Step by step surgery

[fas42]

I knew there was a reason I was going to go through this exercise of optimising the 65" TV image, 😁 ...

 

There are so many similarities in what's going on, to audio rig optimising ... the bigger screen is automatically "more in your face", and every tiny thing not quite right disturbs one; the set, like all sets, no matter how expensive, is not perfect, and has its own quirks, behaviours - trying to treat it like a straightforward "get: all the numbers right!" won't work, because the idiosyncrasies will not be 'masked enough'; each parameter will need to be pushed in some direction to achieve a best compromise, and some of the choices won't make sense, technically.

 

Overall, I have a pink problem - the set keeps wanting to add a pink/purplish veil over the image, and "doing the right things" tends to make it worse. Yesterday I thought I was getting somewhere, by largely doing it by the book - but later that night the colour cast was really pissing me off - in disgust, I switched to another colour temperature ... hmmm, at least that bugs me less ...

 

On a cold start next morning, hey!! Pretty good - the other settings made, with the abruptly chosen temperature was working in my favour, the colours were very much in the right zone. Now, will it hold as the set goes through, yep, you've guessed it - warmup stabilisation? ... this set takes at least an hour before it fully settles ... forget about 15 minutes! Uh oh, the dreaded tint is there, yet again .. another round, to come.

 

Been here a million times with audio, same thing with video - you think you are making a good move, based on prior experience, and "good advice" - but it turns out otherwise; each situation is unique, and you have to learn how to "tame the beast". Been looking up other people's experiences, and there are plenty of horror stories of people going nuts trying to get a picture that doesn't bug them for some reason - literally going through a half dozen sets before they find one that "works", for them. Brand, model, price matter little - each set is an individual, and if you're sensitive to some anomaly that it has, then it will never satisfy you.

[Racerxnet]

i1Pro , Klein or other meter?

Calman, HCFR or other calibration software

Whats your target colorspace?  601, 709, P3, 2020

D65

0-255 or 16-235

4:2:0, 4:2;2, 4:4;4? 

Bit depth? 8,10,12

Gamma 2.2, 2.4 or what??

 

Without a meter and software you are pissing in the wind. You'll never see the movie as the director intended or where your TV's faults are.

 

MAK

[fas42]

I have no trouble seeing where the set's faults are ... the big one is that when one reaches the top end of the white space, it goes from pink to green, 😊 - so, I have to work around that ...

 

I'm using a Joe Kane's DVD, which is showing me the issues, when using flat fields, and gray scale ramps. Quite a bit of DSE, with colour variations in that - it's going to be a compromise, no matter what. Again, I'm not interested in playing numbers games - I have got it to the point where the colour balance has been neutral enough so that each new shot shows the distinct look of that scene, but this was only when cold; what matters to me is getting to the point where I'm no longer aware of the TV's colour cast.

 

The prior set was still well out when "done by the book"; it took some weeks to fine tune it, so that it always provided a 'comfortable' viewing experience.

[semente]

You really should try a calibration device.

 

[fas42]

The theory's good, but I would be very annoyed if it still didn't shape up - I tried another more technical approach today, but it still ended up with a colour cast. So, I threw caution to the winds and tried a variation of what I used to fine tune the previous set - flick through channels one after another, and play with each of the RGB guns as appropriate, to get skin tones and vegetation to ring true, for each new channel colour combo. This worked nicely, and I was able to steadily zero in to quite a good approximation - under daylight conditions - to what I had before. This means that the skin tone matched the lighting conditions present in the scene - how people looked "felt right", for the style of the broadcast. That is, people are pink, purple, orange, white, yellow, or red, even natural - depending upon the channel, and program/ad.

 

Night time viewing still needs work - there are very obvious 'glitches' in how near black is handled, on the sides of the screen; I need to find a good compromise there.

 

Once I've settled on the rough settings for daylight, heavy cloud, and nighttime I'll play with subtle variations of single parameters, and see how that sits, for a day or two - the idea is to squeeze the last ounce of subjective performance out of the set ... the best compromise possible.

 

 

[Racerxnet]

Compare your tv to a screen cap and look at the color variation from the 2. The screen cap is your reference as the director intended. Using a broadcast (OTA) signal for reference is incorrect. You can go to Curt Palme's site or AVS and learn about calibration. AVS has files to download for free. Your style of calibration has been debunked long ago. You can see the debates in "screenshot wars" .  Good luck with subjective calibration. It does not exist in the commercial market or when paying for/delivering services. 

 

MAK

 

MAK

[tmtomh]

Frank's entire calibration regimen is based on a rejection of the notion that there's any objective calibrator or standard or device that can be used as a reference. It's a variation of hyper-subjectivism and it's what fuels a lot of the audiophile industry (although I hasten to add that I don't view Frank as an audio subjectivist per se, because he appears to be after some version of high fidelity/realism/accuracy that's different from the more euphonic goals that some folks have).

 

I'm fine with Frank doing things this way, and even with him sharing his experiences here - free speech, live and let live, and so on. And it's clear that for Frank, analogously to those in our hobby who really enjoy endless equipment swapping and tinkering, the rejection of an objective standard is what opens up all the contrived mysteries and options that make for hours and hours of enjoyable experimentation.

 

The only problem I have is when Frank ventures - as he inevitably does - into "objective references are just 'theory' and don't actually make for good sound/good picture quality" territory. Then he's just like those folks who go beyond liking tube gear, and venture into claiming that tube gear has higher fidelity because of its distortions and nonlinearities.

 

Frank, you seem like such a nice and good-natured guy, and I applaud your spirit of exploration, but dude, your eyes + broadcast TV signals do not equal anything close to a proper calibration app or setup.

[semente]

5 hours ago, fas42 said:

The theory's good, but I would be very annoyed if it still didn't shape up - I tried another more technical approach today, but it still ended up with a colour cast. So, I threw caution to the winds and tried a variation of what I used to fine tune the previous set - flick through channels one after another, and play with each of the RGB guns as appropriate, to get skin tones and vegetation to ring true, for each new channel colour combo. This worked nicely, and I was able to steadily zero in to quite a good approximation - under daylight conditions - to what I had before. This means that the skin tone matched the lighting conditions present in the scene - how people looked "felt right", for the style of the broadcast. That is, people are pink, purple, orange, white, yellow, or red, even natural - depending upon the channel, and program/ad.

 

Night time viewing still needs work - there are very obvious 'glitches' in how near black is handled, on the sides of the screen; I need to find a good compromise there.

 

Once I've settled on the rough settings for daylight, heavy cloud, and nighttime I'll play with subtle variations of single parameters, and see how that sits, for a day or two - the idea is to squeeze the last ounce of subjective performance out of the set ... the best compromise possible.

 

 

 

You put too much trust in the accuracy of the image provided by different broadcasters.

 

Even the news read by the same person and broadcast in BBC Radio 4 doesn't sound like that same person reading the same news at the same time on the Parliament Channel or BBC World News or other BBC Channel.

 

Try using a static photo that you know is accurate (use google).

[semente]

I used to edit photos on a LaCie Electron Blue CRT and the calibration file produced by the Monaco Optix (later bought by X-Rite) was a significant improvement.

[John Dyson]

57 minutes ago, tmtomh said:

Frank's entire calibration regimen is based on a rejection of the notion that there's any objective calibrator or standard or device that can be used as a reference. It's a variation of hyper-subjectivism and it's what fuels a lot of the audiophile industry (although I hasten to add that I don't view Frank as an audio subjectivist per se, because he appears to be after some version of high fidelity/realism/accuracy that's different from the more euphonic goals that some folks have).

 

I'm fine with Frank doing things this way, and even with him sharing his experiences here - free speech, live and let live, and so on. And it's clear that for Frank, analogously to those in our hobby who really enjoy endless equipment swapping and tinkering, the rejection of an objective standard is what opens up all the contrived mysteries and options that make for hours and hours of enjoyable experimentation.

 

The only problem I have is when Frank ventures - as he inevitably does - into "objective references are just 'theory' and don't actually make for good sound/good picture quality" territory. Then he's just like those folks who go beyond liking tube gear, and venture into claiming that tube gear has higher fidelity because of its distortions and nonlinearities.

 

Frank, you seem like such a nice and good-natured guy, and I applaud your spirit of exploration, but dude, your eyes + broadcast TV signals do not equal anything close to a proper calibration app or setup.

I am currently doing some stuff where it has to be subjective -- and it is training me more and more strongly that using normal objective tools and techniques are SOOO much easier and a lot less work.  (Well, you know what I mean.)  It is very tricky to make sense of a subjective measurements when the goal is an objectively accurate result...  Everything has to be compared step-by-step, and is still too error prone to make it easy.

Right now, I am fairly 'money poor', I would be very tempted to spend $1000 on a tool that would accurately measure whether or not the results of the side-project of DA decoding existing commercial materal is correct.  Getting even plausible results is so much work, and almost to the point of being hit-or-miss to get accurate results. (Actual master tapes are easy to decode -- the problem is with undoing a complex set of EQ.)

 

I'd suspect that video would be worse in a alot of ways, because many static colors can be created (mixed) in different ways -- esp when the primaries can vary.   I never did fully calibrate my D9 Camcorder, didn't need to -- because I wasn't doing all that much playing around with multiple cameras and sources -- I was just playing.   Subjectively tweaking a complex system for accurate results reminds me of playing whack-a-mole.

 

John

 

[fas42]

7 hours ago, semente said:

 

You put too much trust in the accuracy of the image provided by different broadcasters.

 

Even the news read by the same person and broadcast in BBC Radio 4 doesn't sound like that same person reading the same news at the same time on the Parliament Channel or BBC World News or other BBC Channel.

 

Try using a static photo that you know is accurate (use google).

 

You miss the point ... I want to watch TV, and not be disturbed by the visual contradictions of the colour in the image. Let's take a specific example - our current PM is being shown a lot, for some strange reason 😁, and the colour of his face is different in every situation - just this morning he was being interviewed, and his skin tone was very pink - why? Because the lighting and background were strongly in that direction, you could pick that every element on the screen was pushed in that direction. He has also been greeny yellow, red, orange, and even at times, gasp!, 'natural'. And each time the lighting of the scene obviously was directed in those particular directions - his skin tone matched what everything else in the frame strongly signaled.

 

I don't expect the TV image to be "correct" - but I expect everything in that image to correlate with what my memory tells me how colours alter, depending upon the lighting conditions.

[kumakuma]

21 minutes ago, fas42 said:

 

You miss the point ... I want to watch TV, and not be disturbed by the visual contradictions of the colour in the image. Let's take a specific example - our current PM is being shown a lot, for some strange reason 😁, and the colour of his face is different in every situation - just this morning he was being interviewed, and his skin tone was very pink - why? Because the lighting and background were strongly in that direction, you could pick that every element on the screen was pushed in that direction. He has also been greeny yellow, red, orange, and even at times, gasp!, 'natural'. And each time the lighting of the scene obviously was directed in those particular directions - his skin tone matched what everything else in the frame strongly signaled.

 

I don't expect the TV image to be "correct" - but I expect everything in that image to correlate with what my memory tells me how colours alter, depending upon the lighting conditions.

 

I'm confused as to why you're not following the same procedure that you use with your audio equipment.

 

Open that sucker up and start replacing everything inside that is degrading signal quality! 

 

Eventually your PM's skin tone will be totally natural regardless of how's it's been captured on camera.

[fas42]

7 hours ago, tmtomh said:

Frank's entire calibration regimen is based on a rejection of the notion that there's any objective calibrator or standard or device that can be used as a reference. It's a variation of hyper-subjectivism and it's what fuels a lot of the audiophile industry (although I hasten to add that I don't view Frank as an audio subjectivist per se, because he appears to be after some version of high fidelity/realism/accuracy that's different from the more euphonic goals that some folks have).

 

If there's an objective calibrator that I can be assured is 100% accurate, and that can take into account idiosyncrasies of the particular DUT, to work towards a best compromise, then I would be happy to use it - provided it is not hideously expensive! My experience of the Joe Kane DVD, which used straightforward colour filters, and test patterns, ended up when "going by the book" to give a subjective PQ (Picture Quality) well down on acceptable. So I then fine tuned this earlier set, purely by eye - which resulted in a picture which always "rang true" ... which is what I was after.

 

7 hours ago, tmtomh said:

 

I'm fine with Frank doing things this way, and even with him sharing his experiences here - free speech, live and let live, and so on. And it's clear that for Frank, analogously to those in our hobby who really enjoy endless equipment swapping and tinkering, the rejection of an objective standard is what opens up all the contrived mysteries and options that make for hours and hours of enjoyable experimentation.

 

The only problem I have is when Frank ventures - as he inevitably does - into "objective references are just 'theory' and don't actually make for good sound/good picture quality" territory. Then he's just like those folks who go beyond liking tube gear, and venture into claiming that tube gear has higher fidelity because of its distortions and nonlinearities.

 

The world of audio is a direct mirror. When the tuning of the output is in the "right zone", you are far less, or no longer aware of the device doing the presenting. When a TV is wrong, every new scene "jars", because the colours don't make sense - the unit is imposing too much of itself ... I go to someone's place, and they're watching TV intently - I'm just thinking, this colouring is impossible to take seriously, and I get bored and irritated, having to keep dealing with the TV thrusting itself at me ...

 

7 hours ago, tmtomh said:

 

Frank, you seem like such a nice and good-natured guy, and I applaud your spirit of exploration, but dude, your eyes + broadcast TV signals do not equal anything close to a proper calibration app or setup.

 

The TV is a conduit to the video images, and programming of interest - if it's doing its job as well as it can then it "disappears" and what I'm interested in takes centre stage, in my mind.

 

[fas42]

1 minute ago, kumakuma said:

 

I'm confused as to why you're not following the same procedure that you follow with your audio equipment.

 

Open that sucker up and start replacing everything inside that is degrading signal quality! 

 

Eventually your PM's skin tone will be totally natural regardless of how's it's been captured on camera.

 

Nice try... 🙂

 

The point is that I'm doing what anyone else can do, just by playing with their remote ... what's different is that I'm after a goal of working out the best compromise of using settings available to me, to deliver a picture that never "rubs me the wrong way". A broadcast may have overcooked colour, be washed out, have a savage colour cast, be blurred, ultra-sharp, poor motion blur - but always "fits" the context ... this is what my mind is most at ease with - I'm not trying to have the TV 'fight' the broadcast, make it look the way "I reckon it should look!"

[kumakuma]

9 minutes ago, fas42 said:

 

Nice try... 🙂

 

The point is that I'm doing what anyone else can do, just by playing with their remote ... what's different is that I'm after a goal of working out the best compromise of using settings available to me, to deliver a picture that never "rubs me the wrong way". A broadcast may have overcooked colour, be washed out, have a savage colour cast, be blurred, ultra-sharp, poor motion blur - but always "fits" the context ... this is what my mind is most at ease with - I'm not trying to have the TV 'fight' the broadcast, make it look the way "I reckon it should look!"

 

I was only half joking.

 

Surely you must see that the fundamental problem is the same one that you believe you have solved in the audio realm.

[fas42]

11 minutes ago, kumakuma said:

 

I was only half joking.

 

Surely you must see that the fundamental problem is the same one that you believe you have solved in the audio realm.

 

Problem? What I'm saying is that the graphic presentation, like an audio recording, is itself - the aim is remove all evidence that the device presenting such is 'doctoring' what you perceive - the TV if "perfect" will do that, but from I've read all sets one normally buys are compromised, they're not "as good as they could be"; and they would be substantially more expensive if "made better" - and why the top end sets are dramatically more costly.

 

Working within those constraints, you can still push a particular set to do the best that its limitations allow - this gives a viewing that is always comfortable .. and of course that's what I say happens with audio ..,

 

Just to add the latest tweaking, the backlight is now set at 100% intensity, and I've pushed the colour higher, for daylight. This was possible because the colours barely irritate now, so I can raise the intensity of the presentation without any downsides. Of course, contrast and brightness are set to squeeze every ounce of gradation from the image ... this mirrors being able to play audio at higher volumes.

[fas42]

8 hours ago, John Dyson said:

 

I'd suspect that video would be worse in a alot of ways, because many static colors can be created (mixed) in different ways -- esp when the primaries can vary.   I never did fully calibrate my D9 Camcorder, didn't need to -- because I wasn't doing all that much playing around with multiple cameras and sources -- I was just playing.   Subjectively tweaking a complex system for accurate results reminds me of playing whack-a-mole.

 

John

 

 

It's not as distinctive as what occurs with audio, but with video you can get it into a zone where it always "feels right" - no matter how tilted the colours, etc, are, they still jell. A clear giveaway is when a completely everyday shot, like a news story of an incident on a street in your area is shown - and everything in the clip is nailed: the sky is right, the colour of the house bricks is right, the trees on the side of the road, the various skin tones of all the people who witnessed it, the concrete of the pavement, the road surface - nothing stands out as having an incorrect colour.

[John Dyson]

36 minutes ago, fas42 said:

 

It's not as distinctive as what occurs with audio, but with video you can get it into a zone where it always "feels right" - no matter how tilted the colours, etc, are, they still jell. A clear giveaway is when a completely everyday shot, like a news story of an incident on a street in your area is shown - and everything in the clip is nailed: the sky is right, the colour of the house bricks is right, the trees on the side of the road, the various skin tones of all the people who witnessed it, the concrete of the pavement, the road surface - nothing stands out as having an incorrect colour.

 

Colors are complicated -- might seem easy, but it best to try to use objective schemes -- subjective can be very time consuming.  I don't like spending time tweaking.  There are better ways to spend time.

Measure twice, cut once?  Trimming takes time.

 

John

 

[fas42]

21 minutes ago, John Dyson said:

 

Colors are complicated -- might seem easy, but it best to try to use objective schemes -- subjective can be very time consuming.  I don't like spending time tweaking.  There are better ways to spend time.

Measure twice, cut once?  Trimming takes time.

 

John

 

 

Agree, trimming takes time - I spent weeks on the previous, low cost TV ... at the end of the period I was moving the Green offset and gain by a single increment, up or down, day after day to absolutely nail the colour of vegetation. Tedious, but it was worth it - I could watch the local gardening show, and every bit of foliage, bark and flowers came across as being spot on ...

[semente]

11 hours ago, fas42 said:

 

You miss the point ... I want to watch TV, and not be disturbed by the visual contradictions of the colour in the image. Let's take a specific example - our current PM is being shown a lot, for some strange reason 😁, and the colour of his face is different in every situation - just this morning he was being interviewed, and his skin tone was very pink - why? Because the lighting and background were strongly in that direction, you could pick that every element on the screen was pushed in that direction. He has also been greeny yellow, red, orange, and even at times, gasp!, 'natural'. And each time the lighting of the scene obviously was directed in those particular directions - his skin tone matched what everything else in the frame strongly signaled.

 

I don't expect the TV image to be "correct" - but I expect everything in that image to correlate with what my memory tells me how colours alter, depending upon the lighting conditions.

 

How your TV displays colour is static whilst casts and hue deviations depend on the programme material.

 

If you get your TV to display colours as accurately as it possibly can then it may not improve the worst cases but the better ones will look as good as they possibly can.

 

The alternative is to ride the tint, hue and saturation control for every single programme...

 

 

It's the same in audio with euphonic colourations. Such equipment may indeed perceptually improve some recordings but it will also tint any thing it plays indiscriminately. Like wearing tinted eye-glasses all the time.

[semente]

9 hours ago, fas42 said:

 

Agree, trimming takes time - I spent weeks on the previous, low cost TV ... at the end of the period I was moving the Green offset and gain by a single increment, up or down, day after day to absolutely nail the colour of vegetation. Tedious, but it was worth it - I could watch the local gardening show, and every bit of foliage, bark and flowers came across as being spot on ...

 

Use your time wisely:

 

 

[Racerxnet]

If the device as shown is not calibrated then it is not worth the effort. It looks like a cellphone. i1Pro is cheap enough and calibrated from the factory, will sometimes come with software, and the process can be automated with a laptop. Frank can listen to his tunes and calibrate at the same time. I've seen i1Pros for 150.00 used. I have looked into using the cellphone long ago and determined it did not meet the accuracy required on the Sony 665. 

 

Honestly, Frank does not know what he is doing and just likes to fiddle. It's his Tv and enjoys the time spent. 

 

MAK

[fas42]

36 minutes ago, semente said:

 

How your TV displays colour is static whilst casts and hue deviations depend on the programme material.

 

If you get your TV to display colours as accurately as it possibly can then it may not improve the worst cases but the better ones will look as good as they possibly can.

 

Yes. The programme material dictates what you see, not the TV. The point is not "to improve it", but for the video footage as prepared to stand on its own feet - you are seeing it as a combination of the artistic decisions made by the creators, and how it ended up in its final source form.

 

Just this evening watched an episode of Maigret, the recent British series with Rowan Atkinson - very good production values, with a highly effective visual style. Lots of sombre, dark tones - which clearly showed that this set has difficulty with close to black colours - an odd colour cast, with a curious bloom appeared regularly ... can this be adjusted away by very careful fine tuning of the normal settings; or is it something related to some "helper processing" in the set? I won't know until I do some careful experimenting, and investigate the service menu further ... one thing I am quite confident of - calibrating by the book would almost certainly not have been of any use here.

 

36 minutes ago, semente said:

 

The alternative is to ride the tint, hue and saturation control for every single programme...

 

 

It's the same in audio with euphonic colourations. Such equipment may indeed perceptually improve some recordings but it will also tint any thing it plays indiscriminately. Like wearing tinted eye-glasses all the time.

 

The previous set allowed me to watch anything, and not be disconcerted by the look - the aim is to to do the same again.

 

In audio even more so the recording can stand on its own feet - with sufficient accuracy of replay the mind can "hear past" what's poorly executed in the copy ... no tinting required.

[Racerxnet]

1 hour ago, fas42 said:

one thing I am quite confident of - calibrating by the book would almost certainly not have been of any use here.

 

You don't have a clue what you are talking about. Good luck. 

 

MAK

[fas42]

Dear me 🤨 ... since you mentioned Calman software, I wandered around their website - okay, talking the same talk as Joe Kane; nothing new here. Pretty obvious there is almost no chance of them having advice on this close to black colour issue - most relevant was this, https://kb.portrait.com/help/2-point-grayscale-adjust, which is the area I have been playing with for the last few days.

 

Of particular note,

 

Quote

When both Gain and Offset adjustments are optimized, perform a Read Series measurement, to view the display’s grayscale tracking performance on the full scale RGB Balance chart and to evaluate the error on the DeltaE chart. Mouse over the error bars on the DeltaE chart to view a visual representation of each grayscale color error.

 

Well, yippee - I now have a set of numbers that tell me how imperfect my set is - where's the extra bit that advises me how to play with the settings so that I can get the best visual result, that minimise the subjective impact of those imperfections, hmmm ... Joe Kane doesn't do it, and I suspect no-one else does ...

 

This is why this thread is called step by step surgery - you do the main, obvious bits, then step back and let it sink in for a bit, then have another go at tweaking it a touch more, trying something you haven't considered before - it's always an iterative process, circling steadily in onto the dead centre of the best that the system can do, as far as the senses are concerned.