Lowest Lag HDTV for 480i I've ever seen (Panasonic plasma)

[xeos]

I got a free Plasma TV from 2008 recently and I was super impressed. Out of the box I suspect there's no better TV for retro gaming that isn't a CRT (and finding a CRT that hasn't worn out is getting tough).

What set the Panasonic apart is two things: first, incredibly low lag, roughly 1 extra frame. Second, it's a very fast deinterlacer, and it's adaptive so for static content it looks much better than bob deinterlacing (as compared to the OSSC). Sadly it's only 720p so by modern standards it barely qualifies as HD. But in a world where CRTs are vanishing these are pretty easy to find.

if I've piqued your interest and you want to see if it's worth searching one out I've got an extensive writeup with full measurements:

https://alantechreview.blogspot.com/202 ... g-and.html

Anybody else have suggestions for a more modern HDTV that looks good with 480i and has minimal lag?

[Josh128]

Nice review. I had a circa 2012 TC-p50-X60 for a while, and while the late model Samsungs beat it in PQ for most things, it did indeed have awesome 480i handling which I preferred over the Samsungs 480i handling. Im not sure about the model you are testing, but the x60's 480i actually "flickered" and had line crawl just like a CRT would, which is a good thing, because still images or motion looked the same, no anomalies like you might have with motion on a set that does
standard deinterlacing.

I'd add that for everything else the Samsung F4500 series was superior, except maybe for lag, which I didnt test on the Panny. The Samsungs measured about 37ms of display lag, which is decent, but not as fast as the S60 Panasonics or perhaps the model you have here.

[NewSchoolBoxer]

I really like the work that you're doing. Genius stuff. I learned from my own technical posts that most people just want the answers and not the scientific explanation. In fact, I dislike seeing the push of expensive upscalers without stopping to suggest if they're necessary. I don't play fighting games so 1 frame of input lag isn't noticeable. I do notice and hate lack of scanlines but obviously not a problem outside of 240p. The massive success of NES and SNES Classic tells you that most people don't care about small input lag or scanlines. They aren't going to play on 20" heavy antiques either. So I think the lag research that you're doing is extremely important.

Can you make another, more accessible / understandable chart of the televisions? On your Dell U2410 page, you compare to a digital stopwatch test of perceivable lag with your [real lag] - [1/2 the scanout]. That's great! I'd like to see a chart of perceived lag vs television, with columns of lag in ms, lag / 60 for NTSC frames and lag / 50 for PAL frames? I want to say CRT delay is [small processing delay] + [0 for top of screen, 1/2 for middle and 1 for bottom] frames. Would you add 1/2 frame more for 240p on CRT since it alternates between color and black lines?

Less important but would be cool to have a chart of the video inputs. You say that the plasma has standard inputs, including 2 HDMI, VGA, and 2 yPbPr. Just would be nice to have in one place for all.

From the link:

I measure at the top and bottom of the screen, but as it turns out these produce the same results: the entire frame updates at the same moment, top to bottom. I've seen at least one other plasma display that drew the screen faster than the refresh rate but this is the only model that does the entire thing simultaneously.

Would like to see another plasma television analyzed! Figure out what the average one can do. Just to be sure, they do interlace 240p as 480i so no scanlines?

[thebigcheese]

Nice for 480i, but for 60 Hz content that's actually 2 frames of lag which is not that great by modern standards (an LG OLED in game mode, for example, is one frame regardless of refresh rate, so ~16 ms at 60 Hz or ~8 ms at 120 Hz). Still, could be nice as a dedicated interlaced content TV.

[xeos]

Nice review. I had a circa 2012 TC-p50-X60 for a while, and while the late model Samsungs beat it in PQ for most things, it did indeed have awesome 480i handling which I preferred over the Samsungs 480i handling. Im not sure about the model you are testing, but the x60's 480i actually "flickered" and had line crawl just like a CRT would, which is a good thing, because still images or motion looked the same, no anomalies like you might have with motion on a set that does
standard deinterlacing.

I'd add that for everything else the Samsung F4500 series was superior, except maybe for lag, which I didnt test on the Panny. The Samsungs measured about 37ms of display lag, which is decent, but not as fast as the S60 Panasonics or perhaps the model you have here.

Thanks! I recently acquired a Samsung plasma to test, but I'm behind so it may be a little while. As for flicker, if you look at the raw brightness profile in my review, you'll see the Panny it does indeed flicker, though not exactly like a CRT. Maybe 120hz? It does deinterlacing, however, so motion and static are not identical looking. Motion has some extra blur and loss of vertical resolution (a bit like bob deinterlacing), whereas static switched to "laced", allowing all details thru with zero blur.

As for dot-crawl that's a function of the input (component vs composite). Personally I'm no fan of dot crawl so I use component video input only. You can filter out dot crawl with fancy "comb" filters, but undoubtedly at the loss of some horizontal resolution.

[xeos]

I want to say CRT delay is [small processing delay] + [0 for top of screen, 1/2 for middle and 1 for bottom] frames. Would you add 1/2 frame more for 240p on CRT since it alternates between color and black lines? [...] Just to be sure, they do interlace 240p as 480i so no scanlines?

That's correct, up to the part about 240p. I'm not qualified to discuss 240p; I've only watched a few semitechnical video's and not measured it myself. Part of the issue is while the Pi is perfectly happy to output 240p over HDMI many TVs don't really like that signal so I've stopped trying to test it. The pi does have an option to produce a 640p/240p composite video signal, and the piLagTesterPRO should work in that mode, but since it requires an extra adapter I don't have I've not tested what happens there either.

If you use an OSSC, however, the 480p timing I report is definitely going to be the same when you use 240p inputs.

Can you make another, more accessible / understandable chart of the televisions?

There's a compromise getting to fit in the space allowed by blogger. I'm seriously considering a dynamic table hosted elsewhere and offering the option of other folks submitting their own content to it. The question is whether anybody else would be interested in submitting content besides me. For now filling in a google spreadsheet on my own is a lot easier

Would like to see another plasma television analyzed! Figure out what the average one can do.

I've tested two other plasmas, both discussed on my site.

1) an earlier panasonic which appears to perform the same although I didn't have nearly as good test equipment built when I did that review so I did not discover that the entire screen refreshed simultaneously, and can't be 100% sure that it exactly matched this display in timing.

2) a vizio plasma, which also updates the screen faster than the refresh rate, but has a total delay that's pretty high, at least 54ms.

so not all plasmas are great. The only thing that truly unites them is how heavy they are.

[Josh128]

When I said "flicker" earlier, I didnt mean visible flashing of the screen, I mean the perceived "jitter" that 480i mode presents on a standard def CRT. That is how 480i looked on the X60 Panny, like a CRT. It was not dot crawl due to Y/C crosstalk, as it was being input with an RGB to component video converter. With any CRT, regardless of the source, pixels on diagonal lines will "crawl" in an escalator-like fashion due to the shifting 240 line fields. This is different than Y/C crosstalk dot crawl.

The X60's 480i mode looked like it was actually shifting fields, just like a CRT. Perhaps the model you are testing here doesnt handle 480i the same way as the TC-P50-X60 does.

[xeos]

When I said "flicker" earlier, I didnt mean visible flashing of the screen, I mean the perceived "jitter" that 480i mode presents on a standard def CRT. That is how 480i looked on the X60 Panny, like a CRT. It was not dot crawl due to Y/C crosstalk, as it was being input with an RGB to component video converter. With any CRT, regardless of the source, pixels on diagonal lines will "crawl" in an escalator-like fashion due to the shifting 240 line fields. This is different than Y/C crosstalk dot crawl.

The X60's 480i mode looked like it was actually shifting fields, just like a CRT. Perhaps the model you are testing here doesnt handle 480i the same way as the TC-P50-X60 does.

Ah, I see sorry for the misunderstanding. I've never seen dot crawl used to describe anything other than Y/C crosstalk. I can see how you might get apparent/phi motion in the scenario you describe, though most games avoid the kind of hard 45 degree line you describe because it looks crappy on a low res screen. I think bob deinterlacing would show the effect even stronger. So maybe that's the kind of deinterlacing that the TC-P50-X60 does. My Panasonic does adaptive deinterlacing, with none of the motion you describe. For the time being I'm hanging on to my CRT but I think the Panasonic will be a decent alternative when my crt finally bites the dust.

is it possible your adapter does bob deinterlacing, like an ultra-cheap/simple OSSC?

[Josh128]

Notice I used the phrase "line crawl", not "dot crawl" as dot crawl is definitely the correct term for chroma/luma crosstalk. Perhaps it would have been better to say "shifting fields effect". The CSY2100 clone RGB to YPrPb definitely doesnt apply any kind of scaling or deinterlacing, its RGB to Component video color / sync conversion only, so its definitely the TVs scaler/deinterlacer that was causing the shifting fields effect.

[xeos]

Notice I used the phrase "line crawl", not "dot crawl" as dot crawl is definitely the correct term for chroma/luma crosstalk. Perhaps it would have been better to say "shifting fields effect". The CSY2100 clone RGB to YPrPb definitely doesnt apply any kind of scaling or deinterlacing, its RGB to Component video color / sync conversion only, so its definitely the TVs scaler/deinterlacer that was causing the shifting fields effect.

oops! You sure did.

In the vision science literature the term for that kind of motion is phi motion. But probably "shifting field effect" or bobbing artifact would be better for this audience. If I truly understand you, the effect should actually be strongest for pairs of adjacent horizontal lines, not angled, though.

[xeos]

the effect should actually be strongest for pairs of adjacent horizontal lines, not angled, though.

I checked with my OSSC using bob deinterlacing and the effect really is pronounced for diagonal lines. Gee thanks, never noticed that before, now I can't unsee it ;-(

[Josh128]

You're welcome, lol! The more it looks like a CRT the better IMHO.

[xeos]

I just finished testing another Plasma from the same era, but made by Samsung. It doesn't even come close to the Panasonic:

https://alantechreview.blogspot.com/202 ... t-lag.html

[Josh128]

Yeah the last model Samsungs, the F4500, F5300, and F8500 series were miles ahead of previous models like the one mentioned above.

[NewSchoolBoxer]

That's too bad the Samsung isn't viable for retro gaming. Thanks for continuing the research. I live in country where people mod 20" CRTs for RGB when we have endless early-mid 2000's LCDs and Plasmas with VGA input, if only lag is tolerable.

I went to two Goodwills to check out this television niche since that's what people donate these days. One Plasma, Hitachi 42HDS69, with all analog inputs that exist and HDMI and caps at 1080i. Issue was $150 price when all other 40"+ models were $100. Not trying to bargain down a charity but I big brain asked the cashier why it costs more than LCDs when it has burn in risk. Starts spiel about how TV prices are based on many factors including quality - they have 2 price points!

Maybe they would let me bring in a console to lag test. Got SNES with 240p test suite and PS2 Rock Band that has similar manual lag test in 480i and 480p. Thanks for reading my story.

[xeos]

I would definitely test before buying.

keep in mind also that about half the TVs from that era I tested have variable lag each time you turn them or the console on, which can make lag vary by up to 16ms, turning a good tv into a not so good one depending on your luck: https://alantechreview.blogspot.com/202 ... -dont.html


the pilagtesterpro has a tool to check this automatically or you could just do the on/off test about 4-6 times.