Free calibration soft (non-excel) - Page 12

Sure it is valid, my point was that it is not practical...

Also thousands of measure and users feedback highlight this is not problem in real life... At least we have other areas to work on first.

--Patrice

Most other users are not using it for CRT front projectors. These are a special category because of the three guns and low light output. It is also very simple to hold a magnifying glass in front of the sensor while sampling the light. This will not yield accurate light level readings, but that is not an issue since reading lux is not the key factor.

There is nothing for you to work on. This idea can be used by anyone who faces this problem and uses a sensor with a diffuser.

I agree and sorry, I wasn't clear.
What we work on is some improvements to our probe...

--Patrice

No problem.

Remember that 0% stimulus is not really 0 light output, or at least it doesn't have to be. this might be wrong, but what I did/do in my calibrations is to set 0% to 0.1 - 0.2 cd/m^2. that's the limit of the instrument and at least for crt's seems to provide a good black point.

As to gamma, I made a post awhile back (you can search for my user name to find it. I'm too lazy to do that right now... but I included a link to a good reference) in the free rader spreadsheet thread regarding this issue, based on a couple of technical articles I found written by a group of international tv guru's. Recall that you're trying to fit a slope with a power law. If one ignores constant offsets, you'll get flakey results. I'm not privy to the details of CalMAN (and if I did it's not my place to share them ;-), but what made sense to me was to numerically differentiate the Y versus %stimulus data, then fit that using the ansatz Y = a I^gamma, for gamma the coefficient of interest and a is a constant of proportionality that we don't care about. I tested this approach and it agrees pretty well with CalMAN and disagrees with the free sheet. From a set of measurements from 0 - 100%, you get a set of gamma which can be plotted, averaged, least squares fit, etc.

the CalMAN site is found here: http://www.calman.tv/
click the Forums link to go to the CalMAN forums.

I believe that CalMAN is accurate when using the dtp-94 sensor. This is the sensor I have. They (Derek and Bill) worked *very* hard making sure it was accurate.

Related to noise in the sensor, the dtp-94 has a command that computes amplifier offsets. this step is crucial to being able to compensate for sensor noise at low light levels. If you don't do that, then you get flakey low light measurements.

As to a winter project (building the DIY sensor), an english version of the single player plans would be very helpful. i've been through the exercise of using a foreign language technical webpage to do technical work and it can be problematic. Maybe Joe can help since he's been through the build process. I'm especially interested in the programming of the processor chip...

Best and sorry for the long post,
jeff

This will work with the Spyder2 Pro?

Yes, it is the same probe. You'll need to copy your cvspyder.dll file to the HCFR directory though.

matt -

page 10 has the lastest firmware for download, the above link you have is the original release

Jeff, here is a good source for electronic components I use for my DIY projects Digi Key

I may just build one as a learning experience in filtered based meters. I could add it to my growing collection of hardware, which also now includes a DTP94

Thanks for the link Derek ;-)

Do you also have a spyder probe? Done any comparisons with the dtp-94 or others that you care to share?

Best,
jeff

I would like to see more comparisons as well. I am trying to decide if I should purchase C----n for my Spydertv probe.

Rick

Greeno, building the probe looks more daunting than it actually is. Although I did(as we do) make life harder with respect to which design to start with.

Key points if building a HCFR designed probe.

The 2 major hurdles.

1. the pcb, there is 2 choices, a single sided and double sided pcb layout. They produce the same result, but the double sided uses less space.
The main issue with using the double sided version is that you really need solder down the holes for components make curcuit contact under the component.
Alas I didn't have solder down the holes so I had to raise some parts to create the curcuit.
So in hindsight the single sided version would be a better choice for the average joe.

The pcb's themselves, the website downloads have all the schematics that you need. I looked in the phone book for curcuit board makers, then emailed them the gerbal files from the website. I asked for 2 pcb prototypes, and ended up with 3. They made a spare in case and the total cost was about $25US. Having spares is very useful btw.

2. Next hurdle the microchip. Basically I ended up buying a PICKit2 microchip programmer(which is produced by the same Co as the microchip, must also be updated to the latest software) and made a small adaptor to connect to the correct pins for programming.
You don't have to programme per say, just load the Hex file already provided by the HCFR team.

You can make a interface that connects to your pc and use free software, which I looked at, but since I am very amature at electronics I decided to go the the slight more expensive sure thing route.

The key problem is that I found is that most of the electronic Co's in my city didn't actually have an interface programmer tool for this particular microchip and for such a 1 off situation couldn't be bothered making or purchasing anything.

I actually left chips with 2 different Co's and both failed once they realized what type of chip it is. I spent more on petrol and got pissed off with the running around that proved to be fruitless.

So if you can find someone who can programme the microchip great!, PIC18f2550 , which is a 28pin chip.

Most ppl have gear for 16pin and 20pin chips. The 28pin have different pin arrangements(so I found).

So once I bought my programmer kit and built an adaptor the actual programming took 3 minutes from the time I fired up the programme I had never used before. I had to confirm the programme serveral times because I couldn't believe it was so easy after all the running around I did. sigh.

The major tricky bit when making the probe is the sensors themselves which are very small surface mounted RBG sensors. DO NOT get them too hot.

You need a fine point soldering tip for this job, anything large will be like trying to pick a eyelash out of ya eye with a lamppost.

I mounted one of my probes in a metal case rather than the plastic one on the website. I used PEM standoffs too mount the pcb and the metal case allowed me to fit PEMS to mount the probe on a stand, like a camera tripod(usually 1/4 fittings).

Other advice.
Get enough parts to make atleast 2 probes. They are mostly very cheap and the freight costs are just as much.

I am going to get some more bits myslef to build another, but this time I am going to mount the sensors on pointers so I can get the sensors closer to the screen without causing too much shadow, a problem with any probe and postioning. In fact I'm going to get a number of the sensors so I can experiment with optical and parabola collectors.

With that in mind, a Question to Laric. I plan to use ethernet cabling and plugs and fit the sensors on leads( because there is 8 wires and is shielded and twisted pairs). Will there be a limit of length of wire I can use that the sensors be away from the microchip. With respect to signal loss. I was thinking around 1~1.5 metres.

Regards
Nigel

ps, to those with spyder2 Pro, I have borrowed one and for the life of me can't get it working with HFCR. It doesn't use the same dll files, that I can find on my pc.
I am not to bothered, on the HCFR website forum there was some tests done on various probes. They all have pros and cons, but I feel the trick is to get used to what ever you are using as it is more the skill of the user than the quality of the equipment. In fact I would hazard a guess that any novice at calibration(me being one) getting within 10% accuracy with any of the probes and software is doing well.
The inaccuracies of the software or probes will be far less than the average ability of the novice.
Anything is better than nothing though and for me the idea of building probes and doing this myself helps me learn and is a great hobby aswell.

I can feel the wife frowning.....

That's nifty. What will this software do? Is it a more powerful tool than the software that comes from the factory? User friendly?

Dererk, I'd also be interested in your thoughts on DTP-94 vs. Spyder2, as relates to low stimulus accuracy and also suitability for various display technologies, since we know the Spyder2 is not well suited for DLP's and such....

Matt -

Its super easy..Just make sure you take the .DLL file from the Spyder folder, copy it over to the HCFR folder where the .exe file is..Open the program up..click on the NEW icon, select if you are going to use your DVD or other (I use AVIA so I select DVD Manual), then select your probe.

Connect your probe, if it hasnt been connected and set it up, selected how many IRE you are going to measure, by default you should see 10, which will start you a 0IRE and 10IRE increments until you hit 100IRE..you can changed that, then I select xyz, to see my coordinates, in the table where the IRE is at, click on the camera to start your reading, it will start as mentioned at 0IRE, then it will tell you to go to 10IRE, etc...when done, your figures will pop up, then you can use the see your greyscale, your RGB, gamma etc..

You misread my point. If I help a competitor get better, then I am most likely making a mistake. If I help a competitor with no profit motive get better, then I'm just being silly. HCFR is a "for profit" home cinema site. This software would be, in marketing terms, a "value added service" something is an addition to the value they offer people (and pretty good advertising given the splashes on the graphics). If they drive more traffic to their site, then their revenues go up and their support costs do not since no support is implied.

For us, our model is the inverse. People getting the software increases our revenues, and if the ask questions in our forum, then our support costs increase. Because we offer support, you see. In other words, this really isn't higher math, and it's the last I'll say on the matter. We are looking into getting our app certified, though. Which costs money...


Search is your friend. You have 4 links in your inbox at our site, and you are free to post a thread on our site asking for clarification on anything not covered to your satisfaction.

Bill

Yes I do have several S2's, along with many other meters and types of meters.

I will on the CalMAN forums once I get all the hardware needed to perform a accurate comparison. Those include a light and temperature controlled environment and the ability to accurately measure those.

I'm sorry but this is totaly wrong !!

HCFR web site is totally non profit ! Self financed by donation.
There is no adverts on site and no advert revenu at all !
(Although there was some adverts testing earlier this year, they were finally not in sync with our spirit and removed, BTW, that was way before software goes out, last it was only visible to un-registered users, registered never seen it).

HCFR organisation is a non profit one, registered and validated as so per France gov. agency.

If you can read french, look at our "change" file and you'll see the reasons to put the splashes on the graphics, it is not to promote site, but to avoid profit usage

We do this on a total non profit basis, and given the thousands hours we spend on this, I take this as an offence if you think there is !

Neither Georges nor I did tell anybody to post on the HCFR Forum, we answer here and will continue to do so, here and elsewhere... You don't have to go to "our" forums...

Last, another revision will be available soon, lots of improvements on GUI and some new probe support. (Isn't that promotional ? )

--Patrice

There seems to be some jealousy here..and I hope this does not get out of control. It seems like HCFR has to defending itself but maybe that is a good thing..its like they say BAD PUBLICITY IS GOOD PUBLICITY...well at least in Hollywood :0)

I personally think HCFR has something going for itself..and for the very least on my end, have not profited from anything from me . If anything, many of us in some way can profit from using HCFR..but that is my opinion..

thanks HCFR for such a wonderful and simple to understand software

btw..I downloaded your DVD patterns, and it seems in line with AVIA (by comparison), and so much easier to use since it is directly patterned to your software...I am assuming correctly that we can use this also and although its PAL, the patterns are the same as NTSC???

Patrice - No offense intended! I have to admit that my French is very rusty, so it seems I mistook your 15 Euro requested donation as a mandatory one. I expect that you guys have spent as much time and money on your site and application, if not more, as we have on CalMAN. Neither of us is likely to get rich making tools for other DIYers to use.

Lest I stick my foot further up my mouth and offend a good group of folks who have put out a nice app and what seems like a decent DIY meter, I'll bow out completely now. If you have any additional questions on CalMAN, you know where to find us!

Bill

That's fine Bill, no offence taken.

richlo, yes patterns will be the same, an NTSC version should be available soon, we wait to have DVD v1.15 validated then will do NTSC version.

--Patrice

Back to topic.

I did some small experiments to try and boost light input from afar too do 2 things, so as to raise the low end light levels above noise and so I can move the sensors further away so I can avoid the sensor case casting a shadow on the screen. With a slightly boosted signal overall possibly will help readings in general.

I tried using some small toy binoculars. These as predicted by some reduced the signal.

Next I tried a lens from an old slide projector. This had potential, but has the issue of being sensitive to tuning to a focal length. Too hard for us in the long run.

Next a neat bit of DIY. Cheap and effective. A light funnel.

Working on the principal that light bounces off surfaces at acute angles, ie greater than 15deg I found a piece of 45mm round clear plastic bar 90mm long. I then lathed it into a cone shape with an angle of 15deg down to 8mm at one end. The final shape looks like an icecream cone. I polished the whole finnished piece as to aid reflectiveness internally.

How it works. The large end area is the collector and the small end is the funnel which is placed in front of the sensor. Basically its about ratio, the collector has a 31:1 ratio in surface area. This is about light traveling in straight lines off a surface, the most powerful for us here. This collected light is funneled to the small end area increasing the light intensity.
The funnel doesn't need focal tunning at all, only direction, but we have to do that anyway.

I ran a proof test on my PC screen with the sensor placed 300mm from the screen.
All lights were off in the dark.
Interestingly I used the device to try a calibration and found it easier to get results, dispite the extra range of 300mm from screen, whereas a sensor on the screen left me off target at the bottom end.

Anyway the proof of the exercise.
In the attached file check the "measures" run I did.
The first half of measurements is without the funnel, then it spiked as I placed it upon the unit, then notice the luminance values go from 0.47 to 0.65 and since this value is not linear the result has potential of my aim of raising the low end response above noise so one can set black point brightness and get low end balance figures.

I was pleased to see the other figures, delta E, temp not change although there was a slight change in red. Maybe responding better.

I suspect by using larger round material and making a larger collector you could gather more light. Dishing the collector end may help collect wide angle light coming from wider afield, without the need to go wider material.

All in all a cheap and effective result and worthy of further tuning. The only thing I don't want to do is get to sensor saturation and haven't really looked into that as of yet.


btw Laric, With the V1.1.5 DVD I did find a small error in the 00IRE part of the "contraste". It strangely worked in v1.1.4 now doesn't in v1.1.5.
I just used the grey scale 00IRE instead.

If it was on auto from probe it would crash.

 

Opticrun001.zip 1.71875k . file

don't forget that any material you use between the sensor and the source will, in addition to aiding as a "collector" here, also affect the color balance to some extent depending on the index at different wavelengths (colors)....

To the designers of the DIY sensor:
What about including an "amplifier offset" calculation much like the dtp-94 has to account for amplifier noise?

to those in the know of the internals of the spyder2 SDK (and hopefully you can answer without violating a NDA):
Does the SDK allow for a similar calculation of amplifier offset?

If not, then this might be a key to improving low light level performance, no?

Best,
jeff

What sensors are now supported, DIY, Spyder 2, and DTP-94? Is that correct? If so any chance the Display 2 will be added anytime soon?

Thanks!!

Youa re correct on the 3 sensors..I have not heard anyothers they will suport

Hello,


Of course it is possible, but as I said earlier we are working on some improvements and then it is likely you'll stuff like this one...


I pass it to the person (David) doing the DVD, he says nothing change in that area and did not fully understand the point, would you mind given more details ?

--Patrice

when i'm measuring gray scale with "DVD manual" as generator, the popup window tells me exactly what DVD chapter i have to use.
unfortunately this is not true for any other color measurement like saturation, primaries or secondaries.

is it possible to enhance this in the next version?

Hi Waho

Works perfect for me..never had a problem with that. It doesnt matter what generator you have, just make sure to hit the camera in the areas for greyscale, primary, secondary, and I get the popup screen for the pattern they request

Laric, when manually selecting the "contraste" black 00IRE value the DVD would freeze and stop.

Starting from the "contraste" white 100IRE allowed forwarding through the series.

I got around the issue by running the 00IRE from the grey scale series then returning to contraste from white.

DVD player is a Panasonic HT520
DVD created on a LG burner using imgburn software.

Relatively minor issue, I would only attend too if/when making other changes.

Regards
Nigel