AVS Forum banner
Status
Not open for further replies.
1 - 20 of 30 Posts

·
Registered
Joined
·
418 Posts
Discussion Starter · #1 ·
I posted last week my experience with the Panasonic RP56 DVD player and "ghosting" of the image when running it with my Sharp 9000 DLP. I also tried a RP 91 and yesterday a Pioneer 47A and had similar results. All 3 of the newer players exhibit this "ghosting" when connected to my Sharp. I do not see this with my Pioneer DV-05. I have screen shots to show everyone what I'm seeing.


Again, I've used long and short cables, component, s-video and composite cables. I've used monster cable and AR cables and an old generic cable. I've tried all (and yes I mean all) the user adjustments on the respective DVD players to see if I can "tune out" this effect. All the players were calibrated equally in terms of brightness and contrast with Video Essentials. BTW I was unable to tune out the ghosting effect even with extreme brightness and contrast levels.


This is my 1st time at trying to post pics, so bear with me.


First, a picture of the Panasonic RP 56 and a video essentials menu


Corey J
 

·
Registered
Joined
·
3,667 Posts
Sometimes mild "ghosting" is observed that is really an MPEG artifact. Typically this occurs from highly compressed satellite channels. Another potential source is edge enhancement. Ideally edges should be critically damped but it commonly gets over-enhanced.
 

·
Registered
Joined
·
418 Posts
Discussion Starter · #5 ·
By the way. I also went to my dealer and saw the same ghosting on their Pioneer 47A hooked up to a Sharp 9000. Therefore I've seen this effect on 3 different DLP's (my Sharp, a store Sharp and a store Yamaha), and several DVD players with different cable setups. Is this something that's on the VE disc? Is this a lens artifact? Is this an inherent problem with DLP's? Inquring minds want to know.


Corey J.
 

·
Registered
Joined
·
418 Posts
Discussion Starter · #6 ·
Bump


Has anyone else observed this effect. Check out the photos on the yahoo photo site listed above.


Is this ringing? Ghosting? Band ripple effect?


Stacy Spears where are you and what do you think of this effect.


Corey Joekel
 

·
Registered
Joined
·
1,005 Posts
That's ringing, and it's almost certainly a symptom of the pre-filtering done by the projector to reduce aliasing when it samples the input signal. I don't think it has anything to do with the DV-05 being a better player than the others, it just has to do with a progressive signal being twice the bandwidth and thus nearer the limit of the filters in the projector.


It's important to note that you are seeing the ringing on menus, which are unfiltered raw pixels, and generate very high frequencies at the edges. Typically menus and subtitles will have the largest ringing artifacts because they aren't filtered, whereas movies typically don't have much, if any, high frequency content. The worst-case edge transitions on a DVD are on the AVIA sharpness pattern (and some of the other Avia patterns), because they are specifically designed to generate the highest frequencies possible and aggravate any ringing. Even so, the menus will typically have worse ringing than even those patterns.


Frankly, it looks to me like the Sharp just doesn't handle the highest frequencies of a progressive scan signal very well. If so, there's not much you can do about that, it's an artifact of the A/D conversion than happens on the way in.


On the plus side, ringing on the menus does not necessarily mean that you'll see the same thing in movies. Compare the results with Avia, and then with the resolution pattern in VE. My guess is that you won't see much, if any, ringing in the VE resolution pattern, because it's prefiltered (I believe to broadcast specs, which doesn't allow interlaced content about 5 MHz).


In short, assuming all of the above is true, any progressive-scan DVD player that didn't show this ringing on this projector would have to be rolled off enormously at the highest frequencies.


Don
 

·
Registered
Joined
·
418 Posts
Discussion Starter · #8 ·
Thanks Don for your reply. I have a couple of questions though.


What happens when you switch the DVD players from progressive to interlaced? When I do this I still see the "ringing" on the menu's.


Is this "ringing" something that could be adjusted or tuned out in the Sharp DLP unit?


Could this effect be minimized by using the SVGA input on the sharp projector (ie using a Faroudja NRS) thereby bypassing the internal scaler?


You are correct that I don't see much if any ringing on the VE resolution pattern. I don't have Avia, but will try to get a copy to compare. I don't see this artifact on movies, however I did seem to notice it a little more when using the progressive machines.


Corey J.
 

·
Registered
Joined
·
1,005 Posts
That's very strange, that you get the same ringing with these other players, even in interlaced mode. My guess would be that the DV-05 is very rolled-off at the high end. The measurements Stacey made for Secrets of Home Theater and Hi Fi only go out to 5 MHz, so this is only conjecture.


The usual way you reduce ringing artifacts is with the sharpness control. You adjust the projector's sharpness, and if that's inadequate, you adjust the player's sharpness control, if any. The sharpness control is essentially a gain control for the high frequencies, like a very special-purpose graphic equalizer with only one slider for the highest range.


And yes, you would probably reduce or eliminate these artifacts if you used a good scaler or HTPC to feed a native-resolution signal to the projector. But first off, get Avia, so you're not testing with menu subpictures; menus don't represent a real-world case (though keep in mind that Avia is just barely a real-world case). Use the sharpness pattern to look at ringing. You may even want to see if there's a factory sharpness adjustment on the projector you can play with.


Best,

Don
 

·
Registered
Joined
·
418 Posts
Discussion Starter · #12 ·
Thanks so much to Don and Stacy for your input. This forum is so cool and so much fun.


Don, in response to your comments about sharpness controls. I have the Pioneer DV-05 set at minimum and the Sharp DLP at -15. The sharp setting is similar to what others have reported also using VE. When I start to crank up the sharpness (on either unit) I see the classic bright white enhancement around the edge of the resolution patterns. This is different, however, from the ghosting that is seen on the menu patterns noted above. Adjusting the sharpness control seems to have no effect on the ghost images.


I've ordered an Avia Disc. Where do I get the S&V disc?


Corey J
 

·
Registered
Joined
·
1,005 Posts
Ringing comes in several related forms. The classic white outline you mention is a mild form of ringing where you basically only see the first-order effects as an overshoot or undershoot before and after strong voltage transitions.


The ringing you are seeing on the menus is sometimes called "band elimination ripple" which is again a form of ringing, but I admit that the shots you took look especially bad. You are seeing the first, second, and third ripples, which is not good. However, given that none of these players have anything like that level of rippling on the scope or on other displays, I'm inclined to blame the projector.


As to why the sharpness control on the player isn't helping, there could be several possibilities:


- The sharpness control may not affect the menu subpictures, but only affect the DVD picture (unlikely, but possible, depending on how and where the sharpness is applied in the signal chain)


- It may not be affecting low enough frequencies. If the nasty ringing occurs above 5 MHz, for example, but the sharpness control is only attenuating the band from 6-6.75 MHz, there's still enough high frequency info in the signal to be problematic.


- It may not attenuate the signal enough. Often the sharpness controls on DVD players are pretty coarse, and are oriented around boosting the high frequencies, not cutting them. People tend to want more sharpness, not less. There are players where their 0 sharpness level is a bit too boosted, and it just gets worse from there.


As to why the sharpness control on the projector isn't working, most likely it happens in the signal chain after the A/D converter has already inserted the ripples, so there's not a lot you can do. If the input signal has too much high frequency information, you're going to get ripples.


Basically, it looks to me like to avoid these ripples, you need a severely rolled-off high end, and there aren't many players these days that do that. Again, this is just my educated guess based on your pictures, your descriptions of the problem and what you've done with your players, and my experience with the players in question. Without actually putting the Sharp through its paces and sending in a lot of test signals, I can't be totally sure.


Best,

Don
 

·
Registered
Joined
·
271 Posts
Great thread. I'll join the fray with another angle.


A previous Sharpvision set, the 34" 16:9 direct view CRT, had exactly the artifact Corey is seeing, also unaffected by Sharpness and SVM settings. It is so pronounced on the 480p YPbPr input as to make that configuration useless - this appears to be due to an undesired and unavoidable

480p->480p or 480p->1080i conversion (the "DDFC" circuit). Even the rather soft VE crosshatch patterns tell the tale; the Avia crosshatches more so. The RGBHV input bypasses the internal scaler and virtually all other video "features", minimizing but not eliminating the problem. Gary Merson had mentioned it in a WR review and on this forum. I also interacted with the exasperating Sharp HDTV service team in Illinois about this. They as much as admitted it was a design issue.


I reckon the 9000's video processing signal chain is very much derivative of that of the earlier circuit and the evil remains. Greg Rogers mentions it briefly in his WR review of the 9000. Basically Sharp still doesn't have their act together with respect to a clean video signal path for wideband signals, SD let alone HD. You begin to understand why so many have turned to driving the 9000's Input 5 at 720p and bypassing the internal scaler.
 

·
Registered
Joined
·
1,031 Posts
Hello -


i wouldn't really refer to this "ghosting" as "ringing". It is actually due to oversampling digital filters somewhere in the playback chain, either in the DVD player, the projector, or both. You can tell this because the "ghosting" appears to both the left (before the transition) and to the right (after the transition). "Ringing" due to any analog phenomenon (such as a peak in the frequency response) will only occur to the right (after) a transition.


This "ghosting" will *only* occur with computer-generated graphics that have not been properly low-pass filtered when the disc was made. It will not show up with any real-world source material. This is because the scanning device in the telecine machine will have low-pass filtering that limits the bandwidth to that appropriate for the format. In the case of DVD, this will be in the range of 5.5 MHz or so, at the most.


If a player (such as the DV-05) doesn't exhibit "ghosting", it is because it either doesn't employ a digital oversampling filter, or because it has very poor frequency response that attenuates the "ghosting" before it reaches the projector. In this case, I am fairly confident that the former is the actual situation.


In any event, this is nothing to worry about whatsoever. It is not due to "ringing" in the video circuitry. It will never show up with anything besides test patterns on AVIA or computer-generated menus.


Hope this helps,

Charles Hansen

Ayre Acoustics, Inc.
 

·
Registered
Joined
·
271 Posts
Very helpful Charles. If I understand your ghosting/ringing distinction, it would be the former on the older Sharpvision I mentioned - both edges of any luminance transition suffer the ghosting. But the problem is so bad that with 480p YPbPr input that it is abundantly evident on regular DVD material, imparting a pervasive murkiness and blurriness throughout the image, not just in the vicinity of very fast luminance transitions in the 5MHz+ range. The Avia 4.2 patch is grayed out as well as the 5.5, and even the upper patches are murky. Basically lots of middle and upper frequencies getting converted to fixed pattern noise - bad news, a flawed design. Ship it anyway!
 

·
Registered
Joined
·
1,031 Posts
Hello Den -


My posting was actually referring to Corey's original posting regarding the Sharp 9000. My observations were based on the pictures he posted. Sorry I didn't make that more clear, as I can see now that you might have thought I was replying to your posting.


Frankly, I'm a bit stumped regarding your posting. If there are "ripples" on both sides of the transition, this must be due to a digital oversampling filter. With the DVD format, these ripples will occur at about 1/2 the original sampling frequency. This puts them at about 6.75 MHz, which is why this behavior won't ever be excited by normal source material, which should be low-pass filtered. It doesn't make sense to me that they would be using a digital filter with a sharp cutoff at some much lower frequency.


If you could post some pictures of what you are seeing, I might be able to figure something out. Or maybe not, but at least I'll try... ;)


Best regards,

Charles Hansen

Ayre Acoustics, Inc.
 

·
Registered
Joined
·
271 Posts
Not to worry, I'm with you through and through. My problem was excited by multiple 480p source material not just DVD but also scaled NTSC sources. That's the point: the TV's video processing was whacked in that topology. An Ayre product or a broadcast piece won't break the rules like this, but you should see the loose ends we find in some of these consumer displays! Never a dull moment for a calibrationist with a sense of humour.
 

·
Registered
Joined
·
418 Posts
Discussion Starter · #19 ·
Den, Charles, Stacy, Don:


Thanks again for your interesting and helpful posts. I hope you all don't think I just sit around all day and watch submenu's on VE!! I really do enjoy DVD movies on my Sharp 9000, I'm just trying to wring as much performance as possible from system and minimize artifacts. I'm also getting quite an education from this forum too!


I don't know if I mentioned in an earlier post, but I did observe this same "ghosting" with a Yamaha DXP1 DLP player with a Pioneer DV 47A or a Yamaha 5 discs carousel DVD player - playing the same VE menu's. And, I think Stacy mentioned he's seen this effect with the Marantz DLP as well - so this effect isn't entirely unique to the "Sharp" line. Also, I can see the ghosting with component, S-video or composite inputs. I've not tried the VGA input yet, but would like to try it with a computer or a Faroudja NRS. Are there any DVD players that output a VGA signal - that might be interesting to try as well.


I have ordered the Avia and S&V calibration discs (from Amazon.com) and will report and maybe post some screen shots when I get them. I also have an old 1st generation Toshiba DVD player - am curious to try it with the Sharp 9000 too. This is going to be interesting and fun!



Thanks, Corey J
 

·
Registered
Joined
·
1,005 Posts
Quote:
Originally posted by Charles Hansen
i wouldn't really refer to this "ghosting" as "ringing".
Hmmm... I'm going to have to disagree with you here. My best understanding, supported by my handy filter reference, is that any overshoot on amplitude changes that decays over time is ringing. My characterization of it as "ripple" is a little fuzzy, as ripple is usually used to describe wavy variations in the frequency domain, not time. But it most certainly is ringing.

Quote:
It is actually due to oversampling digital filters somewhere in the playback chain...
There are oversampling D/A converters, but no oversampling filters, to the best of my knowledge. I don't know what an "oversampling filter" would be. If a device uses an oversampling D/A converter, it can use a filter with a less-steep frequency attenuation response, which will result in less ringing.


In any event, given that the ringing we see in the images is consistent across many DVD players, and Stacey and I have never seen ringing of that level on any other display or on a scope, I think it's fair to put the blame on filters in the projector. Given that the same issue has been seen on multiple DLP devices, it would not surprise me if the problem is in some filter on the reference design that TI provides, or perhaps in a common video ADC chip. This is all conjecture, of course. A lot of analysis would have to be done to definitively track down the culprit.

Quote:
"Ringing" due to any analog phenomenon (such as a peak in the frequency response) will only occur to the right (after) a transition.
Many filters have more ringing on the right, but that is by no means assured. Here's a link to a bunch of impulse-response plots of various filters, and you can see that several of them have preshoot as well as overshoot. The Bessel and Gaussian 12 dB are good examples.

http://www.imcsd.com/filter_technology.htm#performance

Quote:
This is because the scanning device in the telecine machine will have low-pass filtering that limits the bandwidth to that appropriate for the format. In the case of DVD, this will be in the range of 5.5 MHz or so, at the most.
5 MHz is a broadcast limit, not a DVD limit. As far as I can tell, the reason you seldom see high-frequency material on a DVD is that the telecines for movies are done for multiple purposes, including broadcast. I can't get a straight answer on this, and I'm still investigating. It's very possible, even likely, that more DVDs will appear with information all the way up to 6.75 MHz.


I think we're in agreement on several things: the effect is caused by a combination of a lowpass filter somewhere in the display chain and the unfiltered nature of the computer graphics generated internally by the DVD player, and it's unlikely to affect real-world DVD watching. Though I'd caution that it's not entirely clear that DVDs will always be filtered in this way. I wonder, in fact, if Superbit DVDs are filtered. I'd like to see a frequency plot of one. Hmmm...


Best,

Don
 
1 - 20 of 30 Posts
Status
Not open for further replies.
Top