1. Join Now

    AVForums.com uses cookies. By continuing to use this site, you are agreeing to our use of cookies. Learn More.

Just bought SMART 3

Discussion in 'Home Cinema Buying & Building' started by Mr.D, Jan 11, 2005.

  1. Mr.D

    Mr.D
    Well-known Member

    Joined:
    Jul 14, 2000
    Messages:
    11,061
    Products Owned:
    0
    Products Wanted:
    0
    Trophy Points:
    133
    Ratings:
    +1,149
    I got bored , I've somewhat finished tweaking my HTPC so I thought I'd give this a bash and fiddle with my HS20 some more...I'm not bored enough to buy colorfacts yet!
     
  2. avanzato

    avanzato
    Active Member

    Joined:
    Jan 21, 2003
    Messages:
    1,881
    Products Owned:
    0
    Products Wanted:
    0
    Trophy Points:
    51
    Location:
    Norfolk
    Ratings:
    +61
    What is this Smart 3?
    Is it like the Monaco Optix I use at work?
     
  3. Mr.D

    Mr.D
    Well-known Member

    Joined:
    Jul 14, 2000
    Messages:
    11,061
    Products Owned:
    0
    Products Wanted:
    0
    Trophy Points:
    133
    Ratings:
    +1,149
    Monaco optix looks like it profiles a monitor builds a software LUT and applies it to the graphics hardware ( I hope otherwise its just correcting the desktop). More sophisticated systems allow you to first calibrate the display hardware itself ( a closed loop calibration).

    The SMART 3 is a poor man's calibration system which uses a series of idealised targets for a specific display device , in my case an HS20 video projector. Basically you do a full objective calibration on a display , once its calibrated you run a series of test pattens through it and measure it with a tristimulus photometer (RGB values) poor mans colour analyser.

    Rather than try to calibrate the display to a measurable 6500K in this case you calibrate it to give the same values (or ratios given a variance in white point between displays) that you got with the calibrated device on the same patterns.

    This technique is reffered to as curve ripping and is often used as a form of back engineering other peoples designed response curves for things like different film stocks if the manufacturer doesn't make them public. (using this technique for this purpose is technically industrial espionage!)

    The disadvantages are that you can only get as accurate as the variance between your different photometers and the one used to do the original calibration. And you initially need a calibrated device to give you the aims in the first place. And if your particular display is actually a little bit superior to the calibrated one you effectively limit it to the idealised display performance.

    You are not really calibrating to 6500K ( excuse me D.65) you are approximating an already calibrated device. Its a little bit smarter than that to deal with variances over time and its main advantage is that it allows you to use colour correction filters to increase the contrast of your display without compromising the colour balance too much.

    At the end of the day its probaly not as accurate or ideal as a proper objective colour calibration but its probably going to do a better job than just eyeballing the grayscale.
     
  4. Alaric

    Alaric
    Active Member

    Joined:
    Mar 12, 2001
    Messages:
    1,618
    Products Owned:
    0
    Products Wanted:
    0
    Trophy Points:
    66
    Location:
    Peterborough, Cambs, UK
    Ratings:
    +108
    Smart 3....er is that the ForFour (Smart ForTwo, Roadster were before) :)

    Got any web site details....seen the colorfacts and also thought it was a touch pricey, also seen some bits done with a photocell...so this sounds interesting.

    Ta,
    Lee
     
  5. Mr.D

    Mr.D
    Well-known Member

    Joined:
    Jul 14, 2000
    Messages:
    11,061
    Products Owned:
    0
    Products Wanted:
    0
    Trophy Points:
    133
    Ratings:
    +1,149
  6. Alaric

    Alaric
    Active Member

    Joined:
    Mar 12, 2001
    Messages:
    1,618
    Products Owned:
    0
    Products Wanted:
    0
    Trophy Points:
    66
    Location:
    Peterborough, Cambs, UK
    Ratings:
    +108
    Ack....doesn't do CRT. *sulk*
     
  7. avanzato

    avanzato
    Active Member

    Joined:
    Jan 21, 2003
    Messages:
    1,881
    Products Owned:
    0
    Products Wanted:
    0
    Trophy Points:
    51
    Location:
    Norfolk
    Ratings:
    +61
    AFAIK the Optix software applies LUT's to the hardware. Monaco are trying to get me to upgrade to the Pro version which if I'm reading the manual right will measure the monitor (luminance/RGB) while you adjust it to target values. But I could be wrong and I'm not sure how sophisticated to go, or what the expensive pro packages measure.
     
  8. Mr.D

    Mr.D
    Well-known Member

    Joined:
    Jul 14, 2000
    Messages:
    11,061
    Products Owned:
    0
    Products Wanted:
    0
    Trophy Points:
    133
    Ratings:
    +1,149
    Well I had a go with the SMART calibration system on my HS20.
    Apparently the vpl11ht and following models all have some sort of auto black level control that is good in some ways but messes with the cal. Not sure this is correct from what I've seen of my HS20 over DVI from an HTPC.

    I ran a measurement sequence which was useful as it gave me an idea of what the projector was doing (heavy in red beyond 50IRE but quite accurate with regard to tracking and gamma, very blue towards black), then rather than follow the recommendations that smart gave me I applied a CCR20 filter and then eyeballed the grayscale back to where it should be ( I think I might be not too bad at this because of the job I do).

    Going by the theory I managed to significantly up the blue and green ( my HS20 always went red in the whites beyond a certain contrast level). Now I've mangaged to get it so that the contrast can be raised significantly without clipping or any discolouration and the filter only made a slight difference to the blacks so I only had to nudge the brightness up a click.

    So I'm seeing a better contrast range , blacker blacks with no loss of detail , better grayscale tracking and more accurate colour ( this has improved significantly previously I couldn't get the colour and hue quite right which points to a naughty grayscale now its bang on looking at Avia and DVE.

    I'll run through another measurement cycle with the new settings and see if there is any improvement according to smart3 but I sort of know there is already.

    So its a useful system in some ways , not sure if I would say its essential for the HS20 and others but its more accurate than just eyeballing alone.
     
  9. JohnS

    JohnS
    Active Member

    Joined:
    Oct 31, 2000
    Messages:
    2,580
    Products Owned:
    0
    Products Wanted:
    0
    Trophy Points:
    71
    Location:
    Herne Bay, Kent
    Ratings:
    +33
    Let us know how this progresses Keith, this is something I was thinking about buying myself but wondered if, having tweaked a fair bit already, if there is much more to gain.
     
  10. Gary Lightfoot

    Gary Lightfoot
    Distinguished Member

    Joined:
    Apr 6, 2001
    Messages:
    12,464
    Products Owned:
    1
    Products Wanted:
    1
    Trophy Points:
    166
    Location:
    Surrey. UK.
    Ratings:
    +2,114
    Did you take before and after contrast readings Keith? What are you getting now?

    Gary.
     
  11. Mr.D

    Mr.D
    Well-known Member

    Joined:
    Jul 14, 2000
    Messages:
    11,061
    Products Owned:
    0
    Products Wanted:
    0
    Trophy Points:
    133
    Ratings:
    +1,149
    From memory it was about 360 ...what this refers to as far as SMART3 is concerned is anyones guess I assume its 1:360. Smart classified this as "OK" prior to tweaking.
     
  12. Gary Lightfoot

    Gary Lightfoot
    Distinguished Member

    Joined:
    Apr 6, 2001
    Messages:
    12,464
    Products Owned:
    1
    Products Wanted:
    1
    Trophy Points:
    166
    Location:
    Surrey. UK.
    Ratings:
    +2,114
    Did you take a white and black reading and divide the black into the white to get that result?
     
  13. Mr.D

    Mr.D
    Well-known Member

    Joined:
    Jul 14, 2000
    Messages:
    11,061
    Products Owned:
    0
    Products Wanted:
    0
    Trophy Points:
    133
    Ratings:
    +1,149
    Nope I took a peak white reading and fed it into SMART3 , this is before the initial measurement stage probably so that SMART3 can work out what the "unit" values (because SMART3 is subjective rather than objective) are for the contrast range of my particular projector and than probably scales the aim values based on the calibrated model that SMART3 originally referenced. It alos incorporates correctional offsets for the meter itself that the guy(s) at SMART have referenced agains the original one ...I assume.

    Later on in the measurement stage you do measure the tristimulus values for 0IRE but not explicitly the black point itself with the cosine sensor in place ( as you do for intital projector gain value). So either its working out what the contrast is from the tristimulus measurements themselves or its taking the initial white point value and dividing it by the minimun recordable value of the meter.

    So whilst I haven't explicitly measured the blackpoint ( it may not be too accurate with this meter , certainly for the actual tristimulus measurement you are told to place it nearer the projector to increase accuracy in the lower intensities to give the meter something to work with most likely) there is a contrast range worked out by the spreadsheet in the course of the measurements.

    There are a lot of things that I''m having trouble making sense of , gamma response vs grayscale tracking graphs seem to contradict each other a bit but I"m sure its just a question of tying it altogether.

    The advice that SMART gives you to readjust the projector is in terms of percentage , ie increase blue gain by 36%. Lower brightness by 20 units. The problem is what ias it referring to ? Percentage of gain value in the service menu or percentage of blue luminance reading in the upper end? Feel it has to be specifically referring to the projector controls increments as it doesn't talk in terms of referencing a specific IRE level for the 36% increase. Thats effectively what I did plus a little eyeballing of the grayscale so the next measurement run I make will let me know if I've improved things ( with the filter in place as well).

    There is also a different method by setting a specific red point for example and getting value indications of what to adjust blue and green by to recalibrate the high end.

    To be honest the spreadsheet itself is much more sophisticated than I expected but its really all about trying to find ways around not having a colour analyser and getting the system itself to be relevant to more than one specific projector.

    I think its a very good idea but its not the most intuative thing in the world ( it would be better if it gave you more specific targets to set your projector to but I'm not sure it works like this , all about offsets from an ideal based on your projectors mean capabilities type malarkey).
     
  14. Gary Lightfoot

    Gary Lightfoot
    Distinguished Member

    Joined:
    Apr 6, 2001
    Messages:
    12,464
    Products Owned:
    1
    Products Wanted:
    1
    Trophy Points:
    166
    Location:
    Surrey. UK.
    Ratings:
    +2,114
    I find the light meter itself is useful for taking CR readings and for calculating genuine lumens at the screen. It's got quite a range, so you can select the most appropriate one for the situation i.e black reading or white reading (I found taking those closer to the pj got more accurate results), and the results it gets aren't far away from Colorfacts. I find the light meter is better for taking readings like that, and leave the CF colorimiter for colour balancing etc.

    If you had a low gamma reading, say 2.1 and wanted to get 2.2, what would you change to do that? Adjust the brightness of the colours or something else, like the gamma on the DVD player for example?

    Gary.
     
  15. Mr.D

    Mr.D
    Well-known Member

    Joined:
    Jul 14, 2000
    Messages:
    11,061
    Products Owned:
    0
    Products Wanted:
    0
    Trophy Points:
    133
    Ratings:
    +1,149
    If you had a source that had a decent gamma control I'd adjust that and leave the projector where it was. If you wanted multiple sources with no overall or reliable gamma control I'd recal the display most likely.
     
  16. Gary Lightfoot

    Gary Lightfoot
    Distinguished Member

    Joined:
    Apr 6, 2001
    Messages:
    12,464
    Products Owned:
    1
    Products Wanted:
    1
    Trophy Points:
    166
    Location:
    Surrey. UK.
    Ratings:
    +2,114
    Thanks Keith.

    What I did with my HTPC and HT1000 was to adjust the ATI overlay so that the gamma increased and that did help, but I wondered if there was another way to do it without adjusting the gamma.

    Gary.
     
  17. Mr.D

    Mr.D
    Well-known Member

    Joined:
    Jul 14, 2000
    Messages:
    11,061
    Products Owned:
    0
    Products Wanted:
    0
    Trophy Points:
    133
    Ratings:
    +1,149
    Yeah I like this.

    Initially I didn't like what smart was telling me and eyeballed things instead bearing in mind what the initial measurement had told me about the sony's capabilities.

    With CC20R filter in place and re-establishing grayscale I got the contrast range up from 360 initially to 580. However I was pretty sure I'd introduced a bit of a blue hump into the middle of the grayscale ( a bad place to get the hump when you have bias and gain controls).

    Having found where to enter the projectors bias and gain in the spreadsheet I got more precise information out of smart ie set green to 180 set blue to 155 sort of thing. this was close to what I had eyeballed in the gains but smart3 recommends you leave the hs20 bias alone because of how the projector holds black....I hadn't as I thought I knew better.

    After inputting the precise smart values and returning the bias settings to where they were originally I've got rid of the hump . I need to run another measurement cycle but I am very happy with what I'm seeing.

    There is an awful lot of real knowledge and actual experience with the projector model I feel embedded in SMART3...definite thumbs up.
     
  18. Gary Lightfoot

    Gary Lightfoot
    Distinguished Member

    Joined:
    Apr 6, 2001
    Messages:
    12,464
    Products Owned:
    1
    Products Wanted:
    1
    Trophy Points:
    166
    Location:
    Surrey. UK.
    Ratings:
    +2,114
    Having a pj specific Excel sheet can make all the difference. I only had a generic DLP so things weren't as good as I'd hoped. It's still a great piece of kit for the money though.

    Gary.
     
  19. Mr.D

    Mr.D
    Well-known Member

    Joined:
    Jul 14, 2000
    Messages:
    11,061
    Products Owned:
    0
    Products Wanted:
    0
    Trophy Points:
    133
    Ratings:
    +1,149
    Well I finally got round to doing a stack of calibration runs to optomise my HS20.

    Result...very happy. Grayscale tracking is now very smooth. I'm not entirely convinced the colour temp is right on the money ( looks a little cool to me) but I cannot fault the linearity and gamma. Huge improvement from where it started. I could never quite get rid of the pinking in the upper levels from the filter CC20R if memory serves. Smart managed it!

    The biggest problem I had with Smart was not doing as I was told! I had to resist the urge to not follow the instructions at the end of some runs ( dropping brightness an amount that you know is going to clip for example) . However it really is best to just let it follow its course and eventually give you the best results for your machine.

    Well worth a look . Now to see what the new version can offer.
     
  20. Gary Lightfoot

    Gary Lightfoot
    Distinguished Member

    Joined:
    Apr 6, 2001
    Messages:
    12,464
    Products Owned:
    1
    Products Wanted:
    1
    Trophy Points:
    166
    Location:
    Surrey. UK.
    Ratings:
    +2,114
    Sounds good Keith.

    Looking forward to seeing how the new version works out.

    Gary.
     
  21. ROne

    ROne
    Active Member

    Joined:
    Dec 21, 2000
    Messages:
    779
    Products Owned:
    0
    Products Wanted:
    0
    Trophy Points:
    18
    Ratings:
    +7
    I have the new version of SMART III and to be honest, I am struggling to assess the difference.

    You get a new updated excel sheet with the ability to pick projectors, but for me the projector was a sanyo Z3, and the drop down only stated sanyo - so I don't know how a generic projector profile could support all three models as they all perform so differently.

    There are some updated instructions and a paragraph or two about filters.

    I thought it was a bit of "mean" deal actually.
     
  22. Mr.D

    Mr.D
    Well-known Member

    Joined:
    Jul 14, 2000
    Messages:
    11,061
    Products Owned:
    0
    Products Wanted:
    0
    Trophy Points:
    133
    Ratings:
    +1,149
    I think as long as SMART knows that the white point of the projector is as high as it can be without clipping one of the primaries it will try and fit the corresponding desired calibrated intensity response into the available contrast range. Although the actual end capabilities will vary from projector to projector : from unit to unit as well as model to model SMART is only looking for a specific balance within each individual case.

    The only reason for specifying a projector brand is probably to account for the increments in the bias and gain controls even then if you use the luminance match technique as you are actually measuring the result as you adjust rather than relying on specific increment values to adjust by the end result is you just speed things up a bit rather than slavishly work through measurement runs and make changes in ever decreasing circles.

    There are probably some other parameters that depend on the manufacturer setting that boil to being familiar with the limitations o individual models. To stop smart running around in circles to get calibration outside the capabilities of certain units.

    This is similar to how you would calibrate with a full blown objective colour analyser. You adjust RGB bias on say a 20 IRE patch and RGB gain on a 70IRE patch and ping pong back and forth until the unit is as accurate as you can make it .

    SMART is trying to give you this capability without being a 5grand colour analyser.

    I'm wondering if you can even now use it to calibrate direct view devices.
     
  23. ROne

    ROne
    Active Member

    Joined:
    Dec 21, 2000
    Messages:
    779
    Products Owned:
    0
    Products Wanted:
    0
    Trophy Points:
    18
    Ratings:
    +7
    I thought the idea of the projector specific versions is that they were referenced against Steve-smallcombs SMART, which in itself was set-up with a colour analyser as per each projector. So you could reference (albeit a few steps away from the colour analyser), through the spread-sheet. Otherwise you have to assume one of your presets is close to D65 internally and obtain the best spread-of intensities from that.

    Having said all this SMART for me excels in trying to smooth the bumps out of the RGB readings irrespective of it's D65 based starting point.

    Good piece of kit for measuring gamma though, and keeping a check on what the iris of your projector is doing.
     
  24. Mr.D

    Mr.D
    Well-known Member

    Joined:
    Jul 14, 2000
    Messages:
    11,061
    Products Owned:
    0
    Products Wanted:
    0
    Trophy Points:
    133
    Ratings:
    +1,149
    Yes I believe they are but there has to be the option within smart to take into account variance in white point and individual projector's differences .

    Otherwise if every projector was the same all you would need to get is the RGB bias and gain values and the contrast and brightness to replicate the calibrated unit. SMART has to take the initial aims from the calibrated model and be able to apply offsets on an indivdual unit's capabilities.

    This is why you have to do the initial white point measurement , it lets smart know what its got to work with as a starting point. Its a bit like a feedback loop .
     

Share This Page

Loading...