LCDs and 10bit (30bit) output

[chrisdent]

Just wondering if current LCDs have any problems with the extra 2 bits of colour information per channel provided by the 9700 or Parhelia.

 

[Entropy]

Just wondering if current LCDs have any problems with the extra 2 bits of colour information per channel provided by the 9700 or Parhelia.

Other than not being able to utilize the extra colour resolution, no.
Do you want the gritty details?

Entropy

 

[RussSchultz]

Most LCDs can't completely resolve 8:8:8, much less 10:10:10.

 

[chrisdent]

Most LCDs can't completely resolve 8:8:8, much less 10:10:10.

That's what I suspected

 

[RussSchultz]

If you're really worried about color reproduction (i.e. graphic artist using photoshop), you need a high quality CRT.

 

[Dio]

Not sure about that. They have different characteristics, neither is very precisely linear. CRT's and LCD's can both be properly Pantone calibrated, etc.

As far as I can see, low-end CRT's are better than low-end TFT's in this respect, but at the high end it becomes much harder to call.

LCD has the advantage of a full-digital signal path (OK, one of the Iiyama CRT's does it). Recently I've been seeing more and more adverts targetting design and repro houses with the really big TFT panels...

Hard to say I guess. I'm not an artist.

 

[Mariner]

Although older generations of LCDs had problems with colour reproduction and so weren't suitable for colour correction work, I'm pretty sure that most modern LCDs have no such problems.

There are quite a few high end LCDs specifically aimed at repro work (Apple Cinema displays, for example). In fact, after having a very quick look around, I see that Eizo are advertising one of their TFTs as being better than CRTs:

http://www.eizo.co.uk/common/eizo/products/lcd/L685EX/frame.asp

I've got an Eizo TFT at home and the colour rendition is very good indeed.

 

[RussSchultz]

Perhaps they've solved the problem with desktop displays, but projectors have "fixed panel noise", as every pixel does not quite have the same response. This shows up in pans as looking like dirt on the lense (or something like that). It is theoretically calibratable, but with 1M+ pixels, that'd be a pain in the ass.

 

[Dio]

I haven't noticed any big problem on the LG TFT I have here, and that's a cheap one (well, for an 18" it is ). Overall, consistency is somewhat worse than the Iiyama VM Pro 510/512, but given that it's got a huge 0.28 pitch it's better than I expected. And unlike the VM it doesn't induce a hernia whenever it gets moved.

 

[Alistair]

As you seem to be well informed... Im thinking of picking up a Hitachi 17" with a 16ms refresh time. This model appears to only support 16 bit colour however (262k colours). Does this mean Ill see the same artefacts that 16 bit colour generally produces on a CRT (eg the Quake III sky banding) or are LCDs just, uh, different?

Apologies for my ignorance. Trying to make a quick Christmas purchase decision

A.

 

[OpenGL guy]

As you seem to be well informed... Im thinking of picking up a Hitachi 17" with a 16ms refresh time. This model appears to only support 16 bit colour however (262k colours). Does this mean Ill see the same artefacts that 16 bit colour generally produces on a CRT (eg the Quake III sky banding) or are LCDs just, uh, different?

There shouldn't be any problems because the video card will still do all of its rendering at 32-bit. The fact that the display can only show 262k colors won't matter much.

Another example: If I render everything at 32-bit and then copy it to a 16-bit buffer, it will look much better than if I rendered everything at 16-bit.

 

[Nite_Hawk]

Hi Alistair,

The reason why games like quake experience such terrible banding in 16bit colour is because of error propogation when doing multitexturing. Basically, the more layers that need to be combined, the worse the errors will be on the final output. Cards like the kyro II internally render everything at 32bit, and dither to 16bit for the framebuffer. These cards look much better when running in 16bit, than say a TNT2 or a voodooII would look.

As for the screen, it's probably not going to be terribly noticable in most situations. Certain examples such as looking at skin tones (otherwise known at pr0n) might look slightly off. For the most part, newer LCDs such as the hitachi and others should be very comparable to CRTs, with CRTs still possibly getting a slight edge.

If it makes you feel any better, I've recently purchased the samsung 191t LCD and am fairly happy with the picture. Contrast and brightness are both good. Response time is rather low on transitions to or from darker shades, but otherwise it's a great screen. The hitachi looks to be very good as well, and the response time seems to be much better, so I'd say go for it.

edit: one of the sentences didn't really make sense.

Nite_Hawk

 

[Alistair]

Ah - exactly what I wanted to know, thanks. I didn t know about the difference between internal error propagation due to limited precision, as distinct from the depth of the final display. Many thanks.

Think I ve found my Christmas present

A.

 

[Simon F]

Not sure about that. They have different characteristics, neither is very precisely linear. CRT's and LCD's can both be properly Pantone calibrated, etc.

But surely you don't want linear. A nice smooth x^gamma curve (with perhaps linear behaviour at the black end) to match the eye's behaviour must be preferable.

 

[Reverend]

Not exactly an answer to your question but here's what Carmack said to critics about 64-bit color vis-a-vis current monitors having only 32 bits of resolution :

While a monitor may only display this, say, 10-bit resolution on there, the human eye is capable of perceiving well over a range of 64,000 [colors]... like the difference between these lights that are shining right in my eyes here, and the floor... sitting down there, between the aisles. That's a difference of hundreds of thousands of levels.

The way you should be calculating all graphics... the way it ought to be done is: you're basically counting out photons that are, you know, imprinted on a surface. Lights spray out a whole lot of photons, that are collected on surfaces. If you're doing things really right, you have an inverse square falloff, and you have a radiosity map, and all this... So what we want to do is do all of this calculation the right way, and then... we know at the end that it's going to be going on to some... not exactly optimal solution that we want on the monitors there. But there's still a lot of benefit to gain by doing all of the intermediate calculations the way you really should do it.