So, if I well understand, all of us have to review our customized patchsets depending on our patch generator capabilities/settings!

Let me bold it, as that implication was not at all clear to me at first reading!

ebr9999
As I have no idea what your patch sets are, and what signal range you use with your displays, I have no idea, so I can't answer that question.

Steve

I have been using 0-100 RGB patches and full RGB output on my patch generators (RPG and Murideo). By setting levels to 16-235 I could generate video levels.

Then, if you read the page linked to, nothing changes.
Except now the 16-235 range has no rounding/scaling errors.
(I assume you mean your .csv patches are 0-255 range?)

Steve

Assuming my patchset is a regular greyscale, what it is below 16 and above 235 is now clipped.

ebr9999
Sorry, I don't understand the question...
Edit to say: I see you have updated your post - and yes, that is EXACTLY as it should be.

Steve

It is not a question but a complaint and a warning to all beta users: With previous betas, customer patchsets were scaled to the requested interval. I understand introducing rounding errors.
Now your patchset should match the interval (so rounding errors are eliminated at the origin). That implies that patchsets might be modified accordingly.

No - please read what has been written, and the explanation provided on the website.
You appear to be rather confused.

Steve

Sorry for insisting, but
1) I have patchset with levels between 0 and 255
2) Up to now I have set my hardware (Patch generator) to full RGB
3) Up to now I have set patch scale to 16-235 in cs/settings for generating video level.
Now the question is: are levels sent to the hardware still the same in this update (excluding rounding errors)?
Previously 0 was scaled to 16 and 255 to 235, as the whole patchset. Now it is written in the manual (and confirmed by who has pre-tested this release) that values below 16 and above 235 are clipped in such a case. How can you tell that nothing has changed (apart from rounding errors)? E.g. what's about RGB 13 in my patchset intended to be about 5% and now being clipped?

ebr9999
I guess the best is for you to simply try...?

As per the provided information, the set Patch Scale values will be the range the triplet values will be limited to.
No scaling required/implemented.

13,13,13 is not a valid value for TV Legal (when you set Patch scale to 16-235), so will correctly be clipped to 16.
Exactly as it should be.

Steve

ebr9999

With your configuration, with TPG set to RGB Full, Patch Scale 16-235, TV set to expect Video Levels (Black Level Low).........

Previously when using a custom patch sequence CSV list, for example, a 20 point greyscale, you would create and load the sequence in full range (0-255) and the software would scale accordingly.

0 0 0
13 13 13
26 26 26
38 38 38
51 51 51
64 64 64
77 77 77
89 89 89
102 102 102
115 115 115
128 128 128
140 140 140
153 153 153
166 166 166
178 178 178
191 191 191
204 204 204
217 217 217
229 229 229
242 242 242
255 255 255

But with this, the xyY targets would be incorrect due to rounding errors.

Now you still set the same settings in software, but any custom sequence you create for import should be created in the matching levels. So the above sequence becomes

16 16 16
27 27 27
38 38 38
49 49 49
60 60 60
71 71 71
82 82 82
92 92 92
104 104 104
115 115 115
126 126 126
136 136 136
147 147 147
159 159 159
169 169 169
180 180 180
191 191 191
202 202 202
213 213 213
224 224 224
235 235 235

With this, all target xyY values will be correct as there are no longer rounding errors.

Rescaling any previously created custom csv sequence is very simple and is just a case of applying the correct Full Range -> Limited Range formula.

The same principle applies to 16-255 sequences (for LG LUT calibration), and all of this only applies to custom created patch sequences. For the software internal (built in) sequences, there is no need to do anything as the software set the correct triplets accordingly based on the software settings.

If the original 0-255 patch set were used with Patch Scale set to 16-235, patches 0,0,0, and 13,13,13 would generate a 16,16,16 patch, with just the second patch being 'saved' into the profile (as it will overwrite the first 16,16,16 patch).
And patches 242,242,242 and 255,255,255 would generate 235,235,235, again with just the last patch being saved into the profile.
All other patches would be exactly as defined.

Steve

Leon, many thanks for expanding the whole sequence and showing what has to be changed.
I was quite confused by Steve saying:

Implication: Due to my HW settings I have to convert my patchset.

Oh - wait!

Are you trying to perform Manual Calibration, using set % patch values to match a TV's CMS control points?
If so, that is the missing information from your posts.
And the correct answer will depend on your TV, and how it maps its CMS controls differently if you set the TV to Full or Limited range.
But that all depends on your TV and how its CMS works.

Steve

Steve
Hi Steve

Yes, I am aware of that. My point was really regarding the fact that if a specific patch sequence was previously created for specified points (such as 5% incremental steps in a greyscale sweep), the software would previously scale the output based on the Patch Scale setting. You could very easily see this in action by hovering your mouse over the calibration patch window, and it worked exactly as intended.

The difference is now that when you set Patch Scale 16-235 but use 0-255 sequence, as you say, 0 & 13 will be clipped to 16 with only the second reading being recorded, and all other patches will be as specified in the 0-255 sequence.

But these are no longer the same 5% incremental steps, as ColourSpace no longer scales the patches.

This, in my opinion, is a good thing. It means that now all targets are correctly displayed based on the user defined settings in software (no more rounding errors).

The thing that is confusing some people is the fact that when taking this approach, it is best to create any custom sequence relative to the settings you apply in software. So when using Patch Scale 16-235, then create the sequence in 16-235.

For those that know how, it is very, very easy to rescale any previously created sequence to the levels you require. I have already done this for users of the LG OLED LUT calibration guides I created.