Pixel 3 Thoughts
I ended up not buying the Pixel 3 this time around. And while I'm still tempted, because it still has everything I had wanted in the Pixel 2, I'm also finding more reasons not too. It actually all clicked in place when a reviewer pointed out that the Pixel 1 is still a decent phone.
Now, my reasoning is not that, since the Pixel 1 is still a good phone 2 generations later, that I think that the Pixel 2 is still good for another generation or perhaps more. In fact, I think it is the opposite. The Pixel 1 fares favorably because the specs haven't really moved all that much. But, looking deeper I'd put it this way; in terms of most specs like all Pixels were basically on par for the times, except RAM. The Pixel 1 had 4GB at a time when it was still rare to see that in a phone. It was ahead of its time. The Pixel 2 had 4GB at a time when 6GB phones were either just out or on the horizon, so just kind of sticking with the times, and the Pixel 3 brings 4GB again. But, now it is clearly behind the times.
As the Pixel lineup pushes the software boundaries, that 4GB is becoming increasingly more stifling. I'm certain, over time, Google will nail things down and get the performance back under control in the cases where it is running wild on the Pixel 3... in short, Pixel is back to being behind the times in terms of RAM. Google's release cycle also places them right near the end of the life of SoC's. This likely helps them save costs, but it also means that new chipsets are always unveiled in the months following the release. And some of the improvements in the next round of Snapdragon processor's not only sound bigger than some previous iterations. Some sound much more up Google's alley like AI computer power.
In short, it seems highly plausible that the next iteration will feature not only a more modest upgrade in the chipset than in the past, but it will likely be the first new flagship Pixel phone to increase RAM. Now, I suspect it will only go up to 6GB rather than 8GB. But, that is still a 50% increase in RAM. And it will likely be faster RAM as well. As SoC vendors make the base chips more AI friendly, I also expect to see Google leveraging some of that power. I know they have some of their own chips for some tasks. But, more dedicated compute power for AI related tasks is definitely playing into Google's hands.
Now, why I don't think they will jump to 8GB? Two reasons. Supporting past models and keeping the costs low. The Pixel seems positioned rather intentionally to be a late bloomer phone to take advantage of price drops on components and show up the competition on the same hardware and specs. I don't know if this is why both the 2 and 3 kept the same 4GB of RAM or not. But, it seems feasible.
And, the other advantage is that it makes a lot easier to back port the stuff that they want to the older models as an olive branch of thanks to those who helped them get this far. If the spec bumps are too wild and the new software too dependent on the new power available, then less can go back to the older phones. Sure, they keep a few features as a draw to upgrade or to stem the tide of people buying the older model. And some features will rely on hardware improvements.
But, it seems to me like Google is making an effort to show Pixel owners that buying into their hardware is a thing which will be rewarded for a few years. And keeping certain parameters relatively stable helps keep engineers in more realistic borders to be able to deliver on that.
Though, as implied above. I think all good things must eventually come to an end. The Pixel 1 is basically at the end of support. It seems like a fitting time to start raising the bar on the one spec which has been stuck like a stick in the mud. Also, while the Pixel 3 may have been a little behind the times on RAM, it is only really coming into this year that they will really and truly seem far behind.
I also think the next flagship version of the phone may finally see dual back facing cameras. While Google has pulled off some amazing feats with the single lens, the fact remains that the same tactics or similar ones can likely also be applied, and in some cases applied perhaps with even greater success in a multi-lens system. Beating out every single lens phone 2 years running and even beating out multi-lens phones on certain aspects is great. But, when Google described the technology used to achieve this, my first thought was... the shots can be faster, with less distortion and less impacted by motion if a multi-lens system is used.
Take the super res zoom for instance. Right now it takes up to 9 images (I think) and uses the differences between them and some AI to intelligently improve the resolution of the final image. With multiple lens... well, you can get two pictures from slightly different perspectives at the exact same time. You can be guaranteed to always have perfect image pairs, even if the lens moves. Calibrate it right, and some pictures might come out better from a single pair than from 9 successive exposures. And, you can have 18 in the same conditions as 9 pairs.
For low light, you can have the two lenses taking different length exposure shots at the same time. And again, when you stitch the data together, you'll have much better groups of data. A lens with a shorter exposure time can be used to catch details while the longer exposure one can help apply lighting to the those "burstier" images. Or, you can have them both run the same exposure and combine in the super res zoom feature with the night sight feature.
In short, I'm pretty happy with the camera on the Pixel 2. The Pixel 3 does look noticeably better, but not enough to compel me to buy it. Multi-lens camera modules will be more affordable a year on, as more and more phones head that route. Making it more viable for Google to add it in, without breaking the budget. And, they can once again take that technology which is now mainstream and use their AI smarts to out do others.
My predictions therefore would be that the Pixel 4 (or whatever) will have a snapdragon 855 with 6GB of RAM and a dual lens setup on the rear (with both lenses similar, if not identical). While I think a triple lens system is "possible", the only reason I can think would be for some additional specialty lens type like the wide angle one on the Pixel 3 front side. But, I would think Google would prefer to opt (if possible) for some means of simulating the effect rather than paying for the hardware.
Now, it is crazy early to be predicting what will be in that phone (if it even comes into being).
But, looking how long they've tied themselves to the 4GB RAM and single camera specs. Even with the results they are pulling off, it is hard to think that they won't be a joke if they don't push those specs in the next hardware iteration. And, those spec changes will make for some pretty tangible benefits. I suspect it will be the biggest performance gain since the Pixel line started.
Now, my reasoning is not that, since the Pixel 1 is still a good phone 2 generations later, that I think that the Pixel 2 is still good for another generation or perhaps more. In fact, I think it is the opposite. The Pixel 1 fares favorably because the specs haven't really moved all that much. But, looking deeper I'd put it this way; in terms of most specs like all Pixels were basically on par for the times, except RAM. The Pixel 1 had 4GB at a time when it was still rare to see that in a phone. It was ahead of its time. The Pixel 2 had 4GB at a time when 6GB phones were either just out or on the horizon, so just kind of sticking with the times, and the Pixel 3 brings 4GB again. But, now it is clearly behind the times.
As the Pixel lineup pushes the software boundaries, that 4GB is becoming increasingly more stifling. I'm certain, over time, Google will nail things down and get the performance back under control in the cases where it is running wild on the Pixel 3... in short, Pixel is back to being behind the times in terms of RAM. Google's release cycle also places them right near the end of the life of SoC's. This likely helps them save costs, but it also means that new chipsets are always unveiled in the months following the release. And some of the improvements in the next round of Snapdragon processor's not only sound bigger than some previous iterations. Some sound much more up Google's alley like AI computer power.
In short, it seems highly plausible that the next iteration will feature not only a more modest upgrade in the chipset than in the past, but it will likely be the first new flagship Pixel phone to increase RAM. Now, I suspect it will only go up to 6GB rather than 8GB. But, that is still a 50% increase in RAM. And it will likely be faster RAM as well. As SoC vendors make the base chips more AI friendly, I also expect to see Google leveraging some of that power. I know they have some of their own chips for some tasks. But, more dedicated compute power for AI related tasks is definitely playing into Google's hands.
Now, why I don't think they will jump to 8GB? Two reasons. Supporting past models and keeping the costs low. The Pixel seems positioned rather intentionally to be a late bloomer phone to take advantage of price drops on components and show up the competition on the same hardware and specs. I don't know if this is why both the 2 and 3 kept the same 4GB of RAM or not. But, it seems feasible.
And, the other advantage is that it makes a lot easier to back port the stuff that they want to the older models as an olive branch of thanks to those who helped them get this far. If the spec bumps are too wild and the new software too dependent on the new power available, then less can go back to the older phones. Sure, they keep a few features as a draw to upgrade or to stem the tide of people buying the older model. And some features will rely on hardware improvements.
But, it seems to me like Google is making an effort to show Pixel owners that buying into their hardware is a thing which will be rewarded for a few years. And keeping certain parameters relatively stable helps keep engineers in more realistic borders to be able to deliver on that.
Though, as implied above. I think all good things must eventually come to an end. The Pixel 1 is basically at the end of support. It seems like a fitting time to start raising the bar on the one spec which has been stuck like a stick in the mud. Also, while the Pixel 3 may have been a little behind the times on RAM, it is only really coming into this year that they will really and truly seem far behind.
I also think the next flagship version of the phone may finally see dual back facing cameras. While Google has pulled off some amazing feats with the single lens, the fact remains that the same tactics or similar ones can likely also be applied, and in some cases applied perhaps with even greater success in a multi-lens system. Beating out every single lens phone 2 years running and even beating out multi-lens phones on certain aspects is great. But, when Google described the technology used to achieve this, my first thought was... the shots can be faster, with less distortion and less impacted by motion if a multi-lens system is used.
Take the super res zoom for instance. Right now it takes up to 9 images (I think) and uses the differences between them and some AI to intelligently improve the resolution of the final image. With multiple lens... well, you can get two pictures from slightly different perspectives at the exact same time. You can be guaranteed to always have perfect image pairs, even if the lens moves. Calibrate it right, and some pictures might come out better from a single pair than from 9 successive exposures. And, you can have 18 in the same conditions as 9 pairs.
For low light, you can have the two lenses taking different length exposure shots at the same time. And again, when you stitch the data together, you'll have much better groups of data. A lens with a shorter exposure time can be used to catch details while the longer exposure one can help apply lighting to the those "burstier" images. Or, you can have them both run the same exposure and combine in the super res zoom feature with the night sight feature.
In short, I'm pretty happy with the camera on the Pixel 2. The Pixel 3 does look noticeably better, but not enough to compel me to buy it. Multi-lens camera modules will be more affordable a year on, as more and more phones head that route. Making it more viable for Google to add it in, without breaking the budget. And, they can once again take that technology which is now mainstream and use their AI smarts to out do others.
My predictions therefore would be that the Pixel 4 (or whatever) will have a snapdragon 855 with 6GB of RAM and a dual lens setup on the rear (with both lenses similar, if not identical). While I think a triple lens system is "possible", the only reason I can think would be for some additional specialty lens type like the wide angle one on the Pixel 3 front side. But, I would think Google would prefer to opt (if possible) for some means of simulating the effect rather than paying for the hardware.
Now, it is crazy early to be predicting what will be in that phone (if it even comes into being).
But, looking how long they've tied themselves to the 4GB RAM and single camera specs. Even with the results they are pulling off, it is hard to think that they won't be a joke if they don't push those specs in the next hardware iteration. And, those spec changes will make for some pretty tangible benefits. I suspect it will be the biggest performance gain since the Pixel line started.
Comments
Post a Comment