Siglent SDS3000X HD and upgraded SDS1000X HD

[rf-loop]

You tell here that it is zoomed out but I can see it is zoomed in horizontally using a factor of 400.
Least in image where data length is 400kpts (afaik captured using 20us/div) and then you have stopped acquistion and then you have zoomed deep in (zoomed horizontally 20us/div to  50ns/div ).
Also there is vertically zoomed in but no where can see how much because you have not displayed original run time setup with some reason.

What you are trying to say?
Yes it's zoomed out on both direction. It is what is seen when I do zoom to see details of a signal no matter what I do. I don't understand what you are triyng to say.
Software version is 1.1.0.2.

Zoomed IN.
When publish this kind of info it is least good to inform how these have done because many (most) readers can not know it when just looking images.


But good to  know your FW version.

At this time, least  I do not know FW version what is factory installed in these units what are later officially available in western countries. Also availability here can be later than official models information release date in outside China.
ETA: Updade: This Sinc problem (bug) is regognized and solved by Siglent and fix included in later FW.

[skander36]

Zoomed IN.
When publish this kind of info it is least good to inform how these have done because many (most) readers can not know it when just looking images.

The  look of the waveform is what is strange.
The interpolation errors appear only when I zoom in.
But the problem proove that is just a bug.

[tszaboo]

In the case of the 1000X HD series, you can expect a little surprise if you take the original prices of the Rigol DHO1104/1204 as a benchmark.

Good surprise or bad surprise? I think I made up my mind of purchasing either of these scopes. Which one will depend on the prices, and the unlocking abilities.
Few decisions I had to make:
Sacrificing speed, because I'm not going to be able to see high speed with these, but then I just bring the DUT to work to measure it, or buy a 20 year old Lecroy or something to measure those.
AWG: Tried the built in AWG in the SDS1000 series, wasn't impressed, it couldn't even make a 3.3V logic level clock signal that I needed, so external AWG is just much more useful. I can still connect them with USB and make a Bode if both of them is Siglent.
LA: No built in LA, but I tried Sigrok during the Christmas period, and was impressed how useful it was. Most of the decoding I'm doing now are not the usual I2C/SPI/UART signals anyway so the scope's built in decoding options are less useful.
I would be happy if the SDS1000 HD would be in the right price bracket, and it would compete with the Rigol 1000 series. I kinda fear that the SDS800 HD despite of it's name will cost more than most SDS1000 series, but we will see.

[Performa01]

The wiggly, oscillating interpolation clearly looks like a bug. For the blocky lines on a 2*2 pixel grid, one could argue that it's not a bug, since there is not more information available (assuming that the view was zoomed in vertically). But it is a very poor choice of rendering which makes the scope look bad -- especially given that "12 bit, high resolution" is the main claim here.

In fact, I would argue that Siglent should reconsider their "staircase" rendering of traces in general, since it makes the traces look unnecessarily jagged and pixelated. The programmers can do better: The attached close-ups compare the trace baldurn had shared a few posts above (reply #201) with a Bode plot I grabbed from an older 2000X HD thread.

In my view, the Bode plot shows how to "properly" render curved or diagonal lines, and is both clearer (more precise) and visually more pleasing than the way traces are rendered.

Unfortunately, the budget instruments have to make do with the block RAM in the Zynq SOC, so there are no ressources for pixel mapping at that level. This affects the SDS800X HD, SDS1000X HD and even the good old SDS2000X Plus. which has been around since early 2020 and nobody ever complained.

And if you're honest, this is nothing but cosmetics. For math traces (which are inevitable anyway, as soon as you start digging for details in the noise), there is no restriction at all. Also in Run mode nobody will notice it, because of the hundreds to thousands of records mapped to a single frame on the screen.

[Martin72]

For math traces (which are inevitable anyway, as soon as you start digging for details in the noise), there is no restriction at all. Also in Run mode nobody will notice it, because of the hundreds to thousands of records mapped to a single frame on the screen.

If that is the case, I don't find the whole thing so dramatic.

[rf-loop]

But the problem proove that is just a bug.

Yes.
And this bug do not exist, least I can not reproduce it, in SDS800X HD.

(But I know it exist in OLD China only version SDS1000X HD)

ETA: Updade: This Sinc problem (bug) in SDS1000X HD is recognized and solved by Siglent and fix included in later FW.

[2N3055]

For math traces (which are inevitable anyway, as soon as you start digging for details in the noise), there is no restriction at all. Also in Run mode nobody will notice it, because of the hundreds to thousands of records mapped to a single frame on the screen.

If that is the case, I don't find the whole thing so dramatic.

This is something that can only be observed in stopped, single capture. As long as you have scope in run, dithering provides AA... There will always be some noise that will make sure of that...

Like I keep repeating, Keysight MSOX3104T I have here has horrible blocky low res rendering if you start looking at it this way.  But in RUN mode, waveform seems fully analog.. Because of dithering effect.

[skander36]

Like I keep repeating, Keysight MSOX3104T I have here has horrible blocky low res rendering if you start looking at it this way.  But in RUN mode, waveform seems fully analog.. Because of dithering effect.

You miss the ideea ... MSOX3104T is an 8 bit scope. SDS1K HD is 12 bit, but waveform reconstructed on the screen look like an 8 bit scope while his Rigol competitor do a better looking  reconstruction in stop mode. This was the meaning of my first post. Is this dramatic? For me is not, but also not pleasant. Also the LCD has a low contrast versus same competitor.

[2N3055]

Like I keep repeating, Keysight MSOX3104T I have here has horrible blocky low res rendering if you start looking at it this way.  But in RUN mode, waveform seems fully analog.. Because of dithering effect.

You miss the ideea ... MSOX3104T is an 8 bit scope. SDS1K HD is 12 bit, but waveform reconstructed on the screen look like an 8 bit scope while his Rigol competitor do a better looking  reconstruction in stop mode. This was the meaning of my first post. Is this dramatic? For me is not, but also not pleasant. Also the LCD has a low contrast versus same competitor.

No you are missing the point.
You keep repeating something that is not true.. Screen does not look like 8 bit scope.. If it were 8 bit scope, at that magnification trace would be 10mm thick...

Pleasant? Who cares. It is a nitpicking. It is literally not visible in normal use, at least I didn't notice it.

[skander36]

Like I keep repeating, Keysight MSOX3104T I have here has horrible blocky low res rendering if you start looking at it this way.  But in RUN mode, waveform seems fully analog.. Because of dithering effect.

You miss the ideea ... MSOX3104T is an 8 bit scope. SDS1K HD is 12 bit, but waveform reconstructed on the screen look like an 8 bit scope while his Rigol competitor do a better looking  reconstruction in stop mode. This was the meaning of my first post. Is this dramatic? For me is not, but also not pleasant. Also the LCD has a low contrast versus same competitor.

No you are missing the point.
You keep repeating something that is not true.. Screen does not look like 8 bit scope.. If it were 8 bit scope, at that magnification trace would be 10mm thick...

Pleasant? Who cares. It is a nitpicking. It is literally not visible in normal use, at least I didn't notice it.

You can't see because you don't have Rigol DHO1K. This is a comparison between Siglent SDS1K HD and Rigol DHO1K. Not with Keysight, not SDS800 HD not SDS2000-HD. You reach an apex of denying
Go get an read my first post (176). I don't said that is an 8 bit scope but it has an "8 bit" -ish look ...

[Martin72]

P.S. I have done factory reset few times due to a bug with Power Analisys

Do I understand correctly that you have an SDS1000XHD?

No ... I have SDS1204X-HD 

Excellent!
Then it would be great if you could keep this thread alive with your impressions:
https://www.eevblog.com/forum/testgear/siglent-sds1000x-hd-12bit-2gsas-oscilloscope-unboxing-test/msg5261787/#msg5261787

The beta testers have either SDS800X HD and/or 3000X HD at the start, but no 1000X HD, so any new input is very welcome.

[skander36]

Then it would be great if you could keep this thread alive with your impressions:
https://www.eevblog.com/forum/testgear/siglent-sds1000x-hd-12bit-2gsas-oscilloscope-unboxing-test/msg5261787/#msg5261787

The beta testers have either SDS800X HD and/or 3000X HD at the start, but no 1000X HD, so any new input is very welcome.

For what type of coments I have received for extremely simple and pictures documented comparison, right now I'm not very willing ...
but if you have specific questions you can ask. If I'm able I will try to answer.

[skander36]

Could you post a similar comparison of trace appearance on the DHO1000 and SDS1000X-HD in "Run" mode, maybe for the same square wave with the same horizontal and vertical settings? With noise fluctuations and dithering, a trace broadened by one pixel should make very little difference then, I would hope.

Few posts above (https://www.eevblog.com/forum/testgear/siglent-sds3000x-hd-and-upgraded-sds1000x-hd/msg5341559/#msg5341559) are the the screenshots with same settings (1V/div , 50 ns/div), single capture, no zoom. If I corectly understand your request.

[skander36]

Few posts above (https://www.eevblog.com/forum/testgear/siglent-sds3000x-hd-and-upgraded-sds1000x-hd/msg5341559/#msg5341559) are the the screenshots with same settings (1V/div , 50 ns/div), single capture, no zoom. If I corectly understand your request.

Those are in "stopped" mode, which makes the difference between the single-pixel line of the DHO1000 and the double-pixel line & vertical steps of the SDS1000X HD most clearly visible. I was wondering whether one can still see a difference in "run" mode, while the scopes are dynamically overlaying multiple traces. Both scopes can still capture screenshots in that mode, hopefully?

Ok, you are right! I have missed RUN mode.
In run mode the effect may not seem so obvious for everyone. (for me it is). As many said here the benefit that softare capabilities bring make this aesthetic issue to be easily overcome.

[KungFuJosh]

I see the point you're making between the two screenshots, including in RUN mode. Though I'm guessing the bug you found is present still in the Siglent screenshot.

In either case, I think the Rigol line looks blurry and it annoys me. Anti Aliasing is a method of visually lying. It can be tweaked well on a computer for crisper edges, but on the Rigol, it looks blurry. If you want your scope to smooth out your line, and lie to you further, then I find it even harder to trust the scope. 🤷

[skander36]

I see the point you're making between the two screenshots, including in RUN mode. Though I'm guessing the bug you found is present still in the Siglent screenshot.

In either case, I think the Rigol line looks blurry and it annoys me. Anti Aliasing is a method of visually lying. It can be tweaked well on a computer for crisper edges, but on the Rigol, it looks blurry. If you want your scope to smooth out your line, and lie to you further, then I find it even harder to trust the scope. 🤷

It is your opinion and I will respect it.

[KungFuJosh]

Okay, I'll say they're both lying in their own ways. 😉

[DaneLaw]

Is it TFT LCD panel tech used in the SDS1000HD or IPS.?

[Rudgas]

To chime in on the aliasing-discussion, a pixel-perfect representation is the sharpest, with the highest edge-contrast possible, and edge contrast ist mostly what is perceived as an image being sharp. So Siglents choice to go for pixel-perfect fonts makes total sense for me, even at small size, very legible. There are of course PP fonts available for any size. But the tradeoff are jagged edges, especially for small angles.

If the resolution is high enough, anti-aliasing can be used to blend the edges making it even more pleasing visually.
Where it gets hairy though, is when when the resolution is low enough that an anti-aliased font will start to blend the edges and will even cut into the "skeleton" of the characters itself. Resulting in low contrast as most pixels will already be part of the "edge" and feathered out. An offender would be the Rigol 5000, for everyone who has one, a look at the numbers of the LA channels shows that the brightness within the numbers is uneven, depending on where a pixel lands virtually as to where it is rendered on-screen and thus blended. (I'm sure there is a rule of thumb how many pixels at full intensity there should be before edges are blended)

Taking a look at the latest two screenshots, in the siglent screenshot the top row illustrates this nicely as the font is pixel-perfect, put the symbols aren't. The Rigol screenshots seems unfortunately rescaled as not even the grid is rendered Pixel-perfect.


P.s. As a side-note, the wave-form rendering of the siglent seems a bit unfortunate as the pixels are unsymmetrical, making Y2Y slimmer than X2X.

[DaneLaw]

Certainly quite a difference  (From #233)

[skander36]

Is it TFT LCD panel tech used in the SDS1000HD or IPS.?

I realy don't know, but after backlight  brightness I will say that is LCD. But again I have no confirmation.

[tautech]

Is it TFT LCD panel tech used in the SDS1000HD or IPS.?

I realy don't know, but after backlight  brightness I will say that is LCD. But again I have no confirmation.

From the datasheet:
10.1 TFT LCD with capacitive touch screen

[Rudgas]

Certainly quite a difference  (From #233)

To chime in on the aliasing-discussion, a pixel-perfect representation is the sharpest, with the highest edge-contrast possible, and edge contrast ist mostly what is perceived as an image being sharp. So Siglents choice to go for pixel-perfect fonts makes total sense for me, even at small size, very legible. There are of course PP fonts available for any size. But the tradeoff are jagged edges, especially for small angles.

All is well for the font rendering. For the limited resolution of the Siglent screen, I see the benefits of sharp pixels over not-so-convincing anti-aliasing.

It's the rendering of the traces which skander36 demonstrated in various screenshots. You are not supposed to see blocky super-pixels and stepped slopes there. I find that disappointing for a 10" scope with the major selling point of 12-bit resolution. Especially in view of the fact that Siglent is already at a disadvantage vs. Rigol regarding the physical LCD panel, I think they made a bad decision to degrade the trace resolution by another factor of 2.

The Rigol screenshot isn't an accurate representation of what the scope screen actually looks like, I don't have one at hand, but I can't imagine the grid wouldn't be pixel-perfect. As can be seen from the magnification, the grid-dots are often 2-4 pixels in size and quite blurry. Since the original screenshot is 1396 x 872 pixels in size, the image in question was upscaled at some point causing upscaling artifacts. While easily visible on the grid-dots, the trace will also have been altered. And at these small margins, ~9% increase in size, the effect on thin lines is quite destructive and dependent on the upscaling-method. Taking a guess, since everything else has softer edges, the trace will have softer edges as well, as in reality.

[tszaboo]

Is it TFT LCD panel tech used in the SDS1000HD or IPS.?

I realy don't know, but after backlight  brightness I will say that is LCD. But again I have no confirmation.

The white balance will be different on those pictures, because the case color, and different camera.

[DaneLaw]

Certainly quite a difference  (From #233)

To chime in on the aliasing-discussion, a pixel-perfect representation is the sharpest, with the highest edge-contrast possible, and edge contrast ist mostly what is perceived as an image being sharp. So Siglents choice to go for pixel-perfect fonts makes total sense for me, even at small size, very legible. There are of course PP fonts available for any size. But the tradeoff are jagged edges, especially for small angles.

All is well for the font rendering. For the limited resolution of the Siglent screen, I see the benefits of sharp pixels over not-so-convincing anti-aliasing.

It's the rendering of the traces which skander36 demonstrated in various screenshots. You are not supposed to see blocky super-pixels and stepped slopes there. I find that disappointing for a 10" scope with the major selling point of 12-bit resolution. Especially in view of the fact that Siglent is already at a disadvantage vs. Rigol regarding the physical LCD panel, I think they made a bad decision to degrade the trace resolution by another factor of 2.

The Rigol screenshot isn't an accurate representation of what the scope screen actually looks like, I don't have one at hand, but I can't imagine the grid wouldn't be pixel-perfect. As can be seen from the magnification, the grid-dots are often 2-4 pixels in size and quite blurry. Since the original screenshot is 1396 x 872 pixels in size, the image in question was upscaled at some point causing upscaling artifacts. While easily visible on the grid-dots, the trace will also have been altered. And at these small margins, ~9% increase in size, the effect on thin lines is quite destructive and dependent on the upscaling-method. Taking a guess, since everything else has softer edges, the trace will have softer edges as well, as in reality.

Could look like there is some enhancement in play from the blur around the waveform on the Rigol, both screen dumps use PNG, so no EXIF metadata, quite a difference in size 25KB vs 384KB, even though only a minor res bump on the Rigol screen 1280x800 (screen dump 1396x872) vs 1200x600 (1200x600).

I don't have either, the only somewhat bench scope I have with a screen dump (+video) ability is a Micsig, and here the screen dump is native res, but in a JPEG container with EXIF metadata that is possible with the JPEG-std even though its usefulness in a scope is quite limited.
You can work with it in post (fx on the scope's own built-in photo editor) get histograms on color & segment charts, curve the values on fx intensity grading-steps to highlight certain data.  http://tinyurl.com/4zdvkb45
If I just transfer the scr dump over WIFI or to a USB drive or to the cloud by the inbuilt browser - the screen dump are native, and not processed / upscaled.

Skander36@ since your the lucky owner of both, can you make a correlating setup on both units, and take a camera picture..