|Review of the Samsung Galaxy S3 (LTE)|
The new Samsung Galaxy S3 has created quite a bit of buzz with a reported 9 million pre-orders. If that number proves to be true, it blows away the 4 million or so iPhone 4S models that were sold in the first 4 days. But does the device deserve this level of adoration? I got a chance to answer that question when I was able to borrow the pre-launch copy of the phone that Howard Chui had on loan from Samsung Canada.
Last Updated: 16-Jun-2012
Before reading this review, please read Some Thoughts on Phone Reviewing.
The model in question is the
i-747M, which is a dual-core phone with support for LTE (and not the
quad-core HSPA-only device seen outside of North America). Like the
recently-tested HTC One X LTE model, it
uses the Qualcomm MSM8960 chipset, clocked at 1.5 GHz.
Due to this strong similarity to the One X, the HTC phone is the obvious one to compare the S3 to, and I will most certainly do that as best I can. However, itís also not a stretch to compare this phone with the Galaxy S II LTE (which I happen to own, and thus am able to compare side-by-side), and so Iíll be doing that too. Many of my comments, when making the latter comparisons, will be with an eye to those who already own the S II LTE and might have been thinking of blowing a large sum of money to do an early upgrade to the S3 (or even to buy one outright).
Also see Howard Chui's review of the
Galaxy S3 at Howard Forums.
Click on this link for a full description of RF Performance, and how to interpret it.
Try as I might, I couldnít find any major difference in RF performance between the S3 and my S II LTE. Even in the Service Monitor screen the signal strength and signal quality numbers were virtually identical. In real life performance the overall ability to pull in a signal seemed the same too, though the S3 is the first Qualcomm-based LTE phone that finally provides a USEFUL number of ďbarsĒ for a given signal.
When I tested the HTC One X I found the RF sensitivity to be about the same as my S II LTE, though it did provide faster LTE speeds when signals were weak. By association we can conclude that the S3 has similar RF sensitivity to the One X.
As far as RF goes, the S3 doesnít raise the bar in the slightest.
Click on this link for a full description of Audio Performance, an how to interpret it.
Overall the sound quality on
incoming calls is very similar to all other Samsung high-end phones Iíve tested
over the last couple of years. Compared to the S II LTE the tonal quality is a
bit harsher, but not by much. Surprisingly however, the earpiece volume is a
couple of ďnotchesĒ lower on the S3 than it is on the S II LTE.
According to information I was able to dig up on the internet the S3 uses a Wolfson DAC, which if true is good news. However, I spent an inordinate amount of time listening to sample music files on both the S II LTE and the S3 (on both headphones and through speakers) to find out if there was a major difference between the two. The S3 still possesses the annoying pops and ticks that you can hear clearly when you connect to an amplifier and play with the phone while audio is not present. As for overall sound quality and background noise, there is very little difference between the two phones.
In the end I had to say that if push came to shove Iíd given the nod to the S3, but there just isnít that much to judge one truly better than the other. As for comparisons to the One X, it was a bit hard to come to an accurate assessment without a One X to compare to directly to the S3, but given that I hadnít really found the One X to sound much better than the S II LTE, Iíll have to call it a 3-way tie. Now bear in mind that the S II LTE may not have the same DAC as the regular S II, but of that Iím not sure.
Multi-media audio on the built-in speaker is not as good on S3 as it is on the S II LTE. Itís not as loud as the S II LTE and itís much tinnier. This wouldnít be all that bad if the S II LTE happened to have a superior speaker, but it doesnít. I suspect the problem is a result of having less space to cram the speaker in, and thus they went with something smaller. The 2100 mAh better is physically larger than the 1850 mAh battery in the S II LTE and the phone is slightly thinner, thus leave less room for everything else.
The S3, like the Galaxy Nexus and the S II LTE HD, has a Super AMOLED screen boasting a resolution of 1280 x 720 (which had now become pretty much standard for high-end phones). The screen measures 4.8 inches diagonally, which makes bigger than all the rest: One X at 4.7 inches, Nexus as 4.65 inches, and the S II LTE (non HD) at 4.5 inches. Because the aspect ratio of a 1280 x 720 screen is different from that of a phone with a resolution of 800 x 480, the screen is the same width as the one on my S II LTE, but itís taller.
The brightness of the screen is almost identical to the Super AMOLED Plus screen on the S II LTE, but once again if I had to pick a winner Iíd go with the S3 for being ever so slightly brighter. The screen has a slightly red-ish tint compared to the slightly blue-ish tint on the S II LTE. The colors are slightly more saturated, but this might not be a good thing given that the AMOLED display on the S II LTE is already consider a bit over-saturated as it is. Iím not sure many users will find it annoying, but it will be obvious if compared to other phones side-by-side.
Many people have complained that the use of the PenTile sub-pixel arrangement is detrimental to the S3ís screen, but I canít see it. Solid colors look solid, text looks smooth as silk, and there are no strange artifacts in any images I looked at. Sure if you look at the sub-pixels under a microscope you can see the difference, but given the incredibly tiny size of the sub-pixels itís a stretch to claim that anyone can detect this with their naked eye. However, if there are people who really can detect a difference, this isnít a slight against them, but rather a warning to everyone else not to take this too seriously.
However, compared to the One X the display on the S3 is a bit of a disappointment. The One X display (specifically the one on the dual-core LTE model) is probably one of the best presently available on any phone, especially in terms of brightness and sharpness. Youíd have thought that Samsung would have tried to at least match the One X.
The S3, like the One X, uses the Adreno 225 graphics processor. Surprisingly however, the One X managed to match the performance of the S II LTEís smaller screen (and Adreno 220 GPU) in terms of frames-per-second. Given the use of the same GPU on a screen with the same resolution as the One X, youíd have expected the S3 to perform equally. However, the S3 fails to reach the benchmarks achieved by the One X, though the difference isnít all that great (47 vs 51 fps for example). In any event, the One X wins out for having a slightly better implementation of the same basic hardware.
Bare in mind that the S II LTE
doesn't need such a powerful GPU because is has 2.4 times fewer pixels to paint
than the S3 or the One X. As such, it isn't difficult for the S II LTE to
achieve relative high frame rates on benchmarks that target graphics
Processor and Chipset
The LTE version of the S3 comes with the Qualcomm MSM8960 dual-core process clocked at 1.5 GHz, which identical to the LTE version of the HTC One X. This particular processor is a Snapdragon S4 chip, whereas the MSM8260 in the S II LTE is a Snapdragon S3. Now Iíve read a number of reviews by perfectly sane people who seem to think that S4 is the greatest thing since sliced bread. I canít say Iíve seen any evidence to back up these claims. There's no question that the S4 is an evolutionary step up from the S3, but there is no VAST difference between the two.
All of the same things I said concerning the One X and the performance of the S4 Snapdragon processor vs the S3 Snapdragon apply here, and so with a few minor modifications, much of the following 4 paragraphs are word-for-word from the One X review:
In various synthetic benchmark tests the S3 doesnít seem to fair any better than the S II LTE. The Antutu scores are fairly close (6609 on my S II LTE and 6819 on the S3). Yes the score is higher than the one I cited for the S II LTE in the One X review, but I reran the test 3 times on my phone and I got similar results each time. Possibly the newer version of Antutu performs the tests differently.
But synthetic benchmarks donít always tell the true story, and so I ran countless apps on both phones to look for smoother or faster operation on the S3 vs the S II LTE. In some situations I MIGHT have seen SLIGHTLY smooth performance from the S3, but I had to stare might hard to see it and the differences were so small that they werenít really significant (same as the slight differences in benchmark numbers).
Now interestingly the S3 runs Ice Cream Sandwich, while my S II LTE only runs Gingerbread. If youíve read any of the fanboy stuff concerning ICS youíd think that it alone would have resulted in a buttery-smooth phone compared to one running Gingerbread. If that were the case I should have seen a huge improvement in the smoothness of the S3 simply because it ran ICS, but I didnít.
The bottom line is that come the end of the day the slight differences in performance between the S3 and other dual-core phones on the market today is insignificant and mostly imaginary. That doesnít mean that the S3 isnít a fantastic performer, but at the same time it canít walk on water and it puts its pants on one leg at a time just like other smartphones (okay, thatís a weird analogy).
As for RAM, the S3 takes the extraordinary step of providing 2 GB instead of the now-standard 1 GB found in most other high-end Android phones (including, interestingly enough, the international version of the S3 with the quad-core processor). At this stage I doubt that this extra memory will make much difference to most owners, but based on what Iíve seen of late (apps consuming large amounts of RAM, such as up to 150 MB in the case of a long run of Flipboard, and up to 100 MB in the case of Draw Something) the extra RAM will probably future-proof the S3 better than its competitors.
One thing I had expected to see on the S3 (given that it uses the same chipset as the HTC One X) was an increased LTE performance over the S II LTE, but I didnít see this at all. The One X has markedly faster upload speeds and can provide superior download speeds when the signals are weak. However, the S3 had performance that is, well, identical to the S II LTE. I donít know how to explain this, but once again the nod goes to the One X for winning in this category.
For those who live in an area covered by LTE however, and you presently donít have an LTE phone, you really should give serious consideration to this technology. While most LTE evangelists tend to harp on the raw transfer rates (with peak speeds over 60 Mbps), this isnít really the big selling point. The real advantage is in the much lower latency (ping times). Latency is especially important when surfing web pages, because the time it takes to request each of the countless small items on a page incurs a delay equal to the latency and this can compound enormously on complex pages. In addition, compared to HSPA, where latency gets increasingly worse as the signal degrades, LTE suffers virtually no loss of its low-latency almost to the point of no signal at all.
So despite the odd discrepancy between the S3 and One X, LTE on the S3 still leaves HSPA eating its dust.
When I first received the S3 I compared the performance of its GPS with that of my S II LTE inside of the house. I ran the GPS Status app on both phones and I found that the S3 showed more activate satellites than the S II LTE, but the reported accuracy was otherwise the same (3 to 4 meters). It looked good, but only a real test would tell for sure.
By real test, I mean that I installed SportsTracker Pro and I took the phone with me on a bike ride, which Iíve done with previous phones as well. I had the route tracked by both phones and I ensured that each had an unobstructed view of the sky. After the ride I compared the results. Along with all of the expected data recorded by tracker apps, SportsTracker Pro also records the reported accuracy of the GPS, which it can then render as a graph of accuracy against time. Both tracks had virtually identical accuracy graphs (mostly 3 or 4 meters, with a few short-lived spikes above that).
To get a feel for how this accuracy translated into the real world, I played back the trips as a ďflyoverĒ on Google Earth. Sadly this is where the S3 came off rather poorly compared to the S II LTE. In the track recorded by my phone I was always shown on the correct location on the road, or within less than a meter of a trail if that was where Iíd been riding at the time. The S3 track suffered from far greater inaccuracy and was often off by up to 5 meters from the true track of the ride. In many instances I was shown riding 2 lanes over from the side of the road. The S3 was also more easily affected by tall buildings than the S II LTE.
But, Iíll be the first admit that just maybe the GPS chip in my phone happens to be exceptionally accurate by fluke and that virtually every phone I test will look poor by comparison. On the other hand, it might be the GPS supplier Samsung used when building the S II LTE has better quality control and their chips are indeed more accurate.
Now if you use GPS primarily with Google Navigation or just to find your location, then this amount of inaccuracy isnít really of any concern to you. However, if you use the GPS as I do for tracking sports activities (and you value accuracy), then the lesser precision of the S3 GPS receiver will be a problem for you.
After much photo-taking and close inspection on my computer I have to conclude that the S3 has the same 8 megapixel camera assembly as the S II LTE. The image quality is more or less identical on daylight photos, but at low light levels (using the night profile) the S II LTE produces slightly less noise, and the noise it does produce is neutral-colored and resembles film grain. The noise on the S3 is predominantly the red/green type that looked very much like digital noise and is less tolerable.
I also found that the camera software on both of these Samsung phones is virtually identical. However, the S3 does benefit from the faster underlying camera firmware in Ice Cream Sandwich and can take pictures markedly quicker than the S II LTE. This is especially true in burst mode, which the S II LTE doesnít even support. I also found that the single-LED flash on the S3 is substantially brighter than on the S II LTE.
Since the S II LTE has an excellent camera, so by extension does the S3. Color accuracy is terrific (though for some reason it isnít quite as good on the S3) and noise is quite low in well-lit shots and fairly tame in low-light shots (if night mode is used). Compared to the One X, the S3 takes better stills, but I didnít really get a chance to test video enough to make a judgment on it.
What of my original question: does the S3 deserve the following it has already generated? Iíd have to say no. Itís certainly a powerful phone, but so is the HTC One X which hasnít generated anywhere near this level of attention (and the One X arguably looks better). What we have here is a true CULT attraction that transcends the hardware. Samsung has finally entered the realm once reserved for the iPhone, thus creating the first true single-device competition to the vaunted Apple product. Itís no wonder Apple is trying to block the sale of the S3 in the United States. But, like the iPhone, the hardware doesnít really deserve this level of adoration. So donít pay any attention to the popularity of this device and instead judge it for what it is, a capable high-end phone that is an incremental evolutionary improvement over its predecessor.
I therefore find it hard to recommend that anyone with an S II (LTE or otherwise) blow a wad of money to upgrade. However, if you presently have a much lower-end phone and are looking to upgrade to something much more powerful, then the S3 is a good choice, but I would advise you to have a look at the HTC One X as well before you take the plunge.