Seen this? 21.5" and 23" e-IPS are 6bit + A-FRC

Interesting. I didn't know eIPS was made in anything below 8bpc.
 
Makes me wonder, how much does it actually cost to produce an 8 bit panel, compared to an 6 bit panel.

Now I don't have any real insight in how panels are being made but my gut tells me that it would rather have something to do with the manufacturing and that the 6 bit-panels could share some of the steps with TN-panels and thus take advantage of the higher throughput of those, but thats just a wild guess - hopefully someone will set me straight :)
 
Are there any other sources that talk about 6 bit eIPS panels?
 
The NEC EA231WMi is definitely not 6-bit as there are no signs of dithering or banding at all, and the panel module is LM230WF2-SLA1, not LM230WF2-SLC1 like they have listed.

The Dell U2311H, on the other hand, shows evidence of dithering, so that one might be 6-bit. The U2311H also glows orange/purple while the EA231WMi glows white, so that tells me they are different panel models, although based on the same panel. As for the Dell 2209WA, that is true 8-bit as they have indicated.

The Samsung c-PVA monitors also show evidence of dithering, but I wasn't sure if it was 6-bit dithering or 8-bit dithering. It might be 6-bit after all.

They say it's not a massive issue because they saw no adverse effects, but it's a serious issue for Mac users because Mac OS X also enables dithering through the video card, which causes the two dithering patterns to clash and produce significant artifacts. That's how I know for sure which monitors dither and which ones don't.
 
I have only tried to search for one of the datasheets, this yields:
http://www.hy-line.de/fileadmin/hy-line/computer/csv/datasheets/LM230WF2-SLC1.pdf
(used in the U2311H).

It must be a load of bull - the data sheet states:

1) "Gray scale or the brightness of the sub-pixel color is determined with a 8-bit gray scale signal for each dot,
thus, presenting a palette of more than 16,7M colors with A-FRC(Advanced Frame Rate Control)."
I.e. it takes a 8 bit input. It applies A-FRC to "emulate" further colors in excess of 16,7M.

2) Table 7 amongst others tell that the input is from 0-255, i.e. 8 bit for each color, and that each bit is assigned into the range (i.e. not discarded).

EDIT: I don't get it, they even posted a screenshot of that very same quote and they totally missed the point. Ouch. To be completeley sure next time, the whole datasheet should be read, but the introduction of the data sheet is actually pretty clear (for once). :)
 
Last edited:
It might be 8-bit dithering as that document indicates, but whatever the case, the U2311H shows signs of dithering while the EA231WMi doesn't.

Also, NVIDIA and ATI GPUs have different dithering patterns, and the EA231WMi doesn't show artifacts with either one. The U2311H clashes worse with NVIDIA GPUs, while the Samsung c-PVA (and TN) monitors clash worse with ATI GPUs.

I hope the new EA232WMi doesn't have the dithering issue like the U2311H does.
 
Last edited:
Sure, it might be an annoying type of dithering.
I've found no indication that the panel in EA231WMi is LM230WF2-SLC1 except TFTcentral's claim. I've seen much more results claiming it is an LM230WF2-SLA1 - and someone even claimed that TFTCentral once stated LM230WF2-SLA1 was the panel to be found in EA231WMi.

They really should pull that article.
 
:

1) "Gray scale or the brightness of the sub-pixel color is determined with a 8-bit gray scale signal for each dot,
thus, presenting a palette of more than 16,7M colors with A-FRC(Advanced Frame Rate Control)."
I.e. it takes a 8 bit input. It applies A-FRC to "emulate" further colors in excess of 16,7M.
)

if you look at some of LG's TN Film panels in similar sizes they have the same opening statement, but again go on to list AFRC and 6-bit +AFRC in the detail

i notice that they actually updated the article as well with the NEC changes as that was acknowledged as an error which is good :)
 
Last edited:
@ ToastyX: you're right that the NEC uses the LM230WF2-SLA1 in the EA231WMi. We were wrong about that, the Dell however does use the LM230WF2-SLC1, like you already suggested.

@ tk-don: an 8 bit palette yields 16777216 colors. That is “more than 16.7M colors”, however you cannot say that it can display 16.8M colors since it lacks 22784 colors for that. Saying it can “present a palette of less than 16.8M colors with A-FRC” just sounds crap.

The panel manufacturers count any form of FRC as an additional 2 bit color depth per subpixel. The palette you get according to them already includes the FRC.

I have a collection of over 1400 datasheet from 25 different manufacturers, I know how to interpret them.
 
Last edited:
They say it's not a massive issue because they saw no adverse effects, but it's a serious issue for Mac users because Mac OS X also enables dithering through the video card, which causes the two dithering patterns to clash and produce significant artifacts. That's how I know for sure which monitors dither and which ones don't.

Hmm, I wouldn't think that 60 Hz was anywhere near enough to do dithering. The monitor would just barely be able to switch between the colors completely before the next frame is received.

To me it feels like it'd rather be something like conditioning or something, dither on the graphic card just seems utterly ineffective.

And why would they? I thought all 6 bit panels used dithering, so how is dithering through the graphic card supposed to help with anything? To get >8bit precision?
 
It isn't enough, but they do it anyway... FRC only works with a pretty much constant image, because it switches back and forth between two shades close to the shade it means to display with every new frame. That way the display will fool you into seeing the shade in between. Or at least that's the theory.

In practice it doesn't work all that well and a lot of people will see a 30 Hz flicker.

If the image is changing rapidly, it doesn't work at all and can actually make things worse, rather than better. You can get artifacts for instance.
 
Last edited:
Whoisthisreally said:
Do FRC monitors typically dither on all grey shades or a small subset of the 255?
3 out of 4 shades would have to be dithered for 6-bit color to emulate 8-bit color.

brumwald said:
To me it feels like it'd rather be something like conditioning or something, dither on the graphic card just seems utterly ineffective.

And why would they? I thought all 6 bit panels used dithering, so how is dithering through the graphic card supposed to help with anything? To get >8bit precision?
It's definitely dithering through the GPU, also confirmed here: http://tech.groups.yahoo.com/group/psychtoolbox/message/10369

They seem to think the problem only affects NVIDIA GPUs in Mac OS X, but I know for sure that dithering is also enabled for ATI GPUs.

They also have a workaround to turn off dithering on NVIDIA GPUs, which people have used for the Dell U2311H and U2211H to work around the pinstriping problem caused by the two dithering patterns clashing with each other. Unfortunately, I don't know of a workaround for ATI GPUs, which makes the Samsung c-PVA monitors unusable on ATI-based Macs. I don't know how the Dell monitors perform with ATI-based Macs.

As for why they would do that, I'm not sure, but the dithering allows the video card LUT to be calibrated in Mac OS X without significant banding, so that might be one reason. Apple displays and laptops don't have any color controls, so everything has to be done through the video card.
 
3 out of 4 shades would have to be dithered for 6-bit color to emulate 8-bit color.


It's definitely dithering through the GPU, also confirmed here: http://tech.groups.yahoo.com/group/psychtoolbox/message/10369

They seem to think the problem only affects NVIDIA GPUs in Mac OS X, but I know for sure that dithering is also enabled for ATI GPUs.

They also have a workaround to turn off dithering on NVIDIA GPUs, which people have used for the Dell U2311H and U2211H to work around the pinstriping problem caused by the two dithering patterns clashing with each other. Unfortunately, I don't know of a workaround for ATI GPUs, which makes the Samsung c-PVA monitors unusable on ATI-based Macs. I don't know how the Dell monitors perform with ATI-based Macs.

As for why they would do that, I'm not sure, but the dithering allows the video card LUT to be calibrated in Mac OS X without significant banding, so that might be one reason. Apple displays and laptops don't have any color controls, so everything has to be done through the video card.
Whoa! Thanks a lot for this link! It's finally solved dithering on my F2380 in Mac OS X
 
@ tk-don: an 8 bit palette yields 16777216 colors. That is “more than 16.7M colors”, however you cannot say that it can display 16.8M colors since it lacks 22784 colors for that. Saying it can “present a palette of less than 16.8M colors with A-FRC” just sounds crap.
You're indicating that it is a native 262,144 color panel with FRC to achieve more than 16.7M colors (probably 16,777,216).
The problem is that if we assumed it to be an 8 bit panel, the sentence would still make completely sense.
There is still no proof in the wording or anywhere else in the particular datasheet which says that it in fact is 6 bit - only that A-FRC is applied to make it show "more than 16.7M" colors.
But i agree that the phrasing might be written with the intention of covering something up due to this uncertainty, but why rely on this? Why not get in touch with LG and get a clear confirmation?

EDIT: And no, i don't have any U2311H or similar, so i'm not trying to defend anything. I'm sure you know how to read the data sheets you have, but i don't see how this can be anything but assumption (as in: non-confirmed). :)
 
Last edited:
the spec sheet for the LM230WF3-SLB1 even mentions 6-bit+AFRC in the features table as in their picture, so seems pretty clear cut.

http://www.tftcentral.co.uk/images/lg_lm230wf3-slb1.jpg

I've had a look at other IPS panel sheets i could find online and they have "8-bit, 16,777,216 color" in the general features table, and "palette of more than 16,7M(True) colors." instead in their description
 
The Dell U2311H, on the other hand, shows evidence of dithering, so that one might be 6-bit. The U2311H also glows orange/purple while the EA231WMi glows white, so that tells me they are different panel models, although based on the same panel. As for the Dell 2209WA, that is true 8-bit as they have indicated.
We should add that using FRC is nothing new (and is also used in the very few displays with a real 10bit panel (e.g. the LG W2420R); at least when feeding a 8bit signal). All better displays with extensive electronic have used it for many years (as implementation in the scaler) to avoid a loss of tonal values after high bit processing of the signal. LG has started a quite new trend (apart from implementations in TN-panels) in the last years when implementing a FRC stage in the panel itself. However: "Pseudo 10bit" or 8bit panel with previous FRC implementation makes no difference for the user (while it has some impact on the electronic): Good FRC only leads to (from my point of view) minor drawbacks without annoying spatial artifacts at least for the 10bit(12bit)/8bit conversion while feeding a 8bit signal. You will see some sort of temporal noise when looking very close to the display.

Regarding the mentioned 6bit IPS and VA panels with internal FRC: While testing the Eizo FS2331 I couldn't make out artifacts that extended the typical Eizo implementation while the full range of tonal values (regarding the input signal) was conserved in the output (that means no banding with alterations via OSD settings). So there was a 10bit/8bit conversion after processing of the signal. Thinking of another panelinternal stage for 8bit/6bit conversion is quite problematic but I will contact Eizo in this case. Maybe they provide some information about their actual implementation in their C-PVA displays.

Best regards

Denis
 
Last edited:
Well Denis, what I am actually observing on my EV2333W is dithering effects on a small subset of grays in test images. The LTM230HP01 panel is stated to be 6-bit + Hi-FRC by Samsung and yet Eizo obviously quotes the monitor as being 16.7M of 1.06B colours, typical of a 10-bit LUT implementation.
 
I've actually pulled out an old TN junker. The dithering on both monitors are comparable which leads me to believe the output stage is 6-bit + FRC. I am looking forward to Eizo's reply if you do decide to contact them.
 
Huh - well I guess this explains why a see very very faint bands when I pull up a gray ramp on my Dell U2211H. I always assumed it was because I didn't optimize the RGB gains and the contrast/brightness.

My NEC20WMGX2 has no bands as long as I keep the RGB gains at 255,255,255 (Native preset) and contrast and brightness at very specific values.
 
I had no idea some e-IPS panels were 6-bit. I've recommended e-IPS monitors to people, I feel duped.

According to this the panel in my EA231WMi is thought to be true 8-bit and I have not seen any color banding or dithering. Looks like I dodged that bullet.

Edit: Woops, sorry for the bump of a two month old thread.
 
14 month old thread. ;)

I have an 8 bit panel and a 6 bit + frc panel and the difference is effectively 0 for everything I use it for.
 
I had a dell S2340L ips, i was very happy with the colours, and saturantion, everything went great until i played games that could not keep the 60 fps and when it touched 24-30 fps i did get very masive ghosting, it's like the caracter almost dubled again and the enviroment, this could be because the FRC?
 
I had a dell S2340L ips, i was very happy with the colours, and saturantion, everything went great until i played games that could not keep the 60 fps and when it touched 24-30 fps i did get very masive ghosting, it's like the caracter almost dubled again and the enviroment, this could be because the FRC?

no, it would be due to the rtc overshoot that all of these dell s-series displays have.
 
But then i ordered the S2440L and i didn't notice it anymore but i wasn't pleased with the colours. So how colud this be possible? I know that the s2440L is AMVA but the diference between these two monitors are just 1 ms and the s2340L is ips. So what's the explanation?
 
Last edited:
there could be a panel lottery or firmware differences going on, as some reviewers have units with pwm and others don't. so it's possible that the overdrive implementations among different units may also vary a little. also, the motion artifacts may not occur to the same degree on the same color combinations between different panel types (a-mva vs. ips).
 
Back
Top