It is not like GW Instek is making scopes to not sell them and they have been in the test equipment business for a long time (probably longer than Rigol and Siglent). Also keep in mind that the GDS1054B is the only one in this price bracket with individual channel controls. Discarding it based on 'a feeling' is a mistake. In my experience the service department from GW Instek is responsive and their quality is just fine (lots of GW Instek gear is sold as rebadges by several A brands).
The best thing for you to do is make a list with pros/cons for each model where it comes to functionality and decide on what is important to you. Quality wise they are all the same.
If you're in the U.S., the GDS1054B comes in under $300 with the EEVblog discount. The SDS1000X-U is $100 more than that. Unless one of the limitations of the 1054B mentioned below is a showstopper, the Instek is the clear winner due to price alone.
The Instek has per-channel controls, while the Siglent does not. But this comes at a cost: the Instek is substantially larger in its width. If space is at a premium for you, the Siglent is a better choice, all other things being equal (which, of course, they're not).
The UI on the SDS1000X-U, if it's anything like the 1204X-E that I have, is good, and reasonably responsive. But the UI on the GDS1054B is
blazing fast, smooth as butter. Save for a couple of times when I managed to make the UI crawl (I haven't been able to figure out how to reliably repeat it), it's been stable in that respect. The Instek shows exactly what a UI implemented on top of the Zynq architecture is capable of when done properly. The Instek also separates out the select button from the multifunction knob, while the Siglent uses a press of the multifunction knob for selection. The latter results in errors while selecting items. It would have been better for Siglent to use a detented knob for the multifunction knob, as that would eliminate the problem. Instek's solution is superior even to a detented knob, since it allows for fast scrolling through values while also eliminating the problem of the knob moving on you when you select something.
Both units give you excellent serial decoding capability. The 1054B gives you i2c, SPI, CAN, LIN, UART, and parallel bus decoding (tough to see how useful parallel would be with only 4 channels, though).
The 1054B has a frontend that limits the bandwidth to below 100 MHz, so even though it can be hacked to claim that it's a 100 MHz scope, it isn't really. See
https://www.eevblog.com/forum/testgear/possible-gw-instek-gds-1000b-hack/msg3306636/#msg3306636. The -3 dB point seems to be at around 80 MHz. What I don't know is whether or not the hack results in any changes to the bandwidth characteristics of the scope. I suspect not, but I don't have a proper signal source that I can use to test that.
When hacked, the 1054B has segment capability. It also has event search capability that ties into the FFT and makes it possible to display a table of peaks. But Instek seems to have implemented segmented memory as an afterthought. With the Siglent, it is a first-class property, baked into the design, and is always active. Why do I say that Instek has implemented it as an afterthought? Simple: because you can't have it active while search is active at the same time. Similarly, you can't enable search (and thus see FFT peaks) if segmented memory is active. This means you can't see the frame-by-frame peaks in the FFT. This is a limitation that has no rational justification that I can think of.
Worse, if you are using segments on the Instek, the FFT mechanism will
not show the FFT as it varies by segment. This means that you cannot examine how the spectral content of the waveform changes over time. The Siglent does this correctly, precisely because it implements segments as a first-class always-on mechanism.
Additionally, while the Instek will allow you to see the FFT peaks in a table, the Siglent implements this better: it allows you to see the table overlaid on the waveform and/or the FFT. The Instek's table covers up the entire screen and thus makes it impossible to see the peaks in the FFT graph as you move through the table.
The Siglent implements a "what you see is all you get" approach to acquisition. In particular, the unzoomed screen boundaries define the limits of the capture. This has advantages and disadvantages. It has the advantage of clarity, in that it makes it clear exactly what you're getting for any given capture, and it then is on you to set your timebase appropriately to capture what you need. It has the disadvantage that if your screen size (in terms of time) isn't big enough to show the data you want, you can't just zoom out in order to see the data that you missed. The Instek implements a more traditional capture mechanism whereby the amount that is captured is defined by the memory size relative to the sample rate, and the scope will wait until it fills the buffer before enabling another trigger event. If you want a fast trigger rate, you have to manually downsize the capture buffer to accomplish that. There's an indicator at the top of the display that shows the size of the waveform on the screen relative to the size of the capture, so you can always tell whether you're viewing a subset of an entire capture or the whole thing. Perhaps the easiest way to illustrate the difference is to perform a digital protocol capture where you're decoding the results. With the Siglent, you have to set up your timebase so that the entire capture is on the screen. If the amount of necessary data per capture event is high enough, you won't be able to examine relevant subsets of the decoded data without using zoom mode (which works well enough). With the Instek, you can just set your timebase to directly see what you're interested in and the scope will still capture everything that will fit into memory. But the downside of the latter is that you would have to Just Know that the part of the waveform that represents the start of the digital message is within the buffer range, which may or may not be the case. But: nothing prevents you from setting your timebase in exactly the same way you would the Siglent and using zoom mode on the Instek, either, at least when the amount of data you need to capture is much larger than what you need to display. Finally, the pass/fail mask mechanism works only on the screen, and this gives the Instek's approach a massive advantage, because it means you can define a specific portion of the full capture as what should be tested, while the Siglent utterly lacks that capability (the 2000X+ fixes this by making mask testing in zoom mode possible, but that capability is not present in the 1000X-E and, I presume, the 1000X-U).
Now, the Siglent's approach here
does have an advantage in theory. Suppose that you're capturing a large series of bursts of communications traffic, such that the amount of time per burst is, say, 40 microseconds of time. At 1GS/s, that's 40000 sample points. The Instek will force you to acquire 100000 points in order to capture this at the full sample rate. If the amount of time between bursts is more than 20 microseconds and less than 60 microseconds plus the Instek's trigger re-arm time,
the Instek will miss the subsequent event. But with the Siglent, you can just set your timebase to 5 us/div, so that the screen covers 70 microseconds (the Siglent has 14 horizontal divisions), and as long as the delay between bursts is longer than 30 microseconds plus whatever the Siglent's re-arm time is, it'll capture the entirety of all the bursts. And because segmented memory is always on, it will always store the captures for later review. The Instek has segments, but you'd be limited to segments of 100000 points in this case, and that may cause you to miss the trailing portion of some 50% of the bursts. If you intentionally drop the capture size to 10000 points, the sample rate drops to 200MS/s, and that
does take care of this hypothetical problem, but you take a sample rate hit in order to achieve that. Both scopes have some limitations here. They're just different limitations.
All in all, I'd say the Instek gets the win. But not by as much as you might think. Its main advantage is price and that fast UI. At the current price, the value of the Instek is off the charts.