Home › Forums › General Discussion › profiling restricted to Rec709??
- This topic has 12 replies, 3 voices, and was last updated 2 years, 9 months ago by S Simeonov.
-
AuthorPosts
-
2021-05-27 at 10:28 #30343
Hi,
there is an interesting phenomenon which I don’t clearly understand.
I did first a precal profiling of the display of Lenovo P17 (4k, DolbyVision, “color calibrated”). (attachement 1, Screenshot 101411.gif). The profile was measured with colourspace software. The surrounding borders are the AdobeRGB colourspace and it can be seen that it is nearly 100% covering precal.
Then I calibrated/profiled with DisplayCal. XYZLUT+Matrix. As I understood this process is just profiling this individual monitor and will not limit the profile ICC to any special colourspace. After profiling, the ICC profile was activated. See the results in the second attachement (Farbumfang..png) and the printout below.
All looks good so far and I profiles again postcal the display with colourspace again, same way as precal, just lesser points. The interesting result you can see in the third attachement:
As before, the white borders are the AdobeRGB colourspace which is nearly identical with the native one from the display. You can see that after calibration (using the same pattern generator and patches) all are now in a much smaller area within this native garmut and interestingly, the inner rectangle reflect extremely well the Rec709 colourspace (which was never selected and not wanted).
What I would have expected also in postcal is to see again all points within the AdobeRGB colourspace, just with less errors. But is it normal that display automatically “shortens” the native colourspace to be Rec709? Again, the colourspace software for pre and postcal are running on a different computer, all patches are exactly the same, nothing wrong here.
Or how do need to configure displaycal (see our profile below) to end up with a best calibrated/profile ICC for this monitor without restricting to a smaller garmut then the native one?
very best, thanks so much!
displaycal: profile:
Größe 1565784 Bytes (1529.09 KiB)
Bevorzugtes CMM 0x6172676C ‘argl’ ArgyllCMS
ICC-Version 2.2
Profilklasse Anzeigegeräteprofil
Farbmodell RGB
Bezugsfarbraum (PCS) XYZ
Erstellt 2021-05-27 01:15:04
Plattform Microsoft
Ist eingebettet Nein
Kann unabhängig verwendet werden Ja
Gerät
Hersteller 0x0000AE30 Lenovo Group Limited
Modell 0x00004125
Medien-Attribute Reflective, Glossy, Positive, Color
Standard-Farbübertragung Wahrnehmung
PCS-Leuchtmittel XYZ 96.42 100.00 82.49 (xy 0.3457 0.3585, CCT 5000K)
Ersteller 0x6172676C ‘argl’ ArgyllCMS
Prüfsumme 0x4270ED1AD888B2657DB56F03409F82C2
Prüfsumme OK Ja
Beschreibung (ASCII) NE173QUM-N42 #1 2021-05-26 23-58 0.3127x 0.329y S
XYZLUT+MTX
Urheberrecht No copyright. Created with DisplayCAL 3.8.9.3 and
ArgyllCMS 2.1.2
Gerätehersteller (ASCII) Lenovo Group Limited
Gerätemodell (ASCII) NE173QUM-N42
Helligkeit 347.06 cd/m²
Medien-Weißpunkt
Ist Leuchtmittel Ja
Leuchtmittel-Bezug XYZ 95.32 100.00 107.32 (xy 0.3150 0.3304)
Leuchtmittel-Bezug CCT 6372K
ΔE 2000 zu Tageslicht-Farbort 0.74
ΔE 2000 zu Schwarzkörper-Farbort 4.05
Medien-Schwarzpunkt
Leuchtmittel-Bezug XYZ 0.1083 0.1144 0.1221 (xy 0.3142 0.3319)
Leuchtmittel-Bezug CCT 6405K
ΔE 2000 zu Tageslicht-Farbort 1.05
ΔE 2000 zu Schwarzkörper-Farbort 5.49
Farbmittel (PCS-Bezug)
Rot XYZ 61.53 31.45 1.54 (xy 0.6510 0.3327)
Grün XYZ 20.32 62.95 6.44 (xy 0.2265 0.7017)
Blau XYZ 14.31 5.36 74.39 (xy 0.1521 0.0570)
Grafikkarten-Gammatabelle
Bittiefe 16
Kanäle 3
Anzahl Werte pro Kanal 256
Kanal 1 Gamma bei 50% Stimulus 1.00
Kanal 1 Minimum 0.0000%
Kanal 1 Maximum 100.00%
Kanal 1 eindeutige Tonwerte 256 @ 8 Bit
Kanal 1 ist linear Ja
Kanal 2 Gamma bei 50% Stimulus 1.00
Kanal 2 Minimum 0.0000%
Kanal 2 Maximum 100.00%
Kanal 2 eindeutige Tonwerte 256 @ 8 Bit
Kanal 2 ist linear Ja
Kanal 3 Gamma bei 50% Stimulus 1.00
Kanal 3 Minimum 0.0000%
Kanal 3 Maximum 100.00%
Kanal 3 eindeutige Tonwerte 256 @ 8 Bit
Kanal 3 ist linear Ja
Gerät zu PCS: Anpassungsart 0
Matrix 1.0000 0.0000 0.0000
0.0000 1.0000 0.0000
0.0000 0.0000 1.0000
Eingabe-Tabelle
Kanäle 3
Anzahl Werte pro Kanal 2048
Farbwerte-Tabelle
Raster-Schritte 33
Einträge 35937
Ausgabe-Tabelle
Kanäle 3
Anzahl Werte pro Kanal 2048
PCS zu Gerät: Anpassungsart 0
Matrix 3.9250 -1.2213 -0.6827
-1.9575 3.8322 0.0668
0.0574 -0.2813 2.6985
Eingabe-Tabelle
Kanäle 3
Anzahl Werte pro Kanal 4096
Farbwerte-Tabelle
Raster-Schritte 33
Einträge 35937
Ausgabe-Tabelle
Kanäle 3
Anzahl Werte pro Kanal 2048
Charakterisierungsdaten [331703 Bytes]
Gerätefarbwerte [331703 Bytes]
Farbmesswerte [331703 Bytes]
Absolute to media relative transform Bradford (ICC-Empfehlung)
Matrix 0.8951 0.2664 -0.1614
-0.7502 1.7135 0.0367
0.0389 -0.0685 1.0296
Roter Matrix-Farbwert
Leuchtmittel-Bezug XYZ 58.66 30.29 2.13 (xy 0.6441 0.3325)
PCS-Bezug XYZ 61.72 31.56 1.55 (xy 0.6509 0.3328)
Grüner Matrix-Farbwert
Leuchtmittel-Bezug XYZ 18.62 63.20 7.48 (xy 0.2085 0.7077)
PCS-Bezug XYZ 20.38 63.06 6.46 (xy 0.2267 0.7015)
Blauer Matrix-Farbwert
Leuchtmittel-Bezug XYZ 18.04 6.52 97.70 (xy 0.1476 0.0533)
PCS-Bezug XYZ 14.32 5.37 74.48 (xy 0.1520 0.0571)
Rote Farbtonwiedergabekurve
Anzahl Werte 256
Übertragungsfunktion ≈ Gamma 2.13 100% (Δ 3.50%)
Minimum Y 0.0000
Maximum Y 100.00
Grüne Farbtonwiedergabekurve
Anzahl Werte 256
Übertragungsfunktion ≈ Gamma 2.15 100% (Δ 3.68%)
Minimum Y 0.0000
Maximum Y 100.00
Blaue Farbtonwiedergabekurve
Anzahl Werte 256
Übertragungsfunktion Unbekannt
Minimum Y 0.0000
Maximum Y 100.00
PCS zu Gerät: Anpassungsart 1
Matrix 3.7668 -1.1229 -0.6171
-1.8605 3.6968 0.1176
0.1062 -0.2739 2.6324
Eingabe-Tabelle
Kanäle 3
Anzahl Werte pro Kanal 4096
Farbwerte-Tabelle
Raster-Schritte 33
Einträge 35937
Ausgabe-Tabelle
Kanäle 3
Anzahl Werte pro Kanal 256
PCS zu Gerät: Anpassungsart 2
Matrix 3.9250 -1.2213 -0.6827
-1.9575 3.8322 0.0668
0.0574 -0.2813 2.6985
Eingabe-Tabelle
Kanäle 3
Anzahl Werte pro Kanal 4096
Farbwerte-Tabelle
Raster-Schritte 33
Einträge 35937
Ausgabe-Tabelle
Kanäle 3
Anzahl Werte pro Kanal 2048
Gerät zu PCS: Anpassungsart 1
Matrix 1.0000 0.0000 0.0000
0.0000 1.0000 0.0000
0.0000 0.0000 1.0000
Eingabe-Tabelle
Kanäle 3
Anzahl Werte pro Kanal 2048
Farbwerte-Tabelle
Raster-Schritte 33
Einträge 35937
Ausgabe-Tabelle
Kanäle 3
Anzahl Werte pro Kanal 2048
Gerät zu PCS: Anpassungsart 2
Matrix 1.0000 0.0000 0.0000
0.0000 1.0000 0.0000
0.0000 0.0000 1.0000
Eingabe-Tabelle
Kanäle 3
Anzahl Werte pro Kanal 2048
Farbwerte-Tabelle
Raster-Schritte 33
Einträge 35937
Ausgabe-Tabelle
Kanäle 3
Anzahl Werte pro Kanal 2048
Chromatizität (Leuchtmittel-Bezug)
Kanal 1 (R) xy 0.6441 0.3325
Kanal 2 (G) xy 0.2085 0.7077
Kanal 3 (B) xy 0.1476 0.0533
Gerätehersteller und Modell
Hersteller 0x30AE Lenovo Group Limited
Modell 0x4125
Metadaten
prefix EDID_
CMF_
DATA_
MEASUREMENT_
OPENICC_
MAPPING_
ACCURACY_
GAMUT_
EDID_mnft LEN
EDID_mnft_id 12462
EDID_model_id 16677
EDID_date 2019-T0
EDID_red_x 0.6455078125
EDID_red_y 0.330078125
EDID_green_x 0.2158203125
EDID_green_y 0.7138671875
EDID_blue_x 0.1455078125
EDID_blue_y 0.0576171875
EDID_white_x 0.3349609375
EDID_white_y 0.3525390625
EDID_manufacturer Lenovo Group Limited
EDID_gamma 2.2
EDID_model NE173QUM-N42
EDID_md5 f88322788775b8d7767f170576580314
CMF_binary DisplayCAL
CMF_version 3.8.9.3
CMF_product DisplayCAL
Lizenz Public Domain
Quality high
OPENICC_automatic_generated 0
DATA_source calib
MEASUREMENT_device i1 displaypro, colormunki display
MAPPING_device_id xrandr-Lenovo Group Limited
ACCURACY_dE76_avg 0.318030
ACCURACY_dE76_max 1.668955
ACCURACY_dE76_rms 0.374568
GAMUT_volume 1.48852434878
GAMUT_coverage(dci-p3) 0.8576
GAMUT_coverage(srgb) 0.9962
GAMUT_coverage(adobe-rgb) 0.9654
Attachments:
You must be logged in to view attached files.Calibrite Display Pro HL on Amazon
Disclosure: As an Amazon Associate I earn from qualifying purchases.2021-05-27 at 10:53 #30351This is not related to DisplayCAL but to a 3rd party app (pic 1 & 3) that seems to be malfunctioning, or that you have misconfigured (simulate on teh fly some colorspace).
2021-05-27 at 11:18 #30355no unfortunately, it is not. the 3rd party app is just measuring the native profile and the triangles I just putted in for clarificaiton. What you can clearly see is that there is a huge difference in the garmut precal and postcal with displaycal. It might be random that this is close to Rec709 what is left, but the main question is still, if there is a way to “limit” this huge loss of garmut after calibration. And no, you can say many things about lightspace, but they measure accurately… Again, the “triangles for colourspaces” are just for demonstration, they have NONE influence on the measurements… and the measurements are taken the SAME WAY precal and postcal… thanks!
2021-05-27 at 11:28 #30357DisplayCAL is not doing it. IT CAN’T. It’s your 3rd party app fault or that you changed OSD mode in display to some gamut limitation or that you enabled AMD’s sRGB emulation of the fly from EDID data.
Easy to test, use command line spotread on a 255 green patch in MS Paint application.
2021-05-27 at 11:40 #30363The only thing which was changed in comparison to precal is that we loaded the displaycal generated ICC and activated it. Before for profiling with displaycal we used a null (or linear) icc which resulted in the large native garmut measurement. BTW, does displaycal disables the ICC automatically during profiling or also should load NULL icc?
2021-05-27 at 11:48 #30365The only thing which was changed in comparison to precal is that we loaded the displaycal generated ICC and activated it. Before for profiling with displaycal we used a null (or linear) icc which resulted in the large native garmut measurement.
No, it’s your app’s fault or user misconfiguration (simulation profile, using device link profiles… or such).
Run a spotread on a non color managed app. It’s the smoking gun. A display ICC profile does not change display colorspace, it just describes it. It can change grey & white (VCGT tag loaded into GPU)
BTW, does displaycal disables the ICC automatically during profiling or also should load NULL icc?
In OS? yes, no need to change it.
2021-05-27 at 12:09 #30368what do you expect to get from the spotread? Is it an app by itself?
if the VCCGT is set (what should be), even in non color manageed application the display changes?!
2021-05-27 at 12:46 #30370what do you expect to get from the spotread? Is it an app by itself?
Spotread just measures a patch: on a display (emissive), on paper, on a facbric (reflexive) and prints CIE color coordinates.
YOU provide the patch. Hence if display has a current colorspace ~AdobeRGB and MS paint is not color managed, 255 green is close to AdobeRGB green.
Spotreading such patch on screen should return that.It’s YOUR 3rd party app that is producing sRGB like RGB patches encoded in display RGB colorspace. Why? you change OSD, app misbehaviour, app misconfiguration, GPU colospace simulation… etc.
DisplayCAL measurement report can do the same, for example testing of your AdoberGB display behaves showing sRGB images on Ps. So YOU or 3rd party app are providing a set of modified patches encoded in display colorspace.
Your issue is not related at all with DisplayCAL, jus 3rd party app misbehaviour or user misconfiguration operating that 3rd party app.
if the VCCGT is set (what should be), even in non color manageed application the display changes?!
VCGT just changes grey. It’s a 1D LUT to make grey have the same color tone as white. Yes, it is applied “globally”, system wide… but just for grey.
2021-05-27 at 12:57 #30371ah ok understood. there is a similar function like spotread in lightspace, can compare this. However, the reason is for sure not in lightspace, but I can’t exclude that the app showing the patches might not properly recognize the ICC generated by displaycal. The patches are just shown in a video (we are using potplayer, which is configured to use ICC). Can you recommend a videoplayer for win10, which is known to properly use the full colormanagement and ICC?
2021-05-27 at 13:13 #30374ah ok understood. there is a similar function like spotread in lightspace, can compare this. However, the reason is for sure not in lightspace, but I can’t exclude that the app showing the patches might not properly recognize the ICC generated by displaycal. The patches are just shown in a video (we are using potplayer, which is configured to use ICC). Can you recommend a videoplayer for win10, which is known to properly use the full colormanagement and ICC?
Then it’s potplayer color management (madVR LUT3D? or ICC equivalent in potplayer) which reads video content colorspace (rec709?) and REENCODES RGB numbers to equivalent numbers in display colorspace (~AdobeRGB)
Application recognizes displaycal profile without issues and taht video player IT’S WORKING AS INTENDED: content colorspace (~rec709) reencoded to display colorspace.It was your fault after all. You are not testing it in the proper way. A color managed video player MUST BEHAVE like what you saw: show RGB values (in display colorspace) that are equivalent to “rec709 only” patches (if video is/is supposed to be rec709).
2021-05-27 at 13:44 #30377ok, got it. have to see how the patch video was encoded but you are probably right. Potplayer was ICC equivalent, not madVR LUT3d, because we wanted to directly test the performance of the ICC generated by displaycal).
For the precal, is no problem, we can use madVRTPG.
For the postcal, you are right, there is a problem. madTPG cannot be configured to use the ICC, right? probably 3dlut only, which then gives not a real idea about the performance of the ICC cal. No idea yet to show the patches in another way, cos u r right with the video…
2021-05-27 at 14:08 #30379For the postcal, you are right, there is a problem. madTPG cannot be configured to use the ICC, right? probably 3dlut only, which then gives not a real idea about the performance of the ICC cal. No idea yet to show the patches in another way, cos u r right with the video…
There is not ICC cal. ICC profiles are a description of a device.
If you make a LUT3D from rec709 (video) to dispaly ICC with displaycal LUT3D maker you are indirectly testing if Rec709 portion of while Display colrospace is accurate.. with soem rounding errors due to the whole chain.
2021-06-11 at 14:58 #30555I have an odd problem with using 3dlut in windowed overlay mode, the movies look kinda dark like crushed blacks or clipped blacks. This problem doesn’t occur in the other modes. The 3dlut has a gamma of 2.4 relative with black output offset set at 100%.
-
AuthorPosts