Does it makes any sense? It might be easier for a rookie instead of stepping into a programming language...

Any suggestions?

Does it makes any sense? It might be easier for a rookie instead of stepping into a programming language...

Any suggestions?

We use image editors (like Photoshop) to make a lot of the patterns on AVS HD.

Producing the image in Photoshop is the easy part (technically), especially if you are good with Photoshop and have some patience. The hard part is getting that image onto a disc without throwing the levels off.

Hwjohn (does hw stands for "hardware"???),
I made some testing with HCFR DVD and AVCHD.... maybe you could help with your experience?

http://www.avsforum.com/avs-vb/showthread.php?p=13124154#post13124154

Hwjohn (does hw stands for "hardware"???),
I made some testing with HCFR DVD and AVCHD.... maybe you could help with your experience?

http://www.avsforum.com/avs-vb/showthread.php?p=13124154#post13124154

I responded to your other post. I think what you are asking for has already been done for you. There is at least one free SD cal disc in TomHuffmans CMS guide, then there is GetGray, Avia, etc. AVS HD is a free HD cal disc, and I think Avia and DVE will have HD discs out soon.

The only disc I can personally vouch for as being 100% correct (level wise) is AVS HD, mainly because I helped make it. But I would highly recommend GetGray for a full featured SD disc, and TomHuffmans disc for the basics. I like Avia and DVE for the tutorials alone, but I believe that GetGray (the person not the disc) found that one or the other had some inaccurate patterns... you might want to check on that.

The new (RC1) version of AVS HD should be available on Monday or Tuesday. This version has a lot of new patterns that make it much more full featured than the Beta versions. If it doesn't contain a pattern you would like to see, then we are always open to suggestions (we have already implemented a lot of forum user suggestions). If you want to take it a step further, you can make the pattern yourself in Photoshop and submit it to us for use on the disc. If we think it would fit in well with the disc and be useful to the majority of users, then we would include it in the next release. If you have questions about how to do it in Photoshop, I could answer those for you.

What I wanted to do is to submit a pattern to my satellite provider suitable for calibration purposes.

In a perfect world it would mean to ask the broadcaster to "sacrifice" 16 channels (6 for colors and 10 for grayscale), but due to higher costs, it will be impossible to obtain more than one channel for calibration purposes so I will have to submit a one-for-all pattern. (unless it would be easy for the broadcaster to send on the test channel a sequence of patterns.... who knows?)

This the current pattern:

http://img104.imagevenue.com/loc898/th_48452_Sky_Italy_Pattern_channel_997_122_898lo.JPG (http://img104.imagevenue.com/img.php?image=48452_Sky_Italy_Pattern_channel_997_122_898lo. JPG)

This is a draft:

http://img107.imagevenue.com/loc821/th_48460_Sky_Italy_Pattern_channel_997_proposal_122_821lo.JP G (http://img107.imagevenue.com/img.php?image=48460_Sky_Italy_Pattern_channel_997_proposal_1 22_821lo.JPG)

It misses a background among others details, but it could:
- allow primaries and secondaries measurements
- allow grayscale measurements

* The patterns would be big enough to place correctly the sensor
* The patterns would be mostly centered to minimize light border-falloff (although having no warranty about light uniformity)
* Being complitely different from windowed and fullscreen, it would need some (calculated) compensation for measurements (plasma power drain)
*.............

Any comments appreciated...

Hwjohn, I'm using Photoshop 7.... it has thousands options when setting color reference....

Can you help?

I need info about:
- image size depending on target (1920x1080, 720x576, etc)
- suitable image format (Targa?)
- bits per pixel or bits per channel
- pixels per inch
- mode: RGB color or CMYK
- color profile?

I've also just used paint for some quick patterns if needed.

Hwjohn, I'm using Photoshop 7.... it has thousands options when setting color reference....

Can you help?

I need info about:
- image size depending on target (1920x1080, 720x576, etc)
- suitable image format (Targa?)
- bits per pixel or bits per channel
- pixels per inch
- mode: RGB color or CMYK
- color profile?

You are starting to get into the difficult part. The most difficult part is getting your image from whatever format it is in (Targa, bmp, etc.) to a video stream without messing up the levels. I'll tell you how we do it, but it is a bit involved.

The original image is made as a 24 bit Targa (that is 8 bits per channel RGB). You always want to work in RGB and 8 bit per RGB channel. Our images are 1920x1080 because our disc is targeted at 1080p displays, but you could do others just as well. Keep in mind that your RGB levels need to be within the 16-235 range. 16,16,16 is black and 235,235,235 is white.

The next step we do is to use a special program written by dr1394 (Ron). This program is absolutely crucial to our process. It takes the 24 bit Targa image and converts it to the UYVY space. The output is a "raw" UYVY file. We still only have one frame.

The next step involves some Avisynth scripting and MPEG encoding. Avisynth is a frame server. There is an Avisynth plugin called "RawSource" that will allow Avisynth to read the raw UYVY file we created. I write a script to repeat the UYVY file(s) as needed. The script also sets the frame rate, as well converting to YUY2 space needed for most freeware MPEG encoders. The good news here is that our levels have been preserved to this point and the MPEG encoder doesn't have to do any further color space conversion.

The next step is to use a special build of HCenc (freeware MPEG encoder) to encode the clip. Avisynth reads my script and frameserves to HCenc. HCenc uses the settings I provide and encodes the clip to MPEG2.

The final result is an HD MPEG 2 file.

From here alluringreality takes the files and does the HD DVD authoring. He also uses Ulead DVD Factory to transcode the MPEG2 files for the BDMV and AVCHD versions.

There will probably be other issues you have to tackle. If you want to make your pattern 1080i for broadcast, you will have to deal with interlacing. The first thing you need to do is see what your broadcast provider would require you to submit. If it is MPEG4/AVC, then you are probably in trouble. If they can take some image format straight from Photoshop and do the work for you, then you are on easy street.

Keep in mind that the whole process isn't exactly trivial. Its not rocket science either, but it has taken alluringreality and I hundreds of hours of work between the two of us to produce the current (RC1) version of AVSHD.

Thnaks for explaining.

You are starting to get into the difficult part. The most difficult part is getting your image from whatever format it is in (Targa, bmp, etc.) to a video stream without messing up the levels. I'll tell you how we do it, but it is a bit involved.

I think I'll just slip into the pattern/bitmap creation and let the broadcaster do the video conversion.

The original image is made as a 24 bit Targa (that is 8 bits per channel RGB). You always want to work in RGB and 8 bit per RGB channel. Our images are 1920x1080 because our disc is targeted at 1080p displays, but you could do others just as well. Keep in mind that your RGB levels need to be within the 16-235 range. 16,16,16 is black and 235,235,235 is white.


Some questions:
- how do I ensure in Photoshop 7 the 16/235 values?
- should I assign a profile to the bitmap? If so what profile (SMPTE-C, sRGB, CIE RGB etc)?

I think I'll just slip into the pattern/bitmap creation and let the broadcaster do the video conversion.




Some questions:
- how do I ensure in Photoshop 7 the 16/235 values?

By hand. You make sure that all the values are 16-235 when you create them. Percentage gray boxes have to be calculated based on that range and drawn accordingly.


- should I assign a profile to the bitmap? If so what profile (SMPTE-C, sRGB, CIE RGB etc)?

This might be better answered by Bear5k or someone who knows a little more about monitor profiling, but in short I don't think it matters. The profile is basically to tell your specific computer monitor "how" to display that pattern. As far as the output is concerned, it shouldn't matter. If you are worried about it use sRGB. sRGB uses the same primaries as Rec. 701. The only time I can think this would matter is if the satellite company reads the file and it has the profile embedded as SMPTE-C or something other than sRGB and they try to compensate for that.

ok. creating a 0,0,0 pattern would mean BTB and 255,255,255 WTW?

Edit: And what about R,G,B,C,Y,K?

R->R=235,G=0,B=0
G->R=0,G=235,B=0
B->R=0,G=0,B=235

Correct?

Edit 2: some values must be rounded? Example: white=235,235,235 means gray_10%=23.5, 23.5, 23.5 .... Photoshop disallows this
and CYM???

ok. creating a 0,0,0 pattern would mean BTB and 255,255,255 WTW?

Yes.

Edit: And what about R,G,B,C,Y,K?

R->R=235,G=0,B=0
G->R=0,G=235,B=0
B->R=0,G=0,B=235

No. Full "red" would be 235,16,16, and so on.


Edit 2: some values must be rounded? Example: white=235,235,235 means gray_10%=23.5, 23.5, 23.5 .... Photoshop disallows this
and CYM???

Yes, you will have to round some values.

When this was brought up in the AVS HD 709 thread two comments were that the Average Picture Level (APL) of the existing pattern might be too high for plasma and that in measuring the colors the adjacent colors might be picked up:

1. I think maybe it's in http://www.avsforum.com/avs-vb/showthread.php?t=852467 that how to actually determine APL is discussed (either that or it's one of the other contrast threads). It's somewhat complex because every pixel has to be used to determine APL. Looking at the draft I'll guess that the APL isn't any lower than existing. This item doesn't exactly make sense to me because I figure color bars are still used with plasmas and I would guess they would have an APL around the existing pattern. In short, if APL is a real issue then I would certainly have a lot to learn to address that item.

2. If picking up adjacent colors would be a real problem on the existing pattern, then I'm not sure that the draft would address that. Again I can't throw out anything further, because I would have to do a lot of learning in that area myself to determine if it was a real issue.


The following are just some random thoughts:

- You don't necessarily need a full 10-point grayscale. Getting rid of that would allow you to lower the APL.

- Your pattern might not address the items that the broadcaster would want in their signal, so I question practicality. Asking them to run a video of different patterns might make more sense to me, but again not knowing anything about their intents or setup this seems somewhat incomprehensible to me as anything more than something to do simply if it interests you.

When this was brought up in the AVS HD 709 thread two comments were that the Average Picture Level (APL).......

Thanks for commenting :) I'll investigate later on

hwjohn: what about YCM?

Y->R=235,G=235,b=16 and so on?

BTW I'm getting strange colors.... blu looks violet.... 100% white, looks with a "point" of gray in it.... I'll show later the bitmap

Thanks for commenting :) I'll investigate later on

hwjohn: what about YCM?

Y->R=235,G=235,b=16 and so on?

BTW I'm getting strange colors.... blu looks violet.... 100% white, looks with a "point" of gray in it.... I'll show later the bitmap

Yes, that should be correct.

The colors may look weird on your monitor of your monitor is off.

When this was brought up in the AVS HD 709 thread two comments were that the Average Picture Level (APL) of the existing pattern might be too high for plasma and that in measuring the colors the adjacent colors might be picked up:

1. I think maybe it's in http://www.avsforum.com/avs-vb/showthread.php?t=852467 that how to actually determine APL is discussed (either that or it's one of the other contrast threads). It's somewhat complex because every pixel has to be used to determine APL. Looking at the draft I'll guess that the APL isn't any lower than existing. This item doesn't exactly make sense to me because I figure color bars are still used with plasmas and I would guess they would have an APL around the existing pattern. In short, if APL is a real issue then I would certainly have a lot to learn to address that item.

2. If picking up adjacent colors would be a real problem on the existing pattern, then I'm not sure that the draft would address that. Again I can't throw out anything further, because I would have to do a lot of learning in that area myself to determine if it was a real issue.


The following are just some random thoughts:

- You don't necessarily need a full 10-point grayscale. Getting rid of that would allow you to lower the APL.

- Your pattern might not address the items that the broadcaster would want in their signal, so I question practicality. Asking them to run a video of different patterns might make more sense to me, but again not knowing anything about their intents or setup this seems somewhat incomprehensible to me as anything more than something to do simply if it interests you.

I haven't read through that thread, but I always figured that APL was calculated as an area-weighted average. Say you had three gray levels, one at 16, one at 128, one at 235. You would do something like take the area of each gray "bar" as a percentage of the total screen, and multiply that percentage times the value of that bar, then sum up all the values from the three bars to get an "average" picture level. So say the bars where 20, 30, and 50 percent of the total screen area, respectively.

APL = (.2)(16) + (.3)(128) + (.5)(235) = 159

I know that probably isn't the right equation, but it is a lot more complicated than something like that?

Wow! the same colours now in acdsee look perfect! The blue is blue.....70% to 100% looks too gray...

what do you think? Do you like it?

http://img221.imageshack.us/img221/1944/testcopyif0.th.jpg (http://img221.imageshack.us/img221/1944/testcopyif0.jpg)


Flame On!

P.S. - I think it could be suitable for a broadcaster, but dunno really how to cope the APL factor....

I know that probably isn't the right equation, but it is a lot more complicated than something like that?

That's the general idea, but to put it into context there might be some level above which plasma acts differently according to some of the comments on the forum. The average picture level listed for movies in general in that linked thread is somewhat low. I don't remember exactly, but it's something like maybe full-screen 20% gray. I would guess the last image would average to around maybe 50% and might be a higher APL than the original.