-
Posts
6,692 -
Joined
-
Last visited
Posts posted by NotMyFault
-
-
3 hours ago, thomaso said:
Kannst du denn jeden der Punkte bei dir lokal reproduzieren?
Ich hab das Problem gar nicht selbst! Ein notwendiger trigger ist wohl Apple Display (MacBook), ich habe einen MacMini mit LG Display und ganz anderen Farbprofilen (Spider Kalibriert oder DCI-P3 Standard).
-
-
7 hours ago, thomaso said:
Ich habe vor Monaten die Erfahrung gemacht, dass ich den Mac neustarten muss, um ein bestimmtes Profil für Affinity unzugänglich zu machen. (als wäre es bei Benutzung in einen Serif-eigenen temp od. cache Ordner kopiert worden)
Der verlinkte Issue-Thread erwähnt nur /Library/…, doch außerdem können auch in System/Library/ColorSync sowie in ~/Library/ColorSync noch weitere Profile liegen. Das Profil-Handling der Apple-eigenen Profile (Apple-Hardware) finde ich verwirrend. Der auch hier im Thread im Nov. 23 erwähnte Unterschied in den Varianten zur Profilbenennung (Display" <–> "Farb-LCD", plus weitere Sprachversionen) macht es für mich auch etwas unklar, welches Profil Affinity definitiv verwendet, wenn es ein Bild öffnet, das ein Profil enthält, das mit selbem Namen "Display" bereits auf meinem Mac existiert. Evtl. greift Affinity bei Namensgleichheit auf bereits geladene Profile zurück, zB aus "Performance-Gründen".
Ich sehe bei deinen zwei Screenshots in Vorschau (Preview.app) keinen Unterschied – aber einen (geringen) in Affinity: Dort erscheint das Gelb deines Screenshots mit deinem Apple-Profil etwas grünlicher während das Bild mit sRGB etwas wärmer erscheint (wie beide Bilder in Preview).
Nicht nur die Varianten von Profil-Benennungen sondern auch die Erstellungs-/Änderungs-Daten der Apple-Profile sind irritierend: Es scheint, dass sie jeweils das Erstellungs-Datum vom Zeitpunkt des Screenshots bekommen, also dem Moment des Einbettens des Profils. Das macht es schwieriger, ein "Original" zu identifizieren. Bei anderen Profilen, zB sRGB oder mein mit X-Rite erzeugtes Monitorprofil, bleibt das Ursprungsdatum des Profils auch im Screenshot erhalten. – Seltsam ist auch, dass die Apple-Profil-Dateien im Library/ColorSync/Profile-Ordner system-eigene Namen (lange Ziffern-Ketten) sowie geänderte Erstellungs- und Änderungsdaten haben, obwohl ich sie nicht benutze. In diesem Ordner liegen auch Profile für meinen externen Eizo-Monitor, die mit X-Rite erzeugt aber von mir anders benannt wurden. Deren "Originale" scheinen dagegen in ~/Library/… zu liegen, mit den Namen und Daten wie ich sie erwarte. Als würden sie durch macOS gelegentlich vom User-Library-Ordner in den /Library-Ordner kopiert werden.
Ich werde also momentan nicht schlau aus dem Apple-Profil-Handling. Interessant bleibt aber, dass das Problem bei dir in Affinity "vorher" nicht auftrat und es offenbar möglich ist, die korrekte (oder erwartete) Farbdarstellung in Affinity zu erhalten. – Da ich mit meinen Screenshots + individuellen Profilen keine Farbabweichung in Affinity sehe, ist für dich vielleicht auch eine Chance, wenn du ein eigenes Profil erzeugst, dass dann eine mögliche Namensverwirrung mit "Display" in Affinity vermeidet. (wie es ja für dich auch mit dem sRGB-Profile nicht auftritt). Wenn du keine Kalibrationsshardware hast klappt es eventuell, mit der macOS-internen Software ein Profil zu speichern, das dann nicht automatisch "Display" heißt.
Das wäre es wert in die FAQ Liste zu kommen.
Falls die Moderatoren nicht darauf ansprechen: Tutorial, damit man es bei neuen fällen schnell als Referenz findet.
-
Another workaround:
- use transform tool to temporary rotate layer to intended degree (center anchor point)
- apply gradient fill by gradient fill tools, using constraints modifyer key
- Reset layer rotation to 0
-
Windows preview app does not support color profiles correctly. The file uses GREY/8. if you want to see it correctly with those app, export with sRGB profile.
-
-
9 hours ago, AcidArrow said:
Affinity apparently destroys the color data immediately upon loading.
No. There are only „potentially lost“ if you don’t use the method described above, and export the file.
9 hours ago, AcidArrow said:This is a bug, it needs fixing.
Affinity stated consistently it does work as designed. This thread is a feature request, and it will become more likely to getting implemented if more users add their vote and provide arguments how this would improve their workflows and why workarounds are not sufficient.
-
16 minutes ago, AcidArrow said:
I can open a tga with alpha in affinity, do nothing, save, and the color channels will have changed drastically.
You need to recover RGB channels by the method described here:
only 2 steps, can be recorded as macro and assigned a keyboard shortcut.
-
Can you please provide a screen recording, full window covered including the layer panel visible?
I cannot reproduce -
-
13 minutes ago, loukash said:
I think the confusion arises from the object name: It should be called Symbol Group, akin to the Warp Group.
I would object in this case.
a symbol allows only one single child object when creating it (otherwise multiple symbols will be created, one per selected layer), so I mostly group layers first to be able to create a symbol. It makes no sense to call it symbol group if it is not grouping at all.
-
-
Seems solved in V2.3
thank you Serif team.
-
Both statements are true:
- by default the UI shows „none“ as gradient type
- you can simply start dragging a linear gradient without choosing a type before. The type is set automatically to linear.
- walt.farrell and ecureuil
-
1
-
1
-
3 hours ago, Tiberius Kraemer said:
It would be much more important if the Affinity Publisher program could save files as INDD and the Affinity Designer, Illustrator files as ai. That would really be a leap forward.
In addition, we would finally need an Affinity PDF reader/editor that is structured exactly like Adobe Acrobat so that PDF data can be checked or modified for print/digital printing.
This is totally off topic to this thread and should be moved to feature requests / feedback
-
Similar to vector layers and other layer types who show blend mode passtrough.
There are more bugs related to symbols mixed up with groups when used with constraint groups. Symbols have special abilities which normal groups don’t have.
Why are only symbols misinterpreted, but vector layers not?
-
Thank you for the workaround, never the less this should be fixed.
-
2 hours ago, firstdefence said:
Just tried this and while it does store the adjustment layer; shows as a red square and is named as the adjustment filter, when dragged back onto the workspace, the adjustment layer appears and the settings are registered in the filter but the adjustment isn't applied, but, if you go to Blend Options and nudge either of the ranges in or out to 99% the adjustment is applied and obviously only 99% of it is applied and anything below that is a percentage overlay of the filter. Duplicating the asset'ed filter maintains this anomaly.
Grouping multiple adjustment filters to asset as a collective adjustment, when dragged back doesn't apply any of the filters, the Blend Options trick doesn't work so you have to ungroup and then apply the Blend Options trick.
Makes me think this might be an obscure bug.
Can confirm this bug on iPad.
instead of blend range, you may use gamma slider. Effect vanishes for default gamma 2.2 but becomes effective for any other value.
please keep in mind that sliders for numeric have another bug. For many you can mircro-adjust the value to a bit higher / lower, but the displayed value remains unchanged. So you often have 3 numeric values all giving the same UI value, e.g. color or opacity sliders.
-
1 hour ago, 凌剑君 said:
It is not work,I would like it works like Photoshop
Works for me on iPad. The command uses letter O (shift-cmd-O), not digit zero. But all shortcuts depend on language settings and may do not work if non-English keyboards are used. Happens to me with German keyboard, especially when modifier keys are used for letters which require a modifier key to reach in my non-en keyboard layout like []
-
4 hours ago, Ldina said:
For whatever reason, I had to add an extra Parameter (labeled "h"). It doesn't DO anything, but if I remove it, the formula doesn't work and I get the same result as you did.
Thank you so much for sharing. That is insane. The input variable h is not even used in the formula, and normally h is a reserved name (giving the height of the document) that is not even allowed to be used as name according to Affinity doc (like x and y).
good as workaround, but even more confirms this is a bug.
and yeah, this is one of my major complaints: for several functions from PT filter, the number and/or type of parameters, or the suitable value ranges are not (correctly or fully) documented. It make a hell of difference if you need to use 0-1 (like bounded color values), small integers, positive / negative values etc. this must be documented.
-
Selection from Layer, Shift-cmd-O should work.
-
I spent countless hours to work with PT filter, apply image, equations filter etc. Really nice in principle, but so many shortcomings, bugs, lack of documentation. Even after specifically requesting documentation the answer was "this is a standard library we use" without any further details (which library, where to find more documentation).
I searched more and found loose relation to OpenGL and other GPU APIs.
-
None of the cellnoise functions work in V2.3 on Mac
https://affinity.help/photo2/English.lproj/pages/Filters/filter_proceduraltexture.html
Voronoi Noise cellnoise cellnoise(S x, S y)
cellnoise(S x, S y, S spread)cellnoise2 cellnoise2(S x, S y, I degree) cellnoisedist cellnoisedist(S x, S y)
cellnoisedist(S x, S y, S spread)cellnoiseedge cellnoiseedge(S x, S y, S sz, S softness) Tried to reproduce cellnoisedge from the tutorial:
No matter what parameters you specify, the output of these PT functions is always 0.
-
On 1/13/2024 at 6:23 PM, Ldina said:
I'm not sure what I am doing wrong, but I can't get the Voronoi functions (i.e., cellnoise, cellnoise2, cellnoisedist, cellnoiseedge) to do anything at all.
Some of the functions are simply broken and do not work at all, or documentation is simply wrong about number / type of parameters.
despite the procedural texture is listed as filter, it is actually less powerful and more an adjustment.
For Vornoi functionality, you would need a blur functionality: calculate an average color over multiple x/y positions. This is not possible in PT filter. You can only access the actual x/y position but no color values from neighboring pixels.
PS: just raised this one.

Farbabweichung beim Import
in Affinity on Desktop Questions (macOS and Windows)
Posted
Genauer: gelöst ist (nur) das Farben auf Gruppen angewendet für Vector/pixel Layer unterschiedliche Ergebnisse liefern.
Adjustmenst in CMYK ist eher grundsätzlicher Natur, bzw. Unzureichende Dokumentation und Anleitung.