Signal Processing on Sony HDTVs - Page 2  

I had the bar fix done a couple of weeks ago. They didn't even know about the HDPT fix only the board replacement. When I told them to try HDPT it was not in video 1. It was 6 or 7 I believe. Could one of you guys describe to a layman what the heck these guys did to my set? Also they did a bunch of other service menu adjustments but not the HDPT until I told them to try.
HAs it screwed up something doing the fix incorrectly. I've noticed some PQ differences but had not thought much about it until reading this thread. Thanks ,

Adu,
Can you explain what adjustments you made to what? I have been away from the forums for awhile now and am missing out on a lot! My 36XBR800 is in the shop right now (getting a new tube unfortunately) but I would like to know what settings you are using and how they will effect my set. Thanks a lot.

Thanks for the info ADU. I think I see what you are talking about. The 1 setting seems to be a little "crisper " or something like that, I can't quite put my finger on it. Like the seams I saw in a pair of stonewashed jeans on Leno the set seemed to "grab ahold" of the small details. 0 seems to be a little softer and smoother. I think I would prefer 1 except when I do that i start looking for the sweeping bar and I've had enough of that already. One other question, I don't know much about this stuff at all and from what I've read I'm confused about the true ability of this set to show "true 1080i". Exactly what is this set or group of sets (xbr's) capable of displaying vs. what Sony advertises. I can't help but feel a little ripped off after all this. After what you have learned about this stuff , do you think this set is the set it was supposed to be when I bought it (march)? Guess that's two questions, thanks again

ADU,

I don't know where exactly DPSW causes a break in the video path. But I do know that the on-screen display is injected during the final CRT drive processing. The injection is in the form of the 3 video and 2 sync signals that are merged with the selected input within the same CRT drive chip on the B board.

The 2 sync signals (hor. & vert.) are generated upstream by the main cpu which already knows something about the frequency that the CRT will be scanning at (either for the internally generated 1080.0001i or for the external 1080i/540p going through the bypass circuit). I still doubt that a 480p input goes directly to the screen without being up-converted by DRC/MID. Whenever you see 480p and 1080i in twin view, it means at least at that moment, they are going through DRC/MID and not the bypass circuit.

Sorry about me confusing the RF input with video 1.

And a final reminder to all who change the HDPT value:

On the HS500/XBR800 direct view chassis, the change is done after the RF input has been selected (you at least have a cable or rabbit ears connected to the RF input so that there is some signal for the set to lock onto).
On the XBR2 direct view chassis, the HDPT value is changed after having selected a components input that is supplying a 1080i signal. DPSW is also changed at the same time as HDPT.

We are activating a bypass circuit ( like CD direct on some audio receivers). Sony has provided this circuit on all but as few Hi-Scan chassis but never intended to make this feature ever available, or at least immediately available, to the North American market. They probably included the bypass circuit to be used in other countries where for technical or marketing reasons the use of this circuit is desirable.

Sony intended all video signals to go through the DRC/MID circuits. Through a design oversight, they failed to adequately shield the input signals from the high voltage CRT drive output signal. This feedback caused the creation of the infamous scrolling vertical bar in 1080i. Sony's quick solution was to activate the bypass circuit which fixes the bar problem but also provides a PQ improvement for 1080i/540p inputs. One should not expect 2002 HS/XBR chassis to rival the 2003 XBR910 once the formers have been patched. But the gap between the generations has been slightly narrowed and this has accidentally extended the life expectancy of 2002 models.

Let's show our appreciation of Sony by all being grateful for this unplanned gift.

This all sounds really cool but i haven't got a clue what's your saying. I have the 34XBR800 and the most tweaking i've done is the in the servce menu and i also tweaked the picture settings using the Sound & Vision home theater guide DVD.

Should a novice(like me) even attempt what you guys are doing? I'm afraid i'd screw up my beautiful XBR? I use my XBR800 for watching TV shows and DVD movies. I want to get a DVI player but i think i'm going to wait for better build quality.

Oh, Question, when i get a DVI player will it bypass that extra D/A A/D conversion your talking about?

Thanks(in advance) for dumbing down the this thread.

ADU,

now that you tweaked your TV is watching your Samsung 931 look better?

just curious.

eye_candy,

The answer is no. Using the video7 input (DVI) will not allow you to bypass the extra A/D, DRC,MID, and D/A circuits. But instead of having the player convert the numbers on the disk to an analog components signal, it will be the DVI decoding chip in the TV that will convert the numbers to analog using a 10 bit D/A converter. Also there will be no loss of signal strength if it travels a long distance of say 30 feet from the player to the set.

However, should you activate the bypass circuit, then your DVI decoding chip will send the analog derivative directly (through several switches) to the screen bypassing the extra circuits.

Will 1080i/540p DVI in bypass mode be better than 1080i/540p components in bypass mode? My guess is yes. Both methods have only one D/A conversion. But Sony is using a very high quality Genesis GM7030 DVI decoding chip to do the conversion and you would have to compare its performance to that of the DVI converter inside a STB or player with components output. Also some STB and players may not want to convert encrypted digital signals to 1080i components for copyright reasons.

Finally you ask if you should attempt the patch? I say wait until there is more folklore built up on this subject and Sony gets a chance to make some pronouncement on the subject. I suspect they won't do it until their new generation of products has had enough time to get a foothold in the market. Then they may comment in order to maintain consumer loyalty.
After all a happy XBR800 customer is one who tells his/her friends that he/she feels secure with Sony and he/she will make a future Sony purchase.

In a Nutshell, if that is even possible, if you are having problems with the
DVI-1080i line sweep and the year you bought your TV then this INformation is INvaluable and could save
you a few hundred dollars. I too don't know if I would attempt some of
these fix's. If I were going to do any adjustments after reading (from the
beginning again) the info that some of these guys posted and made sure
I follow every step there is know doubt in my mind that it is all correct and
very well thought out. I am all thumbs when it comes to fixing things but
i might give it a crack with this info on hand. Or not.

So. let me get this straight. I have a 36XBR800. If I want to perform the bypass, all I would need to do is have a coaxial input (say my cable) connected to the RF input and then seelct that input. Then go into the service menu and change HDPT to 0 rather than 1. Is this correct? I am just trying to make sure of a few points before I try them out (whenever my set gets back from repair). Were any other adjustments needed on the XBR800 in order to "fix" anything? As far as reducing the puchiness of the picture is concerned, is it THAT noticeable? If so, does VM help considerably to "fix" some of this? Sorry for all the questions but there seems to be a lot of extra discussion in here that is making getting the real details a little less obvious. I think some folks think that there is more to it than there really is...not sure though.

Quick question for a prospective owner of the 36HS510. Am I correct in assuming that ALL incoming 480i input signals are upconverted to 480p using the Genesis chip, regardless of whether they come in via DVI, S-Vid, Component or Composite? I currently have a Samsung TS160 DVI equipped STB and would prefer to watch SD D* programming in it's native interlaced format to prevent picture softness. The tv that I am replacing, RCA MM36110 showed all incoming 480i content as 480i as long as it did not come in through the VGA input. All VGA input signals were upconverted to 540p and let me tell you... to say the PQ was soft is a major understatement.

Thanks,
Jeff

I'm not sure if i have a 1080i scan problem because i've never had anything hooked up to the DVI!

How difficult and dangerous is it to do the bypass myself? If we do it ourselves does that void the warranty? If it's easy to do. Can you give a step by step to do it? I'm still afraid though.

Does it still give you a better on 480i DVD player?

In regards to the SMPTE color bars... I believe that HD NET is still doing this two times weekly via the DirecTV HDNET channel. Unfortunately, the last I looked it was like at 10 am on a Tuesday (at work) and 2 am on a Saturday (in bed). Check their DF format schedule for more info. Their HTML versions don't usually show it. I'll post a link to it if I get a chance.

Later,
Jeff

Added: Upon further review, it looks like HD NET is no longer doing this... at least according to their current programming schedule. I may send a message to their programming manager and see if it is just not being listed any longer.

Iceblade,

The Genesis chip is the actual DVI decoder chip. It only connects to the video 7 input, the DVI connector. It converts whatever arrives in an encrypted digital form to analog. Only HD formats are converted, 1080i and 720p. Neither 480i nor 480p are HD formats and neither will ever be transmitted over a DVI wire.

If you have a DVD player with a DVI output, then the player will rescale the numeric image from the disk to 1080i (or possibly 720p - a bad choice for CRTs because a subsequent conversion to 1080i is required) HD format and also do a pseudo encryption.

The Genesis chip in the Sony will decrypt the numbers and create an analog signal in 1080i because that is the format that the player is told (by you or the manufacturer) to output. This 1080i analog signal waits to be selected at the same 'gate' as all the other inputs (RF, components, composite, S-video, memory stick).

If you also have the bypass activated, then whatever is selected at the 'gate' by your Sony remote control will go directly to the CRT through the bypass circuit only if whatever is coming out of the gate is in 1080i or 540p, they having essentially the same scan frequency.

Hope this helps you

Another thing to be done is to tune up the image after changing HDPT in RF mode. You need to select a 1080i source entering the set either as a DVI signal or a components signal (ideally both) in order to tune up the picture. The signal can be from a test generator or any 1080i HD program arriving at your STB.
If your DVD player has a DVI connector forced to 1080i, then this signal will work as well.

Oops... I think I mispoke. It's the SI chip that does the scaling not the Genesis chip, right?

In regards to what I can transmit via DVI... I can set my TS160 to output 480p, 720p or 1080i. The Faroujda FLI2300 does the scaling from what I recall. In regards to the S-Vid and Composite or Component outputs, however, a 480i signal is sent out to the tv. My inquiry was basically... what happens to 480i content that ISN'T coming in through DVI. Re-reading my original question I didn't make that very clear, did I? So basically what it sounds like is that there is no way to defeat the upscaling of 480i material via S-Vid input, as they are all sent to the scaling chip, regardless of the bypass mode setting, correct?

Bummer. I have the same deal with my Samsung 61" DLP. Can't defeat the upconversion of D* content. Never thought I would pine the loss of the RCA's features. :(

Thanks,
Jeff

The SI chip, like the Genesis chip, is a DVI decoder. Sony uses either one in the same circuit, the choice depending on the model of the TV. If you look up the specs for the Si 905 chip you will find it as impressive as the Genesis GM7030.

Yes, they are routed to the DRC chip for conversion to 1080i regardless of the bypass setting. The CRT drive is not designed to operate with a scan rate of 15 khz for the 480i. We're not even sure if it operates at 30 Khz for the 480p. There's some debate as to whether 480p is also up-converted to 1080i so that there is one and only one format presented to the CRT, and therefore only one geometry calibration required for all signal types. If this is the case, one might imagine how the DRC does the anamorphic squeeze if there is no vertical compression at the hardware level, only in the DRC memory.


The Samsung DLP is a fixed pixel display. It can only display 720 lines vertically. Everything that enters the DLP must be rescaled to 720 lines.
There is no point in defeating anything in the DLP. Just make sure that whatever talks to the Samsung does so in 720p and over a DVI wire.
You can't loose with that combination.

ADU & montreal,

Thanks for the interesting info! This thread makes me want a Sony HD-CRT just to fiddle with all day long.

Just some notes about scanning frequencies and why most HD CRTs are dual scan:

The ubiquitous nature of 480i in society makes 31.5kHz an important frequency for HD tv's because it is easier and cheaper to de-interlace (rather crudely when it comes to video based material) than it is to scale up (non linearly)

In other words: If the tv only displayed at 33.75khz then any 480i signals would have to be de-interlaced to 480p and then scaled to 540p which would reduce picture quality.

480i to 1080i would be even worse because you would need to de-interlace to 480p, scale to 1080p, then re-interlace to 1080i.

It is better to display natively 480p/31.5khz (especially film material!!!)

Also, remember that many CRTs including the Sony line have SVM so they are technically multi-scan (even though only two are used). I think one of the SVM frequencies is acutally 31.5Khz (ie 480p/960i)

As for the 540p vs 1080i comparison it is good to remember that 1080i has 1080 lines per frame (2X540 fields per frame). Our eyes see 1080 lines on the screen even though only 540 lines are there for each scan. 540p has only 540 lines per frame and our eyes only see 540 lines.

So 1080i has twice the picture information as 540p but takes twice as long to show it. So in the end the data rate/scan rate is the same for both even though 1080i has more picture information

eye_candy
Some of the (fixes) have to be hard fixes (someone please tell me if I am at least on the right track please) depending on when your set was
manufactured. Some can be soft fixes Sony has updates on a web site.
If you are not experiencing any problems- don't fix it if it's not broke.
Some of this stuff is just trying to achieve the best picture you can and
that seems to have much to do with *Digital to *Analogue or *A to*D
conversions i.e keeping the signal Digital for as long as possible (how am
I doing)

XROX,

The DRC chip is already well equipped to convert 720p to 1080i. I don't see what would be so difficult about converting 480i or 480p to 1080i. You keep introducing the terms de-interlacer and re-scaler as if they were pieces of hardware. Once you have your video signal in a digital form, you can paint the numbers into the volatile memory using any old algorithm and at any sync frequency and independentally grab chunks of data at a different rate and convert them back into analog.

The only way to know if the CRT is a dual scan type is to scope the horizontal final drive or put a AM transistor radio near the fly-back transformer and listen for a change in pitch when selecting 1080i vs. 480p.

If the CRT can scan at 31khz like 33.75Khz, wouldn't 480p look better if it also used the bypass circuit and avoided the extra circuits.

Perhaps it already is but we don't see the difference between a bypassed 480p signal and a non-bypassed signal because this 13 mhz signal is already too low rez. for the DRC/MID circuit to erode it further.


DNINE,

Yes, some sets manufactured in Mexico were assembled with B boards missing the bypass circuits, so a board swap is a prerequisite before any soft fix can be applied. Some B boards work correctly even without the use of the bypass circuit because their owners have not noticed the problem or they don't use 1080i or their boards are adequately shielded.

It's more like A to D followed by reformatting followed by D to A. That's a lot of processing. We are trying to keep the signal completely analog all the way through using the bypass. The only time when we're forced to deal with a digital signal is when it enters the TV via the DVI connector and immediately changes to analog.

According to ADU, the 2003 Sony fixed pixel products may be trying to keep the signal as completely digital for as long as possible. I have no idea what the XBR910 does with the digital signal arriving by the DVI port.

But you can see the copyright problems if a DVI decoder chip were to release decrypted video in a numeric format that someone could tap into.
To prevent this you would need some kind of super chip that does everything that is currently achieved with discrete chips within the XBR chassis. I believe Genesis already has such a beast.

Man. What happens if the service tech did the HDPT fix with NOTHING in the RF input? He in no way had RF selected or anything in the input. He was watching up-converted 1080i with NOTHING in the RF input when he switched it.

montreal,

It is very, very difficult to convert 480i to 1080i properly. Why would all progressive DVD players do 480p, why not just output 1080i?

-because it is extremely difficult video processing (to do well)

I was just trying to point out that 480i to 480p requires no scaling only de-interlacing. Converting 480i to 1080i requires de-interlacing, scaling, and re-interlacing. These are all video processing done by the DRC. To me 480i to 1080i seems like a lot more processing that would not only be more difficult but also reduce picture quality.

Also, remember the DRC interlace setting is 960i (ie 31.5khz).

I honestly don't know for sure but if you still don't believe the Sony (and every other manufacturers HDTVs) don't natively scan 480p then you're AM radio experiment could tell you more. Or you could call Sony and ask them :)

It would seem that the multiple conversions are the problem with the *A
signal or how good the processor that is doing each conversion is-
a big issue. What I do not understand is if the signal is *D and get's
converted X amount of times that signal should always be your best signal
( and best quality PQ) a *D signal is not subject to change 1s and 0s are and will always be that no matter how many conversion this goes thru unless the same rule applies as with the *A conversion what kind of processor is doing each convert still then it could not remain a true *D signal. I understand most of this would not apply to CRT'S but what am I missing? or not I also know I am on page 2 and you guts are on page 2002 or 2003 for ADU, but i am trying.