AVS Forum banner

SpyderTV Review

193K views 932 replies 166 participants last post by  glenndelaro 
#1 ·
With all of the discussion surrounding the proper set-up of HTPCs, and the confusion surrounding PC versus Video levels, Datacolor, under its Colorvision brand, has finally released its SpyderTV product that was first shown at CES in January (2005). The goal of this product is to automate the set-up of Home Theater components using the Colorvision Spyder2 sensor. As a beta tester, I felt fortunate to be allowed to have early access to the product and to some of the people who were bringing the product to market.

HOW DOES IT WORK?

The SpyderTV package comes with the Spyder2 sensor, a universal mount for both tripods and display faces, a CD with the software and a DVD with test patterns on it. Once installed on a PC that can be put relatively near the display to be calibrated, the software walks the user through setting up the display. First comes color temperature, then contrast, then brightness and finally the various color controls, should your unit have them available (e.g., most displays being fed digitally disable many of their color controls like saturation and hue).



The process goes by relatively quickly (expect about a half hour or so), and the results can be merely good to dramatic (my CRT was set pretty closely to the final values the SpyderTV had me set, my LCD was nowhere close). The results of using the SpyderTV software as it is currently shipping with my H77 DLP front projector was not sufficient to overcome the poor factory calibration on this unit. More work is definitely needed on the H77 than just adjusting the simple, user-available controls, and Colorvision indicates that actual support for FP will be released later.


At the end of the calibration, the SpyderTV will print a report (see attached) that gives you details of what changed and how your display improved.

NET-NET

Is it better than DVE or Avia? It is certainly faster in my experience. The DVD does also include additional test patterns to set Contrast and Brightness visually before starting the calibration procedure, so you definitely do not need either of those discs if you are going to be SpyderTV-centric. However, the price differential between DVE ($20), AVIA (~$40) and the SpyderTV (MSRP: $269, about $240 on the street) is significant.


So, who is this for? People who want a rock-solid basic set-up, but who do not want to spend the time and hassle dealing with DVE (bad menu structures) and Avia (color errors in some patterns, no BTB/WTW).


The real question you are wanting answered, though, is will it do grayscale? After all, at $250, this is competitive with having an ISF tech come calibrate your display. Well, hang on for part 2 for that answer...

 

SpyderTV Report - Sony.pdf 279.3671875k . file

 

SpyderTV Report.pdf 134.6982421875k . file
 

Attachments

See less See more
3
#2 ·
Well, early on in my testing of the Spyder2, I had a sneaking suspicion that it was the same sensor that I had gotten with my Spyder2 (plain) package. Since I also had a copy of OptiCAL (now discontinued, see FleaBay for resellers), I decided to test it out. The answer: sure I could get xyY data out of OptiCAL, but the results did not inspire as much confidence as I had hoped. I asked myself the question: are the results valid? are they consistent? what do I DO with the data, anyway?


The answers came with a lot of nosing around, a lot of research and a lot of luck. The research was easy, for several months, I had immersed myself in the works of Charles Poynton and Keith Jack to try to make sense of the data. Being something of a professional spreadsheet monkey, I had begun crafting a model that would give me usable results from the data that OptiCAL was providing. However, I then joined the beta program for the SpyderTV, and discovered one night how to get xyY data out of it. Unfortunately, this method would break with the shipping release, but an AVS user (Lee Bailey) came to my rescue. He told me how to enable the xyY data that my now broken interface no longer supplied:

SUCCESS!

A new shortcut was needed with a command line switch:
Code:
Code:
"C:\\Program Files\\SpyderTV\\spydertv.exe" /support
I found 9 seconds gave me better results, versus the default 4 seconds, especially for low-light measurements.




With the included tripod mount, using my SpyderTV with my DLP front projector is now easy to do:




BUT, WAIT, WHAT ABOUT ACCURACY?

Well, that is what has taken the longest. Thanks to Jeff Meier (avs userid: UMR), I have software that will work with the EyeOne Pro, the instrument that is the reference for field work.


Over two different nights, on two different weekends (did I mention this has been a heck of a lot of work??), I ran the EyeOne and both of my Spyders through their paces. My Spyder2 software has now been upgraded to the PRO, and I tested each unit with the PRO software the first go round. I would have also had a set of comparisons, but by the time I got to the EyeOne, it was four o'clock in the morning, and operator error kicked in.


Some kind advice from Jeff later, and the next weekend I was hitting on all cylinders. Of course, the ambient light problem in my space means that I had to start at 10pm, but hey, this is sort of like work, right?

The results you can see in the attached spreadsheet . Using both the PRO and SpyderTV apps and sensors, I had very consistent resultswhen used on the same day. There is some disagreement between the Spyder2 and the EyeOne at lower light levels, but as you can see, there was divergence with the EyeOne on the two measurement passes I took below 40% stimulation (by the way, the second pass was much more believable to me given the Y values returned, and it matches the Spyder down to 30% stim pretty well on x and y coordinates).


So, now who is this thing for? The answer, anyone with $300 to spend who wants a solid grayscale, that may or may not have a bit of error in the lower reaches, and who wants to get to 6504K. All that you need is a good spreadsheet model and some patience (a dark night helps, too, in environments like mine).


Red really is the limiting color on my H77:



Too bad that it is also oversaturated, and yellow does have a strong greenish cast to it:



This is where the rubber meets the road. I definitely need to play with the RGB cuts and gains:



6504k? I don't think so!



But wait, you ask, what about the PRO? Should I get the PRO or the Spyder TV? For that, we should turn our attention to part 3 of the review...

 

Colorimeter Comparison.zip 12.7578125k . file
 

Attachments

#3 ·
A key question for users will be which model to get and how to determine which sensor you actually have. My answers to these two questions are summarized below:


Purchase the SpyderTV if:
  • You have displays (e.g., TV, projector) that are not connected to a PC and that you do want a more sophisticated calibration than what DVE/Avia provide.
  • You are looking to buy both the SpyderTV and the PRO for both PC and stand-alone displays.
  • You want to purchase the lowest cost solution for mixed-use, and you have a good quality software DVD player on your PC.
  • You are budget constrained and need the cheapest comprehensive solution available (full calibration, and do not want to also buy DVE/Avia).


Purchase the Spyder2PRO v2.0 if:
  • Your HT is predominantly PC-based, and it is not important to you to calibrate the non-PC displays.
  • You have a front projection based HT using a digital technology (e.g., DLP), and you are more comfortable with formal support for FP in the currently released product.
  • You are looking to buy both the SpyderTV and the PRO for both PC and stand-alone displays.
  • You want to purchase only one product for mixed-use, and you feel the profiling capability of the PRO is worth the price premium.
  • You need the Spyder2PRO for a PC environment, and your HT is a secondary consideration.


As for how a definitive method of how to tell which hardware you have, that is still an open question (as of 9/16/05). My retail SpyderTV that came straight from Colorvision showed as having software version 3CL and hardware version 3CL. I did not have software version 4CL with my sensor. I have asked the folks at Colorvision for clarification on this, and the answer was that software version 4 does not indicate a different sensor. Thus, there is no definitive way of determining which sensor you have if you do not have the shipping full STV or PRO v2 packages.


I am hoping that the upgrade comes back and is easy to purchase. My bet is that the upgrade was pulled because it would validate against the STV serial number, and that was an undesirable attribute. There may have been another issue, and I have not confirmed this hypothesis with Colorvision yet, but it seems plausible to me.

TRAINING THE SPYDER

It has been noted that the results of my initial colorimeter comparison showed unacceptable (read: noticeable) color deviations in the Spyder2 with my Optoma H77 DLP front projector. Having been in contact with Colorvision, I received a new, retail packaged sensor. Thus, I set about clearing my living room for an evening to re-run my colorimeter comparison.


However, since the refugees whom we sheltered wanted to cook us dinner (I had not been home for two weeks), I could not start measuring until after we had all had a good meal and watched Shrek 2 on the big screen. As a result, I did not get started until 11pm, so I only had time to run two sets of measurements for both the STV and the EyeOne.


The results are fairly interesting. For my DLP FP, the Spyder2 had fairly decent repeatability and accuracy. With the new STV sensor, repeatability was dead-on for my test. For all measurements above 0% stim (~0.3 Lux), x and y measurements rarely deviated by as much as 0.001! When I ran the same comparison for the EyeOne, .001 and .002 deviations from the average were the norm. Lacking a dedicated, NIST traceable light meter, I cannot validate the Y measurement, but given the issues I have had with the EyeOne at low light levels, I do suspect this measurement more than I do the xy data. Of course, given that x and y are transforms of XYZ, my suspecting Y does not necessarily make sense from a color science standpoint. It's really just a gut feel, that indicates someone with better gear than I have will need to give this a similar go (Colorvision, theoretically here soon...).


Given the high repeatability of the STV hardware, I decided to "train" my measurements against the results I received from the EyeOne. To do this effectively, one should have a reasonable set of measurement data for both the training device and the reference device. I would consider three points a minimum for this, personally, but I went ahead and used two to illustrate how to do it.


The results can be seen at the bottom of the attached spreadsheet. I used the measurement data from 10% stimulus through 100% and also the primaries and secondaries. This gave me 16 data points, which is a decent population on which to run a regression to give me both a measure ment offset (regression intercept, or alpha) and a scaling factor (regression coefficient or beta). More measurement points would be helpful for those trying to get real accuracy, though the t-stats on the slope (regression) coefficient are both highly significant.


Corrected data can then be generated by plugging the "x" data into the regression model for the "x" data set (don't get confused by Excel'slabeling here!!) and the "y" data into the regression model for the "y" data set. Those wanting to try this at home should use the SpyderTV as the "independent variable, and the reference instrument's data as the dependent variable. There should be a separate model for each of x, y and Y.


One important note that goes without saying, but just in case: if you are going to train your sensor, you need to have access to a reliable reference and do it independently for each sensor. One should not use someone else's training data on an untested sensor. If it was that easy, the factory would have included that information in the firmware itself.


Later,

Bill

 

Colorimeter Comparison2.zip 25.8037109375k . file
 

Attachments

#4 ·
Looking forward to part three -- though if I am reading the above entries right, though the SpyderTV may be able to work for doing grayscale on a FP system, the software included is not adequate? (One assumes that one could do grayscale calibration of a flat panel display too/even easier BUT that again the software is not adequate?)
 
#5 ·
RE: Optical


Do you feel that doing grey-scale with xyY data with The TV or Pro software is that much better than using Optical or did you switch from Optical to Pro just because you could ??


I have a Spyder and a Spyder2 and Optical but NOT the new PRO....

 
#6 ·
Nathan - I did have to hack my own spreadsheet model to generate usable info from the xyY data.


Jim - I will try to get some numbers from OptiCAL to go along with the ones from the PRO and the TV. However, since I do not like ICC/ICM profiles for digital inputs (e.g., rounding issues being most prominent), OptiCAL provides the same level of functionality in this respect. The major question is whether it returns the same numbers. The TV, and presumably the next iteration of the PRO, are based on the ColorFacts codebase. I do not believe the same is true for the last release of OptiCAL.


However, since I have all three...



Later,

Bill
 
#8 ·
To further clarify:
  • The upgrade from TV to PRO will be for a fee.
  • The upgrade for the FP functionality for the TV will be free, but not soon. However, my results seem to indicate the TV does pretty well right now.
  • The major distinction between the functionalities appears to be how the automation is done.

If you can settle for a manual calibration, then either solution seems to be working well right now.
 
#10 ·

Quote:
Originally Posted by Ursa /forum/post/0


To further clarify:
  • The upgrade from TV to PRO will be for a fee.
  • The upgrade for the FP functionality for the TV will be free, but not soon.




  • I am feeling oh so



    When you state "the upgrade from TV to PRO will be for a fee"; what exactly is being upgraded?


    It is my understanding they would release a software upgrade in October (which allows FP calibration) for FREE if you own the PRO and for a nominal FEE if you own the TV.
 
#12 ·
Mark - You've got it. Just like one can purchase an upgrade to go from the base Spyder2 package to the PRO package, there will be a commercial upgrade between the PRO and TV. I do not know what "nominal" means in this respect ($50? $100?).


Scott - Colorvision is also planning on building the same wizard-type functionality for FP as they have for CRT, RPTV and Plasma and putting it into the SpyderTV package. This has not yet been done, and given the environmental factors involved, it may not be done for a while. It also does not help that the developer for the SpyderTV is also the developer for ColorFacts, with a new version of ColorFacts scheduled to ship this fall/winter.


The upgrade for a fee is for SpyderTV owners to get the PRO software. If you want to wait for the wizard, then it will be free - but it may be 2006 before it is available.


JimP - The above data was taken from my Optoma H77 DLP front projector. Unfortunately, I do not own either a plasma nor a rear projection TV, but if someone is willing to buy me one, I will take a boatload of measurements to satisfy people's curiosity!



The LCD in question is a five year old Viewsonic that is rated at 400:1 CR. It by no means looks as good as my dedicated old analog TV. However, it did look much better (just not in direct sunlight!).


Later,

Bill
 
#14 ·
Hardvark - One thing to note, I am not exactly using the shipping PRO software in these comparisons. That being said, you guys are going to completely obviate part 3 of my review with all of these questions about which one to buy. Right now, I would give the nod to the TV simply because it can do plasmas and you get the tripod mount. If you can wait for v2.0 of the PRO software (October), then there is a bit more total functionality there with the ability to create profiles. Basically, the net recommendation is if you are HTPC-based, go for the PRO, and if you are not, the TV is probably for you (unless you want to profile other PC-based displays in your house).


Clear as mud?


As for using Jeff's software, I prefer my own for obvious reasons (e.g., it works the way I want it to, it is laid out the way I like to see it). Mine is also a bit cheaper (quite a bit cheaper if Jeff eliminates his introductory pricing) than Jeff's since I am really only providing the math in an easy-to-use format (did I mention my interface broke from the beta to the release?). I do still have a couple of open issues to check out (e.g., PAL/SECAM gamma), so I won't be releasing mine until after Labor Day. However, this is all I'd really like to say on this subject since I do not want to turn this into a sales thread. Hit me on PM or at bill 'AT' datapopuli.com if you want to discuss this.
 
#16 ·
Bill,


All the references here are to the Spyder2Pro, yet there are also cheaper Spyder2 and Spyder2Plus packages that only differ (AFAIK) in their software packages. Is there something special in the Pro package (similar to OptiCAL in v1) that provides the needed xyY values, that isn't in the cheaper packages?


Alternately, could a Spyder2 sensor be used with my OptiCAL software (from the Spyder 1 Pro package)?


Lastly, what did folks pay when having their v1 Spyder "refurbished" to a v2 Spyder sensor (Milori version, since apparently their are two different v2's)?


I'm glad that ColorVision has done nothing to make this confusing to potential customers.



- Tim
 
#17 ·
"Alternately, could a Spyder2 sensor be used with my OptiCAL software (from the Spyder 1 Pro package)?"


I've been waiting for the answer to this one (see above).... I haven't even opened the box yet on my Spyder2, hedging on an upgrade if required....


"Milori version, since apparently their are two different v2's"


HUH ??????????????????

 
#18 ·

Quote:
Originally Posted by Ursa /forum/post/0


Basically, the net recommendation is if you are HTPC-based, go for the PRO, and if you are not, the TV is probably for you (unless you want to profile other PC-based displays in your house).


Clear as mud?

I may have pulled the trigger too early on The Spyder2 Pro, but I was concerned about the ability to get xyY data with the Spyder TV. So I contacted Colorvision to ask if there is an "upgrade" path from the Pro to the TV and this is the response I received:


"Unfortunately, the method in which the colorimeter (Spyder) is

calibrated are different, a new product is required. To show our support

to existing loyal users, we are offering a 25% discount off the purchase

of your home theater product. For this offer to apply you must live

within the USA and purchase directly from ColorVision."


So, although they may be the same sensor, the calibration process seems to make all of the difference and it appears from Ursa's test that the TV may be better at lower light readings. Bummer.
 
#19 ·
Alright, now that we are all back together...



Tim - My answer to your above question was this:
Quote:
Originally Posted by ursa /forum/post/0


Neither the Base model nor the Plus will give you xyY data, AFAIK. I will hopefully be testing OptiCAL this weekend against the PRO and TV software, so I can tell you then what I get. One important point of note is that the TV is designed not to have the light pipe/grate/filter attachment be removed - even for CRTs. This speaks to there being a bit of math that may be going on in the background that OptiCAL does not have. However, I will post my results when I get them.


I think getting a Spyder 1 refurbished at this point may no longer be possible. Heck, there are some Spyder2s out there that are not compatible with the TV software (the Kevin's point about there once being two different Spyder2s - though that is no longer the case).

Jim - If you bought the Spyder2 relatively recently, odds are that you have the latest sensor. I, too, originally bought the base Spyder and a copy of OptiCAL as a cheaper (
 
#20 ·
Here is the definitive word from the product manager from Colorvision:
Quote:
SpyderTV hardware is universal; it works with ColorFacts and the Spyder products in our digital imaging line. Spyder2 colorimeter that is included in Spyder packages prior to SpyderTV release do not work with SpyderTV software. By the end of September, all the Spyders will be universal. ColorFacts Professional 6.0 that we will start shipping in September will include the universal Spyder. So, you are right in your argument, there is no such a thing as Milori Spyder2. Spyder2 hardware will work with all of our software products. Software packages will be protected.


We will offer upgrade packages to our customers. So, for example, STV customers will be able to upgrade to Spyder2 Pro by just buying the software. And, we will offer loyalty programs to existing Spyder2 pro customers so they can upgrade to STV. Again, ColorFacts Professional 6.0 is a free upgrade to all ColorFacts Professional customers. As I mentioned, we will publish meter comparison studies on our website so everybody can see how Spyder works against other meters. This is going to be very valuable information. We have full confidence in our Spyder2 hardware.

A few things from my own experience:
  • My Spyder2 seemed to work okay with the SpyderTV software, but then I have only just recently begun working with shipping software.
  • I had originally thought there would be a direct upgrade from the PRO to the TV, but apparently one has to buy the software and sensor package again. Seems like a better deal to get the TV and then upgrade to the PRO.


Hopefully this helps. Please keep in mind that I believe that I am pushing the boundaries of where Colorvision is going with this product, so if some of the answers seem confusing, it is probably because we are outside of their marketing plan.


Later,

Bill
 
#22 ·

Quote:
Spyder2 colorimeter that is included in Spyder packages prior to SpyderTV release do not work with SpyderTV software.

Thanks for the update. I suppose then the question becomes, at what point did the Spyder2PRO line become universal - for existing Spyder2 users that is? Meaning, what hardware and software version of the sensor corresponds to the sensor included in the SpyderTV?


Also, if one were to purchase the Spyder2PRO package today, I gather they would be able to sidegrade to SpyderTV or upgrade to Colorfacts 6.0 in the future. Again, I think the confusing part is that you have two different sensors included with the same product - the Spyder2PRO package.


Mark
 
#23 ·
Just so I am clear....


If I purchase the SpyderTV today I will receive the latest hardware sensor and will be able to calibrate every device known to mortal man except for a front projector. However, Bill's testing will show us that we can get "darn close" by doing some manual math and such. Colorvision will offer a FREE software upgrade "sometime in the not too distant future" that will allow the SpyderTV to calibrate a front project via a wizard interface.


If all the above is true, is there a reason why someone should purchase the SpyderTV AND also purchase the Spyder2PRO software "upgrade"? What does the Spyder2PRO software offer that the SpyderTV software doesn't?



Thanks!
 
#24 ·

Quote:
What does the Spyder2PRO software offer that the SpyderTV software doesn't?

The Spyder2PRO Studio offers the ability to calibrate and generate ICC profiles for you computer monitors among other things.


Ther package also includes:

Quote:
nik Color Efex Pro 2.0 Standard Edition, PANTONE® colorist, and ColorVision® DoctorPRO to enhance, edit, correct, and specify colors on your images, screens, and printers.

Mark
 
#26 ·
Well, you can to a certain extent. Its just that the software isn't designed for it. Using the Spyder2PRO software is something of an automated process. It analyzes your CRT or LCD display and generates an ICC profile for you. I don't belive the SpyderTV software will generate an ICC profile. Spyder2PRO also has a colorimeter and other tools that can be used to obtain xyY data, but it doesn't include tools to model this data per say.


Bottom line, if you have an accurate way of modeling the data obtained from Spyder2PRO/SpyderTV readings, you can (with certain caveats) calibrate a display for your home theater (but you won't be able to generate an ICC profile based on this data for you computer monitor). The software just doesn't make this an easy process. Along those lines, the SpyderTV doesn't do grayscale out of the box. From what I understand, you would need to enable the SpyderTV to generate xyY data. In this case, you are back to obtaining additional software to model the data. Colorvision's Colorfacts (originally Milori) will model the data obtained from the sensor.


I think the confusion comes from the notion that there are now essentially two Spyder2 sensors. The sensor included with the SpyderTV and newer version of Spyder2PRO and the sensor included with older version of the Spyder2PRO. That and consumers are looking for a "one box" solution to do both tasks. I have a need for ICC profiles for my work on the computer, but I am also interested in calibrating the display in my home theater. It seems silly to need two *slightly* different sensors in this case.


Mark
 
This is an older thread, you may not receive a response, and could be reviving an old thread. Please consider creating a new thread.
Top