CurtPalme.com Home Theater Forum

[Melifluonze]

Register to remove this ad. It's free!

I was reading through the thread on the new VNBs and I came across a statement that a higher refresh rate at the same progressive resolution caused scanlines to disappear... I can see that happening... but...

Wouldn't the only explanation for this be that at the higher scan rate, you're starting to reach the limits of the frequency response of the video path, and your signal is getting a bit noisy or inaccurate, which results in a less focused or controlled scan line, which ends up widening them and blending them so you no longer see them? That would be a reduction in effective resolution...

My common sense and experience would suggest that running the highest resolvable resolution at the lowest acceptable refresh rate (acceptable to the user...) is where you want to be. I would think you'd want to increase the resolution (not the refresh rate!) until the scan lines *JUST* disappear. Running a higher refresh rate just increases the video bandwidth, uses more power, heats things up, runs things at their limit, with no real benefit, and can result in decreased effective resolution...

That is, unless you have motion estimation on and you want to watch a soap opera instead of a movie.. but 60 Hz is good enough to get that done...

Ok, the soap opera thing is seriously just humor... don't flame.

Ok, maybe you can flame a little, but I want you to answer... I know some of you like that effect... but it makes my old fashioned, trained for 24 fps flicker, that REAL film-based cinema used to have brain hurt... I guess I'm in the "make the home theater feel like the old film cinema, with awesome picture and great sound... but make it FEEL the same as it was...

So, what's the deal? Why crank the refresh up to insanity? Especially if you're not necessarily looking for the "TV Effect"? Doesn't 48 Hz give you a better resolution than like, 72+ Hz? Do you guys really see annoying flicker that wasn't in the original source at anything less than 72 Hz? Does the green phosphor on these non-3D tubes really respond faster than 60 Hz? Is more light being put out because you're drawing more times? (I doubt this, because you're moving the beam faster at the same drive = correspondingly less time and energy transfer per pass = not more brightness)

Wouldn't you want to run at 48/60 Hz (depending on 24 Hz or 30/60 Hz ORIGINAL footage) and just de-focus a touch if you need to close the scanlines, instead of maybe needlessly over-driving the electronics? Or is it all about the fluid smooth of motion estimation that the scalers provide...

Hmmm???

- M

[Melifluonze]

See that.. I'm replying to myself... Don't laugh.

Maybe you're looking to get to 72 Hz to do 3X the 24 fps??? Get rid of flicker, but not the studder/judder of 24 fps?

(Still trying to think out why...)

- M

[jbmeyer13]

[Melifluonze]

Frame drop? Do you mean like false frames, etc. when running 60 Hz?

[gjaky]

Why you don't see the scanlines at higher refresh is caused by the non-linear nature of the phosphor's luma response (see attachment), with higher refresh the phosphor's output isn't decaying that much. A good related question: is the phosphor burn depending from the refresh rate at a given resolution?
As for the flickering: when I had my NEC 6PG xtra in the begining I used with 720P-72Hz refresh, and whle it was fine I easily could notice the flicker of it, later switched to 1080i-96Hz, and I see zero flickering, only the scanlines can flicker sometimes. Still find that the best compromise even on my XG135, now that my eyes are used to the 96Hz refresh I notice the 72Hz's flicker more extensively. On a 9" machine one should play with interlaced resolutions even further: 1080i-120Hz or even 144Hz, those resolutions should hide the line structure of the interlaced picture...

[jbmeyer13]

[ecrabb]

It's not frame drop; nothing is being dropped. The anomaly you're talking about is called judder, and it's the result of displaying the original film frames for alternating, but different lengths of time, which is required if the display is running at 60z.

I made this illustration probably 6 or 8 years ago to show why judder happens with a 60hz refresh.



You can see how some film frames are displayed for two display refreshes, while others are displayed for three display refreshes, which results in an odd "pulsating" motion. As Justin mentioned, it's much more noticeable on scrolling titles, slow motion, or large pans. Often you won't really get a sense for how distracting it can be until you see the same scene displayed with and without judder. Another way to think about it is, "ignorance is bliss."

Cheers,
SC

[Melifluonze]

Thanks guys!!!

I understand the judder. I very much notice it and am aware of it. Sadly, I'm living with it until I figure something out, since I am a casualty of the DVDO 1080p60. Not sure why the DVDO DUO couldn't have output 1080p48... I was thinking of finding an older scaler that could simply scale the refresh up to 48 from DVDO DUO 24 without processing, but then we're talking additional lag...

The phosphor decay is interesting. The theory (or fact...) behind that as a justification for higher refresh must be that even though the beam is moving faster, and imparting less energy per pass, somehow the phosphor is being excited to close to the same level as a lower refresh rate with every pass.... But that would only be true in the case of saturation... Or is it that the same average amount of energy is being imparted to the phosphor over time, just at a more even rate, which brings you up the decay curve, but reduces the peak of the curve? And because of that, the parts of the beam that you normally don't see come into view, like the hazy edges of the beam (which reduce resolving power, I think) because they are being brought up the curve. That's interesting. There's some math there somewhere... I'm going to have to play around tonight... I'm still thinking that actually hurts resolving power, though...

On the phosphor as well... if the refresh rate is increased, the perceptible brightness (or brightness perceived by us humans) should decrease, because the instantaneous brightness of each higher speed redraw is diminished. We perceive instantaneous brightness as being brighter than average brightness with the same energy over time. So, if you increase the refresh rate, don't you perceive a drop in brightness? (I'd say yes, from experience) An equivalent would be high intensity LEDs... They are pulsed at a high frequency to emit more instantaneous light but keep the heat down, versus a continuous on with the same average current. Continuous on would be perceived as dim compared to the higher power flashes. There's dark time with the flashes, but we don't perceive it (unless we move our eyes and the flashes spread out...).

I'm still not convinced that higher refresh is better for the electronics, or for the image. I do understand it is better for perceived smoothness, and I do understand that 3:2 pull-down is bad stuff. I also understand the lower resolution to decrease the video bandwidth at higher refresh rates... but that strikes me as a resolution decrease as well, if I can actually see the scanlines at 1080p48, but not at 72 or 96...

Goes back to if 48Hz doesn't bother you with flicker, it's better for the projector, resolution, and brightness than 72Hz, and that's better than 96Hz... Higher refresh is nice for a solid picture though, like on a computer monitor, so I guess it's a trade. I'd also think that you'd have to see what your projector can actually resolve and then use the resolution where it *just* breaks down (i.e. you can't resolve any more detail in a test pattern), and then set your refresh to your comfort level...

Am I still missing something? I'm going to have to experiment.

And, after poking around, projected film has a 50% dark / 50% light ratio (180 degree shutter, closed when the film moves to the next frame, open when the film is still in the gate and projected), and that 50% dark is actually somewhat important to the "feel" of film... Too bad you can't display at a refresh of 144 Hz, and shut off the video for three frames, then enable it again for the next three, for every 24 Hz frame...

[Melifluonze]

Hey, this seems to be from the same presentation...

If the peaks drop, the perceived luminance will drop as well... You'll gain stability (and reduce flicker), but you'll reduce perceived luminance... Not sure that's good. Again, seems like a personal trade!

- M

[gjaky]

[jbmeyer13]

[Melifluonze]

[Melifluonze]

Ok... so now I'm going to make this crazy statement...

If you increase your frame rate without dropping your resolution, you'll end up wanting to boost your contrast (on the 9500, which boosts your plate potential and accelerates your electrons harder, imparting more power into each phosphor molecule) to keep your brightness up. Guess what that translates to? Less efficient use of the phosphor. You're imparting more energy into it by scanning it faster AND boosting the contrast, but you haven't gained any brightness... you just brought it back to what it would be if you scanned it slower...

So what are you doing? BURNING your phosphor faster, because you are actually raising that peak in the phosphor diagram to compensate for exciting LESS phosphor! Not good, I don't think!

Why is this all bothering me so much? Because when I went to see Nash's blend (Which is truly AWESOME), I asked him what his contrast was at, because it wasn't very bright, even blended... He told me he was running his 9500LC's at a contrast of 80! (the beam current density is fixed, but the voltage is increased = more power per phosphor molecule)

I run my ONE 9500LC at a constrast of 70 on my 120" 16:9 screen and it's brighter than Nash's setup. Granted his screen is larger by a bit, but he's running TWO projectors like mine to cover it. That has bugged me since I left, and now I know why!

So, bottom line, it's rock solid, no flicker, looks great... but I don't know how good it really is for the tubes and I wonder if there's a better trade... And maybe that's what you were referring to, jbmeyer13, reducing the resolution, but increasing the frame rate, actually slows the beam down a bit, and may compensate for the higher frame rate.

VERY interesting!

Of course, Nash has a nearly endless supply of tubes, the Bastige! And he probably doesn't care much if he burns a couple every year... You suck, Nash!

[gjaky]

[Melifluonze]

Hmmmm... I also noticed you brought this up earlier...

My question would be... what makes the phosphor "burn", or wear out? If it's the instantaneous impact of the electron with high power that destroys the phosphor molecule, then you're burning more over time by blasting them harder (like murdering each one with one hit), but blasting less of them...

If its average heat, let's say... then blasting them less hard, less often (so they "live longer"), would allow the phosphor to run cooler (just like the LED example)...

Rahhh... wonder what burns phosphor!!!???

And... if you run a CRT at low intensity, does it last correspondingly longer? Or is there a calculation of the amount of total power absorbed by the phosphor per unit area during it's lifetime that leads to burn?

Anybody know?

And, by the way, I forgot to mention, that you're right! The beam speed has no bearing on the peak of the phosphor (if you didn't read that out of the novel I wrote...)

[gjaky]

I was thinking... the brightness difference between the low and high speed scan could be because the phosphor has an exposure time too, a time until it needs to be excited to get its peak, after that if the input remains the same the output remains at peak. This is in conjunction whith most of the things you wrote earlier, but the LED example. With LED we get more lifetime out driving them with fast peaks than a steady lower current. Until you can reach the peak of the response with the phosphor the higher scan should be less harmful, if the phosphor's exposure time is larger than the beam "speed" (mean it right) you ae over the top.

There is a thread by cmjohnson at the very begining of the forum (barclay66 also brought that thread up a few years ago) that says the phosphor burn is caused by the heat stress.

[redfox001]

My opinion is that it gets brighter with higher refresh.

So higher refresh means bandwidth problems but enables lower contrast settings. Lower contrast mean less blooming. So we need to see what is better. There are two things working against eachother. So it is not clear to me what should look better.

The blooming at high contrast might also be a topic of discussion. Is it the focus coil not handling the beam anymore or is it the transistors driven to hard so they are in a worse bandwidth region? I have heared the opinion that the focus system fails at high contrast from experts. That the video path fails I have heared too. So I have no clue. My feeling is why would a focus coil get worse at high contrast? So I opt for the video amps.

[km987654]

[redfox001]

Not sure where the blooming noise comes from. Might be noise in the video amps or noise in the electron beam focus system?

If your focus driver is not noise free some electrons get accelerated the wrong way. With higher contrast that shows. So I do the focusboard mod with the coils Mike suggested. Actualy they just arrived

Hopefully I will not have to replace the thomson coils themselves as they do have benefits.

[noos@xp37+]

Maybe a little off topic, but perhaps interesting in this high profile conversation - thanks for your comments guys.