tab Ultra C Pro and comparison with Note Air 3C? The sale starts tomorrow and I can't decide what to buy... But I'm more inclined to purchase Note Air 3C, because it is cheaper, more convenient, has a more matte screen, although it has Tab Ultra C There are also many small advantages, such as a more powerful processor, more RAM and flash memory, physical volume/scroll buttons, a larger battery (?). I think I’d rather use a full-size wireless keyboard so that my hands get less tired, I’ll mainly use the tablet as a monitor for programming
Interesting. I wonder if the Tab Mini C also had two different "builds" or firmware when it came out. The first one I got didn't have the colored icons in NeoReader (the bluish color around the "A" icon) and the "NEW" indicators on books in Library were just black and white (not red). I bought a second one after the screen of my first one developed a terrible refresh problem due to hardware issues. The second one I got has the colored icons in NeoReader and the red "NEW" indicators. I'm hoping this second one will last more than 3 months without problems.
One thing you want to look at is also the CPU clock used on the circuit; as that could affect the chipset performances too. IF you get the same hardware but switch the clock on it you end up affecting every single speed value on every system (north bridge, south bridge, RAM and so on). So you had a good hunch that the chips may be different due to sourcing/supply issues; but it is also possible that what is different is the speed at which these chip are running, because of the clock set by the chipset on the main board is also different (due to different parts).
@@MyDeepGuide Yes, but the clock is not coming from space. The PLL is exposed on a SOC and that is what set up the clock, in conjunction with either a crystal on the main circuit board or a reference VCC signal coming to one of the pins on the package... If the former, it can be a physical crystal or a resonator or a simple timer powered by VCC; which means that it could affect CPU speed. Same for the latter, if you use a VCC signal on one of the pins, if that value fluctuate there is a good chance that the PLL will normalize the signal, and you end up with a different frequency for your CPU. I never took apart a Boox device so I have no idea what is on the board; but there is not just a SOC; there must be power circuitry, filters and other facilities to drive the SOC; and one of these facilities is the clock signal usually.
Wahoo, my test results made it into the video! I am still scratching my head on how my 2.4 ended up with faster storage benchmarks, only difference I can think of is that I have a Sandisk Exteme MicroSD inserted (1TB U3 V30 A2), but that should not have had any impact on the internal storage speed tests. Perhaps some of the internal storage ICs came from different suppliers as well, though the speed difference is not humanly perceptible for any normal everyday operations. I remain jealous of those who received the 2.8GHz Golden Ticket with their Chocolate Bar, but I am quite happy with my 2.4GHz NA3C so far, I like it much better than my "soon-to-be-hand-me-down-Christmas-gift" Tab Ultra. Battery life is good-enough for my current use, I am recharging every 2-3 days as long as I'm not watching video... Video will eat battery with ravenous appetite. Thanks for following up on all this!
Hi, just had with two verisons side by side. Besides CPU difference, they have slightly different screens. The one with 2.8 has rougher screen. Also white is warmer on it and a bit brighter. Even without back light. The one with 2.4 CPU has smoother screen, colder white and darker screen.
Absurd that a company that sells devices cannot say the specifications his own devices. It seems that boox is not a BIG company, so it cannot choose specifically what components are inside boox devices... Have the manufacturing factories the "power and control" about specifications? Curious.
No, it's probably a simple case of them requisitioning SM6225 SoC's and not specifying which variant they wanted. So they got a mixed supply of SM6225 and SM6225-AD chips. But I do agree that it is absurd that they say that due to licencing agreements they are not at liberty to disclose the exact chips being used in the device. That's unusual.
How do the SD card read and write speeds compare with the internal storage? I've added an SD card with an 80GB Zotero collection and it is so slow, it's basically unusable. I'm wondering if its my card or the limitations of the device. Internal I have 40MB/s write and 535MB/s read. I'm running SD cards speed tests (SD Card Test Pro) now using my SanDisk 256GB Extreme microSDXC UHS-I Memory Card * Up to 190MB/s, C10, U3, V30, 4K, 5K, A2, Micro SD Card - SDSQXAV-256G-GN6MA The quicktest I've run has 4MB/s write and 6MB/s read. Longer test will take way longer on the order of 6 days
A more interesting question would be wrt. battery life: frequency squared is the power consumption of a CPU and a 17% increase in frequency, given the same manufacturing process, might translate into 36% more power consumption...
That would be an F. Try again, but this time gather actual data, and don't base your conclusion on an opinion, but rather verifiable facts. Hint 1:Where did you get the 17% increase in frequency? Are you sure that is a correct observation? Hint 2: What did you use to base the correlation between the supposed "increase in frequency" and power consumption? Under which conditions? You need to use actual data, knowledge and facts, if you want to make claims like those.
@@MyDeepGuideFirstly, 2.8/2.4 = ? (blessed be he who can do math) and f^2 = power consumption is a well-known correlation in electrical engineering... I am not sure why I am attacked here, I asked a very valid question: What about the battery life on the faster SoC?
@@MyDeepGuideAlso note that this f^2 increase is on that component only, not the entire system. So, why didn't you bother to ask people about their battery life experience? Given that the Air 3C is on the shorter end of that and, according to your review, has plenty of power anyway, that was, to me, the only question that mattered.
What you are saying is factually wrong and your assumptions are wrong. It is not a simple relation of 2.8/2.4, because it is an SoC, System on a chip, that contains both the CPU and the GPU. As I have described and talked about in the video, the CPUs are identical 8-core design with 4 efficiency cores running at 1.9GHz on both, and the 4 performance cores turboing up to 2.4GHz or 2.8GHz only in the most demanding tasks. Regular operational frequency is identical on both CPUs.. The GPU is identical, but we have a clock difference of about 145MHz, which is also simply the maximum operational frequency, but regular operational frequency is the same on both GPUs. So, you can clearly see that the calculations needed to actually get to a somewhat accurate result are far, far more comprehensive than what you suggested. As far as consumption goes, well that's wrong as well, because of the same, above mentioned reasons. You are not being attacked, but your statements are wrong, and I am clarifying that for those who might be interested in actual facts, and not in simplified assumptions. As for the Battery life, in regular, normal workloads the differences will be 0%, because nothing, not even writing forces either the performance cores nor the GPU to turbo up to the maximum frequencies. So that means, that both variations of the device will work in the similar frequency range, performing at the same exact speeds, and since both are the same 6nm architecture and are using the same voltage (because of the same operational clock speeds) the consumption will be the same.
@@MyDeepGuideThen go measure it! That's all I'm asking. I know very well what a SoC is, and I know that this is a simplified view. But: The question whether I get more performance out of a SoC that can turbo up to 2.8GHz whether one that can go up to 2.4GHz is irrelevant. All I am asking for is for a comparison of battery life, since 17% of frequency boost may translate ("may" = "yes or no") into up to 36% more power consumption. Also, while we're at it, let me give you a simple business 101: Is it a good idea to attack what essentially boils down to your customer base? Hint: No. The main question, as I see it, is the following. (1) it is clear that 2.8GHz is faster than 2.4GHz. (2) Getting a 2.8GHz is a lottery (since the system is officially specced to 2.4GHz). (3) That's the one: Should I demand a replacement if my device [arriving soon] comes with a 2.8GHz SOC over a 2.4GHz, considering battery life? Speed on an otherwise identical SoC fabric scales with frequency, but like it or not, power consumption scales with frequency squared....
Experience with Note Air 3C... after a weekend. 1. Hardware, etc. * Much quicker than the Note Air 1 (which I gave up in January 2023 because I didn't figure out changes in NeoReader...) * Lovely experience writing with pen on screen * Relative poor battery life: battery dropped 20% in 2 hours by reading a (technical) book in Kindle and extensively using yellow marker * Not easy to distinguish between colors on the screen -- poorer in practice than what it looks like in youtube videos... 2. Software * Google Play -- works fine * I'll minimize installation of apps (Edge + Outlook for occasional use) * Note app: somewhat confusing to use... where is the *delete* choice? After having used lasso tool -- how to get back to selecting pens, etc? * Where is the NeoReader app? Does it only show up when I open relevant files? 3. Design flaws: * Poor location of charging/USB-C Port... necessitates to take the tablet out of the cover for charging * Poor location for storing pen in cover -- too easy to mess up the pen when grabbing the tablet with cover. Why not copy Samsung's covers, where the pen is stored inside of the cover *hinge*? That solution ensures that the pen is completely covered and can not be lost/messed up. * It is somewhat tricky to use finger print reader when the tablet is in the cover. 4. Wishes for more videos... Workflow Tips * Note app... see above (navigation), management, etc. (where are files stored? on tablet, or somewhere on OneDrive?) * NeoReader... overview, where can I start it, etc. * OneDrive: how to manage files. If I want to operate on a file in OneDrive, do I need to download it to the tablet? When I edit a file in NeoReader, how do I store it back in OneDrive in such a way that I can access it from my PC? How to store PDF (or SVG?) versions of Note documents? * MS Teams App... can I join a Teams meeting from the Note Air 3C and share screen where I draw/write in the Note app? Would it work with MS WhiteBoard? OK -- some thoughts and some wishes 🙂
If the package specifications are the ones that matter, mine on the CPU says only "Octa-core". Maybe I can ask for a complete different SoC that has 8 cores :D
Fascinating! Dear Voja, thank you for comprehensive and high quality research! Perfect content. One question. As you are remarkable and very experienced in note taking e-ink devices area expert, did you try to ask Boox about GPU options for battery saving? Or do you plan to ask them when you gather enough data about power consumption in different scenarios? I believe they may hear you as the voice of Boox users and interested persons 🙂
Thank you very much :) I have expressed that several times before, but also yesterday, when I sent them my detailed feedback on the TUC Pro. If they can implement a BSR on/off setting, that would be imperative.
Transparency is very important for a company's integrity. It does not matter if the results would be the same. You are right when you said they were quite defensive and they kind of skimmed the questions. If we switch it around, would they be happy if some consumers bought at a lesser price? No matter how small, it does matter. It also kind of make sense to have the faster better ones out first so the response and reviews would be much better. Having the 2.4 come out after the 2.8 yahoo reviews really smells fishy. Just my 2 cents.
Well, the stuff you said would make sense if the difference was something that was noticable in any way in real-world scenario's. But in reality, outside of synthetic benchmarks, there is absolutely no observable difference between the two.
Also a note of warning; mean and average are very deceiving in most calculations related to statistical measurements... There is a reason why in most scientific publications you look at percentile measurements, and not at median or average, and that is because a 75th percentile or 90th percentile measurement is way more reliable compared to an average, mean or median value, for various reasons that would take hours to explain here (but that can be easily found on any base course on statistics)
Right, well I just checked the 90th percentile against median, and in all cases the value difference was under 1% (except the multi-core of the 2.4 GHz, which was at 2%). So, not really something that changes the outcome in a meaningful way in any of the results, as far as median values go, in this particular data set.
@@MyDeepGuide Thanks for checking the percentile. I suspect you got similar values as the hardware is basically consistent; and the number of submissions was not that high to introduce outliars.
The most impactful difference imo is Bluetooth 5.2 and 5.1. 5.2 allows Bluetooth LE Audio to allow more efficient and more future proofing with earbuds. Also LE Audio is better for music quality.
Right, except that both devices have a BT 5.0, so just because the SoC is capable of that, it doesn't mean anything if the device isn't actually equipped with a BT 5.2 transmitter.
Hi Voya, My 2.8Ghz was replaced , and i recived a 2.8 (and unfortunatly it has the same defact of a gray stripe where the app bar is ) and the 2.4 actionly feels slower. I wish if there is a way that you can try to compare a 2.4 vs the 2.8
I am 99.99% sure that if you would do a blind test, not knowing which device is which, that you would genuinely not be able to see, feel, or discern the difference. It's simply how the mind works.
@@MyDeepGuide well i think i felt a diffarance on my drawing ( wich always give the large notes message) I am.also having an alwful.problems woth notes created on my onld note 3 (which i have sold ) that if i open the here all the note is deleted to blank. )
They are both using the same 6nm manufacturing process and are rated for the same voltage. So the consumption should be, for all intents and purposes, the same.
@@MyDeepGuide they may have the same power consumption (or not), the fact they have the same 6nm or the same voltage will not give us info on their power efficiency.
Seems to me that Onyx is slowly losing control of the many, many, many new device types they keep releasing in what seems almost a monthly schedule at this point.
Hello Voja i got the notification on my note air 2 regarding the 3.5V update , I tried to update but could not download completely due to slow internet speed. but now i am not able to download the new update as the system says ” it is already the latest version” please guide should i wait for the update or should contact boox
this can be controversial depends on different point of view. Yes, what Boox declared in anyway met as 2.4 is declared, but some can have 2.8 which can be around 10% higher performance in bench mark values not actual use difference that everybody can feel. But, this is very limited point of view that consider customers who think in that way. However, no one can limit the type of customers who buy this product. Someone want to buy something better even if it is very minimal. That is why there are people who do some effort to do overclock their CPU and Ram ... we can not judge them as strange or extraordinary. They are all same customers. I love your investigation, and it is great video. but I believe some comments show how people think about this difference that Boox unintentionally made. this is what it is, and customers are not the same. They all have different opinion, and this is free market. And, seller can not decide or limit type of customer. I know what could happen, but it is very pity that there is anyway difference. I think I receive 2.8 after checking CPU-Z. Not sure if CPU-Z tell me the truth, but I know that I would not be happy if I have 2.4.... cause I pay same as others who can have 2.8....
@@MyDeepGuide That is what I figured! Just to let you know, I ran Geekbench 6, and had nearly identical results to the user submitted version, 1 point lower on single core, a bit higher on multicore. What I would REALLY like to know, is if the 2.8 Ghz has worse battery performance due to higher clock speeds generally requiring higher power draw!?
I disagree wholeheartedly: It would not be unusual to equip the first devices with better components so that potential reviewers can achieve better performance results. There are also reviewers who buy the devices directly on the market and don't have them sent to them. As you showed us the numbers look better - and that can definitely be of interest to a company because people look at the numbers and less at everyday performance. Curious if DESTA results would be different - and these are also pure naked numbers. I was able to observe something like this with USB sticks: the first sold devices were incredibly fast. After six months to a year the same new products were significantly(!) slower. They do the same thing with SSDs and when asked about it by tech magazines they at least admit it (WD, Samsung etc). We will never know whether this was intentional or an accident. And the marketing guy who answered your email probably won't know that either. Thank you for your investigation! 🤗
I disagree, there is a quantifiable performance difference and and depending on the workload it will definitely be felt. Also, you don't need to stress all the cores to see the performance difference, although it will be particularly felt more the more cores you use. In most cases there will be a core used more than others and it will be noticeable if one can reach 2.4 whilst the other 2.8. This is why people overclock these things. Obviously people who got the 2.4Ghz will not be happy. They got a much older variant of a processor, with different capabilities. One has the FastConnect 6100, the other the FastConnect 6200. Their default APIs also support different things, different resolution cameras, different screen refresh rates, NFC or not etc.. Being the 685 clearly more capable. Power consumption may also differ. This doesn't mean that all these APIs make a difference if they are not implemented in hardware but the difference exists. I think it's dishonest and needs more clarification than what they provided. Even if it is that some landed with a weaker performance, it is clear that they did not want people returning devices to try to catch their better variant, which is most certainly what is going to happen now. They didn't want to produce two variants because NO ONE would select the 680 and the difference in performance wouldn't allow for any significative cost difference. This way they can just sell them all. Now they pay with the return costs, and learn their lesson. Other brands have done this kind of thing. Samsung is known for releasing different CPU versions in different parts of the world, for the same phone release, but at least they do make it clear the phone is a different specific model and this specific model number is available, and everyone knows before hand. This was just dishonest, they try to let it pass just so their sales don't get hurt in any way. A good brand has something that is called transparency. And this was not transparent at all. They act as if their customers are stupid and cannot decide for themselves. They've purposefully hidden this detail. There is no way to not have done this on purpose. This affects sales directly. This was a purposeful decision and they need redemption from it. The customer has to decide whether the performance difference is something they want or not, it's not their place to decide that for customers. Their role is to provide a specification of the hardware, the customer decides whether the hardware fulfills his needs or not.
Specifications says 2,4 GHz and that's what people got. What's the issue ? And the features you mentioned would be important for a smartphone, but for the eBook reader/Note taker - it's meh... Psychologically it could bother some people - I understand that, but realistically - it's basically nothing.
@@YuryIlinsky You've been lied to. It's not your decision to say it matters or not. Like it was not boox. There is something called transparency and that means letting your customer decide. To me it matters that I have the 10% additional performance. I do 3D scans, both the CPU and the graphics card make a difference. in the frames I get with my older scanners. A 2FPS difference means I can move faster, and have less tracking issues. The customer has to decide, not the brand. They lied for their benefit. There is in my opinion no defense to this. And I can't understand how you can claim that 10% is not a HUGE difference. IT IS MASSIVE. Specifications says 2.4GHZ but I got 2.8GHZ. If I had the 2.4GHZ I would have returned it immediately in order to try to get the 2.8GHZ variant. In all other cases the 2.8GHZ has higher value. This happens a lot as I said in smartphone markets and if you check the prices you see a clear difference in the same model in the ones with aged CPUs vs the ones with the better ones. Check S21FE variants for ex in used marketplaces. Your re-sale value will be lower. You've been deceived. If the deception or not is going to affect your work, most likely not, but that does not mean you haven't been deceived. You thought the device you were purchasing had no options and it does. You got the worst option. In the same sense someone might purchase the C pro instead of the C. Two different hardware specs, two different models. Transparency.... Take not that I do not mean to not be happy or satisfied if you are, that's great then, I just absolutely disagree with any form of defense towards a company that did this. It was for their benefit that they intentionally did not make users aware.
@@duartemachado1158Duarte, what is happening here is that you feel somehow damaged by different chips, though the company is not promising more than what it’s offering. My diesel car has 150 bhp but the engine they had at the time was the one from 170 bhp with the software from the one from the 150 bhp (so, tuned down). For me as a client it’s the same and the same costs as the original one, but it’s known to happen. On this case, it’s the same: you get at least what the company promises; on same cases more speed and eventually more or less the same battery. So? More APIs etc etc? No because the OS will be the same, even if the hardware would be more capable. Zero difference.
you do know that variations of models are in every mass produced products (not to the extent of cpu difference in most cases) but they do exist and if the company dont overmarket the device then it is fine. you could have scientific measurements to every aspect of a device or a product and results will vary ( like the density of the coating or the finish or the screen light uniformity ... etc) the main and most important thing is that THE DIFFERENCE MATTERS(noticeable) OR NOT (for 99% of the consumers) @@duartemachado1158 i agree that there should be more transparency
Excellent journalism Voja, well done!
Thank you Kit! :)
I sent mine back, too many quality control issues... dead pixels, dark shadow on the left side, low brightness, distorted image and constant crashing.
Have they resent you proper one?
@@siafok6960 i ended up getting a full refund and decided to just stick with my Tab Mini C which didn’t have any issues. Would have liked a 10in tablet but oh well, the Note Air 3 C wasn’t it for me and I can’t afford a Tab Ultra C Pro. So for now I’ll stick with my Tab Mini… its fine for my needs honestly
Nicely done! Fortunately, no one got shortchanged. Some users just got a minor bonus.
Thank you very much, Voja. I got 2,4 version and I like it a lot. I was thinking that for e-book maybe it's even better to have slightly lower clocks to have longer battery life.
But after seeing this video I believe you are totally right that the cores will switch to higher clocks qiute rearely in some very specific situations and most of the time both versions will run at the same clocks, so very similar battery life.
In this case synthetic tests to stress the battery don't make any sense - I doubt anyone plays stressful 3D games on their e-reader (maybe some specific math apps could be an exception).
Exactly the point I was trying to iterate here, and I'm happy people understand it :)
Yeah, I have a 2.8 Ghz version, but was thinking exactly the same... for my usage it is more likely to be a tiny disadvantage because of battery usage. Good to have the reasoning spelled out for why that is likely not going to be an issue. Very informative video and comments here.
Boox should've been on top of this when they sent you the review sample. If the actual model they put in is the 685 but they sent you the 680 there's a good chance they would've asked for the unit to be sent back and retry with a 685 variant.
By sending the 685 but actually shipping the 680 it does feel a bit dishonest, even more so as you had to effectively figure this out and inform them.
That said, they did indeed market 2.4GHz and the performance difference is indeed marginal, but it, imo, does set a rather nasty precedent.
What's next? Or is it really a one-off?
2.8ghz vs 2.4ghz maybe is a marginal perf difference, but more importantly for many is how it affects the battery life.
@@gregoriusmike I very much doubt this will make a noticeable difference.
Thanks ! Once again your conclusion is clear and precise. Well done. :-)
Nice findings.
Even as both CPUs are in the 'minimal specifications' I think the answer from Onyx isn't very satisfying. I don't think it's a coincidence that all reviewers got the slightly faster CPU. I wonder if there is a big difference with the battery life of both CPU versions. Furthermore, I suspect the faster one runs a little shorter without recharging.
That being said, for the daily usage it probably doesn't make a huge difference.
When will the video with in-depth analysis be released? Tab Ultra C Pro and comparison with Note Air 3C? The sale starts tomorrow and I can't decide what to buy... But I'm more inclined to purchase Note Air 3C, because it is cheaper, more convenient, has a more matte screen, although it has Tab Ultra C There are also many small advantages, such as a more powerful processor, more RAM and flash memory, physical volume/scroll buttons, a larger battery (?). I think I’d rather use a full-size wireless keyboard so that my hands get less tired, I’ll mainly use the tablet as a monitor for programming
Such dedication and attention to detail Voja! Thank you!
Interesting. I wonder if the Tab Mini C also had two different "builds" or firmware when it came out. The first one I got didn't have the colored icons in NeoReader (the bluish color around the "A" icon) and the "NEW" indicators on books in Library were just black and white (not red). I bought a second one after the screen of my first one developed a terrible refresh problem due to hardware issues. The second one I got has the colored icons in NeoReader and the red "NEW" indicators. I'm hoping this second one will last more than 3 months without problems.
Curious if the screen on the new unit is holding up better? Seems like you commented 3 months ago.
One thing you want to look at is also the CPU clock used on the circuit; as that could affect the chipset performances too. IF you get the same hardware but switch the clock on it you end up affecting every single speed value on every system (north bridge, south bridge, RAM and so on).
So you had a good hunch that the chips may be different due to sourcing/supply issues; but it is also possible that what is different is the speed at which these chip are running, because of the clock set by the chipset on the main board is also different (due to different parts).
This is an SoC, there is no chipset. Everything is contained in the die itself.
@@MyDeepGuide Yes, but the clock is not coming from space. The PLL is exposed on a SOC and that is what set up the clock, in conjunction with either a crystal on the main circuit board or a reference VCC signal coming to one of the pins on the package...
If the former, it can be a physical crystal or a resonator or a simple timer powered by VCC; which means that it could affect CPU speed.
Same for the latter, if you use a VCC signal on one of the pins, if that value fluctuate there is a good chance that the PLL will normalize the signal, and you end up with a different frequency for your CPU.
I never took apart a Boox device so I have no idea what is on the board; but there is not just a SOC; there must be power circuitry, filters and other facilities to drive the SOC; and one of these facilities is the clock signal usually.
Wahoo, my test results made it into the video! I am still scratching my head on how my 2.4 ended up with faster storage benchmarks, only difference I can think of is that I have a Sandisk Exteme MicroSD inserted (1TB U3 V30 A2), but that should not have had any impact on the internal storage speed tests. Perhaps some of the internal storage ICs came from different suppliers as well, though the speed difference is not humanly perceptible for any normal everyday operations.
I remain jealous of those who received the 2.8GHz Golden Ticket with their Chocolate Bar, but I am quite happy with my 2.4GHz NA3C so far, I like it much better than my "soon-to-be-hand-me-down-Christmas-gift" Tab Ultra. Battery life is good-enough for my current use, I am recharging every 2-3 days as long as I'm not watching video... Video will eat battery with ravenous appetite.
Thanks for following up on all this!
Thank you for sharing your results! They were super helpful and insightful! I'm also thoroughly enjoying my NA3C every day :)
What is your SD card performance like?
Hi, just had with two verisons side by side. Besides CPU difference, they have slightly different screens. The one with 2.8 has rougher screen. Also white is warmer on it and a bit brighter. Even without back light. The one with 2.4 CPU has smoother screen, colder white and darker screen.
Absurd that a company that sells devices cannot say the specifications his own devices.
It seems that boox is not a BIG company, so it cannot choose specifically what components are inside boox devices...
Have the manufacturing factories the "power and control" about specifications?
Curious.
No, it's probably a simple case of them requisitioning SM6225 SoC's and not specifying which variant they wanted. So they got a mixed supply of SM6225 and SM6225-AD chips.
But I do agree that it is absurd that they say that due to licencing agreements they are not at liberty to disclose the exact chips being used in the device. That's unusual.
@@MyDeepGuideyou're right, seems logical.
How do the SD card read and write speeds compare with the internal storage? I've added an SD card with an 80GB Zotero collection and it is so slow, it's basically unusable. I'm wondering if its my card or the limitations of the device. Internal I have 40MB/s write and 535MB/s read.
I'm running SD cards speed tests (SD Card Test Pro) now using my SanDisk 256GB Extreme microSDXC UHS-I Memory Card * Up to 190MB/s, C10, U3, V30, 4K, 5K, A2, Micro SD Card - SDSQXAV-256G-GN6MA
The quicktest I've run has 4MB/s write and 6MB/s read. Longer test will take way longer on the order of 6 days
Thank you very much for your work❣️
A more interesting question would be wrt. battery life: frequency squared is the power consumption of a CPU and a 17% increase in frequency, given the same manufacturing process, might translate into 36% more power consumption...
That would be an F. Try again, but this time gather actual data, and don't base your conclusion on an opinion, but rather verifiable facts.
Hint 1:Where did you get the 17% increase in frequency? Are you sure that is a correct observation?
Hint 2: What did you use to base the correlation between the supposed "increase in frequency" and power consumption? Under which conditions?
You need to use actual data, knowledge and facts, if you want to make claims like those.
@@MyDeepGuideFirstly, 2.8/2.4 = ? (blessed be he who can do math) and f^2 = power consumption is a well-known correlation in electrical engineering... I am not sure why I am attacked here, I asked a very valid question: What about the battery life on the faster SoC?
@@MyDeepGuideAlso note that this f^2 increase is on that component only, not the entire system. So, why didn't you bother to ask people about their battery life experience? Given that the Air 3C is on the shorter end of that and, according to your review, has plenty of power anyway, that was, to me, the only question that mattered.
What you are saying is factually wrong and your assumptions are wrong. It is not a simple relation of 2.8/2.4, because it is an SoC, System on a chip, that contains both the CPU and the GPU.
As I have described and talked about in the video, the CPUs are identical 8-core design with 4 efficiency cores running at 1.9GHz on both, and the 4 performance cores turboing up to 2.4GHz or 2.8GHz only in the most demanding tasks. Regular operational frequency is identical on both CPUs..
The GPU is identical, but we have a clock difference of about 145MHz, which is also simply the maximum operational frequency, but regular operational frequency is the same on both GPUs.
So, you can clearly see that the calculations needed to actually get to a somewhat accurate result are far, far more comprehensive than what you suggested.
As far as consumption goes, well that's wrong as well, because of the same, above mentioned reasons.
You are not being attacked, but your statements are wrong, and I am clarifying that for those who might be interested in actual facts, and not in simplified assumptions.
As for the Battery life, in regular, normal workloads the differences will be 0%, because nothing, not even writing forces either the performance cores nor the GPU to turbo up to the maximum frequencies. So that means, that both variations of the device will work in the similar frequency range, performing at the same exact speeds, and since both are the same 6nm architecture and are using the same voltage (because of the same operational clock speeds) the consumption will be the same.
@@MyDeepGuideThen go measure it! That's all I'm asking. I know very well what a SoC is, and I know that this is a simplified view. But: The question whether I get more performance out of a SoC that can turbo up to 2.8GHz whether one that can go up to 2.4GHz is irrelevant. All I am asking for is for a comparison of battery life, since 17% of frequency boost may translate ("may" = "yes or no") into up to 36% more power consumption. Also, while we're at it, let me give you a simple business 101: Is it a good idea to attack what essentially boils down to your customer base? Hint: No. The main question, as I see it, is the following. (1) it is clear that 2.8GHz is faster than 2.4GHz. (2) Getting a 2.8GHz is a lottery (since the system is officially specced to 2.4GHz). (3) That's the one: Should I demand a replacement if my device [arriving soon] comes with a 2.8GHz SOC over a 2.4GHz, considering battery life? Speed on an otherwise identical SoC fabric scales with frequency, but like it or not, power consumption scales with frequency squared....
Experience with Note Air 3C... after a weekend.
1. Hardware, etc.
* Much quicker than the Note Air 1 (which I gave up in January 2023 because I didn't figure out changes in NeoReader...)
* Lovely experience writing with pen on screen
* Relative poor battery life: battery dropped 20% in 2 hours by reading a (technical) book in Kindle and extensively using yellow marker
* Not easy to distinguish between colors on the screen -- poorer in practice than what it looks like in youtube videos...
2. Software
* Google Play -- works fine
* I'll minimize installation of apps (Edge + Outlook for occasional use)
* Note app: somewhat confusing to use... where is the *delete* choice? After having used lasso tool -- how to get back to selecting pens, etc?
* Where is the NeoReader app? Does it only show up when I open relevant files?
3. Design flaws:
* Poor location of charging/USB-C Port... necessitates to take the tablet out of the cover for charging
* Poor location for storing pen in cover -- too easy to mess up the pen when grabbing the tablet with cover. Why not copy Samsung's covers, where the pen is stored inside of the cover *hinge*? That solution ensures that the pen is completely covered and can not be lost/messed up.
* It is somewhat tricky to use finger print reader when the tablet is in the cover.
4. Wishes for more videos... Workflow Tips
* Note app... see above (navigation), management, etc. (where are files stored? on tablet, or somewhere on OneDrive?)
* NeoReader... overview, where can I start it, etc.
* OneDrive: how to manage files. If I want to operate on a file in OneDrive, do I need to download it to the tablet? When I edit a file in NeoReader, how do I store it back in OneDrive in such a way that I can access it from my PC? How to store PDF (or SVG?) versions of Note documents?
* MS Teams App... can I join a Teams meeting from the Note Air 3C and share screen where I draw/write in the Note app? Would it work with MS WhiteBoard?
OK -- some thoughts and some wishes 🙂
If the package specifications are the ones that matter, mine on the CPU says only "Octa-core". Maybe I can ask for a complete different SoC that has 8 cores :D
Fascinating! Dear Voja, thank you for comprehensive and high quality research! Perfect content.
One question. As you are remarkable and very experienced in note taking e-ink devices area expert, did you try to ask Boox about GPU options for battery saving? Or do you plan to ask them when you gather enough data about power consumption in different scenarios? I believe they may hear you as the voice of Boox users and interested persons 🙂
Thank you very much :) I have expressed that several times before, but also yesterday, when I sent them my detailed feedback on the TUC Pro. If they can implement a BSR on/off setting, that would be imperative.
Transparency is very important for a company's integrity. It does not matter if the results would be the same. You are right when you said they were quite defensive and they kind of skimmed the questions. If we switch it around, would they be happy if some consumers bought at a lesser price? No matter how small, it does matter. It also kind of make sense to have the faster better ones out first so the response and reviews would be much better. Having the 2.4 come out after the 2.8 yahoo reviews really smells fishy. Just my 2 cents.
Well, the stuff you said would make sense if the difference was something that was noticable in any way in real-world scenario's. But in reality, outside of synthetic benchmarks, there is absolutely no observable difference between the two.
I believe this was a video to clean the Boox image after all this problem
Also a note of warning; mean and average are very deceiving in most calculations related to statistical measurements... There is a reason why in most scientific publications you look at percentile measurements, and not at median or average, and that is because a 75th percentile or 90th percentile measurement is way more reliable compared to an average, mean or median value, for various reasons that would take hours to explain here (but that can be easily found on any base course on statistics)
Right, well I just checked the 90th percentile against median, and in all cases the value difference was under 1% (except the multi-core of the 2.4 GHz, which was at 2%). So, not really something that changes the outcome in a meaningful way in any of the results, as far as median values go, in this particular data set.
@@MyDeepGuide Thanks for checking the percentile. I suspect you got similar values as the hardware is basically consistent; and the number of submissions was not that high to introduce outliars.
The most impactful difference imo is Bluetooth 5.2 and 5.1. 5.2 allows Bluetooth LE Audio to allow more efficient and more future proofing with earbuds. Also LE Audio is better for music quality.
Right, except that both devices have a BT 5.0, so just because the SoC is capable of that, it doesn't mean anything if the device isn't actually equipped with a BT 5.2 transmitter.
Got it, thanks for explaining that.
Hi Voya, My 2.8Ghz was replaced , and i recived a 2.8 (and unfortunatly it has the same defact of a gray stripe where the app bar is ) and the 2.4 actionly feels slower.
I wish if there is a way that you can try to compare a 2.4 vs the 2.8
I am 99.99% sure that if you would do a blind test, not knowing which device is which, that you would genuinely not be able to see, feel, or discern the difference. It's simply how the mind works.
@@MyDeepGuide well i think i felt a diffarance on my drawing ( wich always give the large notes message)
I am.also having an alwful.problems woth notes created on my onld note 3 (which i have sold ) that if i open the here all the note is deleted to blank. )
i wonder if both chips have the same power consumption. considering one chip is two years newer than the older model.
They are both using the same 6nm manufacturing process and are rated for the same voltage. So the consumption should be, for all intents and purposes, the same.
@@MyDeepGuide they may have the same power consumption (or not), the fact they have the same 6nm or the same voltage will not give us info on their power efficiency.
Seems to me that Onyx is slowly losing control of the many, many, many new device types they keep releasing in what seems almost a monthly schedule at this point.
Hello Voja
i got the notification on my note air 2 regarding the 3.5V update , I tried to update but could not download completely due to slow internet speed. but now i am not able to download the new update as the system says ” it is already the latest version”
please guide should i wait for the update or should contact boox
You should contact Boox.
this can be controversial depends on different point of view. Yes, what Boox declared in anyway met as 2.4 is declared, but some can have 2.8 which can be around 10% higher performance in bench mark values not actual use difference that everybody can feel. But, this is very limited point of view that consider customers who think in that way. However, no one can limit the type of customers who buy this product. Someone want to buy something better even if it is very minimal. That is why there are people who do some effort to do overclock their CPU and Ram ... we can not judge them as strange or extraordinary. They are all same customers. I love your investigation, and it is great video. but I believe some comments show how people think about this difference that Boox unintentionally made. this is what it is, and customers are not the same. They all have different opinion, and this is free market. And, seller can not decide or limit type of customer. I know what could happen, but it is very pity that there is anyway difference. I think I receive 2.8 after checking CPU-Z. Not sure if CPU-Z tell me the truth, but I know that I would not be happy if I have 2.4.... cause I pay same as others who can have 2.8....
Pls do same review for MacBook Pro M1 Pro M2 Pro M3 Pro :D
According to CPUz my note 3c came with a 665 not a 680...
Same for me. 2.4 GHz Snapdragon 665 (Kryo 260 r10p4)
My device got these specs too! @@YuryIlinsky
🤔
CPU-Z is very unreliable at identifying SoC systems. There is a reason why I didn't use it.
@@MyDeepGuide That is what I figured! Just to let you know, I ran Geekbench 6, and had nearly identical results to the user submitted version, 1 point lower on single core, a bit higher on multicore. What I would REALLY like to know, is if the 2.8 Ghz has worse battery performance due to higher clock speeds generally requiring higher power draw!?
I disagree wholeheartedly: It would not be unusual to equip the first devices with better components so that potential reviewers can achieve better performance results. There are also reviewers who buy the devices directly on the market and don't have them sent to them.
As you showed us the numbers look better - and that can definitely be of interest to a company because people look at the numbers and less at everyday performance. Curious if DESTA results would be different - and these are also pure naked numbers.
I was able to observe something like this with USB sticks: the first sold devices were incredibly fast. After six months to a year the same new products were significantly(!) slower. They do the same thing with SSDs and when asked about it by tech magazines they at least admit it (WD, Samsung etc).
We will never know whether this was intentional or an accident. And the marketing guy who answered your email probably won't know that either.
Thank you for your investigation! 🤗
I disagree, there is a quantifiable performance difference and and depending on the workload it will definitely be felt. Also, you don't need to stress all the cores to see the performance difference, although it will be particularly felt more the more cores you use. In most cases there will be a core used more than others and it will be noticeable if one can reach 2.4 whilst the other 2.8. This is why people overclock these things.
Obviously people who got the 2.4Ghz will not be happy. They got a much older variant of a processor, with different capabilities. One has the FastConnect 6100, the other the FastConnect 6200.
Their default APIs also support different things, different resolution cameras, different screen refresh rates, NFC or not etc.. Being the 685 clearly more capable.
Power consumption may also differ. This doesn't mean that all these APIs make a difference if they are not implemented in hardware but the difference exists.
I think it's dishonest and needs more clarification than what they provided.
Even if it is that some landed with a weaker performance, it is clear that they did not want people returning devices to try to catch their better variant, which is most certainly what is going to happen now. They didn't want to produce two variants because NO ONE would select the 680 and the difference in performance wouldn't allow for any significative cost difference. This way they can just sell them all.
Now they pay with the return costs, and learn their lesson.
Other brands have done this kind of thing. Samsung is known for releasing different CPU versions in different parts of the world, for the same phone release, but at least they do make it clear the phone is a different specific model and this specific model number is available, and everyone knows before hand. This was just dishonest, they try to let it pass just so their sales don't get hurt in any way.
A good brand has something that is called transparency. And this was not transparent at all. They act as if their customers are stupid and cannot decide for themselves. They've purposefully hidden this detail. There is no way to not have done this on purpose. This affects sales directly. This was a purposeful decision and they need redemption from it. The customer has to decide whether the performance difference is something they want or not, it's not their place to decide that for customers. Their role is to provide a specification of the hardware, the customer decides whether the hardware fulfills his needs or not.
Specifications says 2,4 GHz and that's what people got. What's the issue ?
And the features you mentioned would be important for a smartphone, but for the eBook reader/Note taker - it's meh...
Psychologically it could bother some people - I understand that, but realistically - it's basically nothing.
@@YuryIlinsky You've been lied to. It's not your decision to say it matters or not. Like it was not boox. There is something called transparency and that means letting your customer decide. To me it matters that I have the 10% additional performance. I do 3D scans, both the CPU and the graphics card make a difference. in the frames I get with my older scanners. A 2FPS difference means I can move faster, and have less tracking issues.
The customer has to decide, not the brand. They lied for their benefit.
There is in my opinion no defense to this. And I can't understand how you can claim that 10% is not a HUGE difference. IT IS MASSIVE.
Specifications says 2.4GHZ but I got 2.8GHZ. If I had the 2.4GHZ I would have returned it immediately in order to try to get the 2.8GHZ variant. In all other cases the 2.8GHZ has higher value. This happens a lot as I said in smartphone markets and if you check the prices you see a clear difference in the same model in the ones with aged CPUs vs the ones with the better ones. Check S21FE variants for ex in used marketplaces.
Your re-sale value will be lower.
You've been deceived. If the deception or not is going to affect your work, most likely not, but that does not mean you haven't been deceived. You thought the device you were purchasing had no options and it does. You got the worst option. In the same sense someone might purchase the C pro instead of the C. Two different hardware specs, two different models. Transparency....
Take not that I do not mean to not be happy or satisfied if you are, that's great then, I just absolutely disagree with any form of defense towards a company that did this. It was for their benefit that they intentionally did not make users aware.
@@duartemachado1158Duarte, what is happening here is that you feel somehow damaged by different chips, though the company is not promising more than what it’s offering. My diesel car has 150 bhp but the engine they had at the time was the one from 170 bhp with the software from the one from the 150 bhp (so, tuned down). For me as a client it’s the same and the same costs as the original one, but it’s known to happen. On this case, it’s the same: you get at least what the company promises; on same cases more speed and eventually more or less the same battery. So? More APIs etc etc? No because the OS will be the same, even if the hardware would be more capable. Zero difference.
you're reaching so hard@@duartemachado1158
you do know that variations of models are in every mass produced products (not to the extent of cpu difference in most cases) but they do exist and if the company dont overmarket the device then it is fine. you could have scientific measurements to every aspect of a device or a product and results will vary ( like the density of the coating or the finish or the screen light uniformity ... etc) the main and most important thing is that THE DIFFERENCE MATTERS(noticeable) OR NOT (for 99% of the consumers) @@duartemachado1158 i agree that there should be more transparency
Minecraft smoother with 2.8😂
Will it bend.😏
You might end up with $500 paperweight 😮
@learningcheme then he/she will perform "will it blend" 😂
First!