-
FSI – trustworthy or not? Maybe plasma better?
Hi all, I know this is a provocative heading. I mean no disrespect to anyone
But… On the Lightspace forum Steve Shaw, who designed Lightspace CMS, says that “he has never seen Flanders monitor with accurate color” and recommends Penta HD2line instead.
When I go to high-end suites, all I ever see is plasma and CRT. It’s the corporate and weeding guys who seem to have Flanders. Alex Von Hurkman uses a calibrated Vt30 and h’es doing this with a $30 000 BM panel, and he is also, I think few would argue, a very smart guy who knows color.
I also have never seen an actual technical report on the color accuracy of Flanders products. Just anecdotal “I love my monitor.” “Great service” etc. This at at some level doesn’t mean much. My aunt loves her TV, even though it has so much Red push all actors look like are blushing at all times. It certainly doesn’t necessarily mean they are actually accurate.
It seems most people buy Flanders because tehy don’t want to bother with calibration, so we have almost no feedback as to their actual measured, delivered accuracy.
I’ve also seen a few complaints here about yellowish highlights – and then a response that they are “perceptually different” from a plasma. I find this suspect., A perfectly calibrated plasma should look more or less the same as a Flanders if the Flanders is indeed perfectly calibrated. One has more contrast, OK, but highlights should look the same color, Otherwise one is wrong. People say “in that case trust the flanders” but based on what? The fact that they say they are accurate? How do we know this is true unless someone objectively measures them?
Sure they us amazing probe technology to calibrate, but what is the accuracy of the result? No matter how good your calibration gear, all monitors have limits. I can use $500 000 of equipment and hire the head calibrator from Company 3 and still not get good color from a monitor which is not physically up to the task.
I’ve seen a picture that showed all Flanders models together, fed the same 8 bit source, and the different models did not have the same color – to the point where I would make different grading decision based on each. This really, really worried me as they claim all are accurate – which necessarily could not be true from that picture (which was part of a very enthusiastic, pro-Flanders review).
Which makes me wonder….
All feedback welcome. I am here to learn, not to accuse.
-
59 Replies
-
Hi Robert,
The way you put it, I think you’ll provoke a few responses…
I color with a Barco DP2K-P DCI projector on a Stewart matte screen – I have it professionally calibrated more often than it needs (we shoot the colors, but it never drifts except after lamp changes). I consider it the “ground truth.”
When I work in Rec. 709 (which is often – and not for wedding videos), I output the same signal to a Flanders 2461w in the suite, just off to the side. We test it occasionally, and I send it back to Flanders for calibration occasionally. The Barco and Flanders, while they have inherent differences being different display technologies, basically match and reinforce each other (I’ll let somebody else tackle why two correctly calibrated displays can look different).
I can say that in two years with this set-up, I never have to make grading decisions based on one and not the other, they line up well. I’m not familiar with the HD2line, are they comparable in price? The high-end suites I’ve been discussing with people are springing for the Dolby PRM4200.
KC
-
This is great info! Thank you!
So the Flanders, right next to a know “perfect” source, looks the same?
-
Hi Robert,
What you are referencing is what is known as Metamerism Failure. This happens when we view different displays with different spectral distributions in close proximity to each other. This short video might help shed some light on the subject as well as provide a brief description of what you can expect from each of the FSI models in relation to their specific types of spectral distribution.
https://www.flandersscientific.com/index/tech_resources.html
As far as objective measurement go, FSI uses over $50,000 worth of industry standard equipment to align and measure each unit. We start with a Minolta CS2000 Spectroraditiometer to measure the panels native response curves and then create model specific offsets that are input into a Minolta CA310 Colorimeter to align each unit to industry standards. We then use a Photo research PR655 with a 5 nanometer bandwitdh option to measure the alignment. The link below will take you to the tech resources on the FSI website and give you the information in much greater detail.
https://www.flandersscientific.com/index/calibration.html
We do have quite a few customers who use the LM2461W in high end finishing suites, using Barco’s and Christie’s, that swear by that unit being as close to a perfect match to their projectors as they have seen with an LCD monitor.
You bring up good points and this is not a sales pitch, I’m just sharing information and I’m sure others here may have more valuable insight on their actual use of the FSI monitors in their suites.
“Think of everything in terms of building capacity.”
Kris Merkel
twitter: @kris_merkel
Product Manager, Flanders Scientific Inc.
http://www.shopfsi.com
Co-Founder, Atlanta Cutters Post Production User Group
http://www.atlantacutters.com2.2Ghz MBP core i7
16Gb RAM
CS 5.5
FCP7 and Studio
Blackmagic Design UltraStudio 3D
AJA IO XT
FSI LM-2340W
-
Hi Kris! I’m sure you use excellent methods for profiling your monitors. But what I would love to know (and am learning better here) is how accurate is the result, seeing the inherent limitations of any hardware technology to which even the most immaculate calibration is applied. Are we talking Delat-E of 1, 2, 3, 4, 0? You must have measured this too (the success of your careful calibration).
No display device can be perfectly calibrated by any method – it’s always a matter of degree. Sometimes that degree is beyond the limits of human vision, sometimes it is not and can lead to colorist being misled by a false representation of the image.
I would also love to understand better why, side by side, the various FSI monitors look truly different. Are the lower end monitors less precise? What is their level of accuracy vs. the flagship model?
I am learning a lot here. Thank you all for your contributions.
(And by the way, just because I see plasmas at the high end suites I happen to visit doesn’t mean I was implying that Flanders do not exists in other ones I have not happened to visit.
-
Robert, Delta E, with respect to greyscale, is typically less than about 1 as measured at 10 IRE increments throughout greyscale range when we do our QC before a unit ships, if it goes over 2 at any level we realign the unit. Keep in mind this is Delta E as measured by a secondary probe (other than the one that did the calibration as we use different calibration and QC probes). If we were to do the QC report with the same probe that did the calibration then delta E would be as close to 0 as makes no difference, but by using separate probes we have a fail safe to make sure the probes haven’t suddenly got out of calibration for some reason. Of course the probes don’t agree perfectly so a delta E tolerance of 2 is actually quite strict as that is as attributable to the probe variation as calibration accuracy.
The video that Kris linked to explains precisely why you can have two units with equal measurements that still look perceptually different, and different to different observers for that matter as well. This is NOT unique to FSI monitors and is true of any monitor brand that utilizes different types display technology (different spectral distributions). I think sometimes we are a bit honest to a fault here as we make it a point to divulge this information instead of sugar coating it. This is actually an important issue we try to stress with people in talking to them about their specific needs for a monitor and here are some important things you can take away from this:
1. A HUGE number of consumer TVs on the market, especially newer ones, are using a basic white LED backlight. These, even when calibrated to objectively (according to spectroradiometer) match a xenon bulb projector, CCFL backlight LCD monitor, CRT, or Plasma display tend to look warmer/redder to the majority, though not all, observers.
2. White LED backlight monitors tend to be particularly problematic when trying to get something that perceptually matches a high-end projector. Again, the numbers can be identical according to the measuring device, but most people perceive the white LED units as looking significantly warmer.
3. Wide Gamut CCFL, EEFL, and RGB LED backlight monitors tend to perceptually match high-end digital projectors for most observers. This is why, IMHO, you see so many manufacturers opt for solutions other than white LED backlight for their ‘premier’ color grading monitors.
4. The big conundrum in all of this is finding one monitor to rule them all. If you are doing high-end DI work then perhaps white LED backlight is not ideal. If you are doing mostly work destined for broadcast TV maybe you could argue that with the proliferation of white LED backlight LCD TVs that having a professional White LED backlight monitor is ideal, even for color correction. Though there are certainly a lot of other consumer technologies with significant market share. What we find in real world use is that when it comes to white LED people are starting to know and expect white LED LCD TVs to look perceptually redder.
5. Most importantly keep in mind that the differences are most dramatic when you make the decision to place displays of different spectral distributions next to one another. If you just color grade on one and then go from display to display individually, instead of side by side, content tends to overwhelmingly look normal as your eye quickly adapts to these perceptual white point differences.
6. If you are going to mix LCD with Plasma, CRT, or Projection my suggestion would be to use LCD with RGB LED, CCFL, or EEFL backlights. White LED is less ideal in such mixed use environments in my experience.
Again, I think the video Kris links to explains this well despite my southern drawl way of pronouncing Metamerism. At the end we even specifically tell you what to expect with respect to how our various models may match various types of other technologies.
Hope that helps.
Bram Desmet
FSI (Flanders Scientific, Inc.)
http://www.FlandersScientific.com -
Your video was great! Thank you for your answer as well!
II will think about all this. Probably smart would be to have both a well calibrated plasma and a Flanders so that one could see how both will look – at least, that is where I am leaning.
-
As I was quoted at the beginning of this thread by Robert I probably should add a comment or two…
First, it is true that we have been called in to calibrate a lot of displays all over the world, and we have come across FSI displays many times.
What we have seen is a lack of consistency, and a lack of accuracy time and again.
I’m sorry to be saying this – we have nothing against FSI specifically, and have seen the same problems with many other manufacturers – but that is just our findings – and that includes new displays direct from the factory and others that have been returned for re-calibration, as well as others that have been in use for a while.
We saw similar with Penta originally too.
This was not down to the probe, but the calibration system used by Penta to calibrate the display from the probe data.
Penta agreed with our findings and have changed their in-factory calibration to be much, much more accurate – they use LightSpace CMS now, with the same probes they were using originally – which prove my point above.
We have also seen very good consistency with the Sony LMD series – if a tad low on Gamut.
But, we do not recommend Plasmas at all for colour critical grading applications. The ABL (Auto Brightness Level) caused by plasma’s power saving operation (which can’t be turned off, regardless what some suppliers say!) causes major problems with colour accuracy.
See: https://www.lightillusion.com/forums/index.php?action=vthread&forum=8&topic=42
As just about all LED displays use screen from a small number of manufacturers calibration is down to the associated image management used, as well as the image processing electronics.
It is true that different back-lights do give different results, and some technologies to suffer Metamerism Failure, but this is more with different display technologies, not different back-lights. OLED vs. LCD for example.
But, with the correct probe offsets you can overcome all but the worse effects.
Often, you get far better results by turning off all the manufacturer’s internal processing, having the panel work in raw native mode, and perform all calibration via an external LUT box.
For HDMI connections the eeColor LUT Box is good.
For HDSDI the Tcube, Pluto or Davio boxes are needed.
Light Illusion is always happy to answer questions on calibration, etc, and we have some truly advanced colour scientists that perform our colour development work – I’m just one of the team!
Steve
-
I personally care less about that last 2% of color accuracy that I think we’re talking about here, and I think we should try to refocus towards the Display manufacturing industry, where every TV & computer monitor (and mobile display device) is a completely different image.
I’m dealing with more and more clients remotely, and post budgets are not getting bigger. I would just kill for a relatively inexpensive & simple radiospectrometer / calibration suite that could be in the hands of consumers. In a world where a teranex costs 2000$, I don’t understand why there isn’t a good prosumer level calibration tool I can give to a relatively tech savvy producer to make sure we are at least looking at similar things.
Joseph Mastantuono
http://www.goodpost.net
Color Grading & Post Production Consulting -
The problem is we are not talking about 2% errors in our experience.
And more of a problem is the procedure needed to get a good level of guaranteed calibration.
See: https://www.lightillusion.com/display_calibration.html
This is just to set-up a display BEFORE calibration.
We have actually run calibration over the net, using a virtual USB protocol, so only requiring a probe and a LUT box at the client’s location – or using SpaceMatch DCM for OS connected displays.
It works, but if the initial set-up is poor, the results will always be poor.
Steve
-
I also want to to add, the the Flanders 2461 is an incredible monitor. The commitment the guys at Flanders show to create the best product in the market is outstanding.
Having saying that, I’ve seen differences between the 2460 and 2461 on factory settings, the last one looks greener but truer to a Barco 2K, the 2460 way redder. For what I understand they changed the calibration equipment, but you can send it for recalibration.But also we had them side by side with a few Panny Plasmas, and what we found, and we don’t know of this is being a limitation of the Plasma technology and the Flanders being a 10bit panel, is that strong red tones, like the ones found in women dresses, nails or cartoons, tends to look more orange in the Flanders than in Plasmas, everything seems to match except when we enter the strong reds. One thing that got me baffled is that the red in the plasmas seems to be the right one as far as color tone, I mean, we al know how red nails should looks like, unfortunately I have no pics to show, but they do look more towards orange in the Flanders. I was wondering if someone in this forum also noticed something similar.
Reply to this Discussion! Login or Sign Up