What's your workaround for transform presets?

I'm curious what folks do to work around the lack of transform presets. Below is a specific use case, but this can apply anywhere that Adjustment Clips don't work -- or simply aren't desired.

By "transform presets", what I mean is a way to store Transform/Crop/Composite/etc. settings, that you can quickly apply to any clip(s). This could be a thing you drag and drop (like a fusion effect), something you save to a keyboard shortcut (like a Color Page Memory), whatever… simply a way to store and easily apply settings.

Adjustment Clips don't always work because they act in a window on top of the original clip. If my underlying clip is a different size and aspect ratio than the timeline (i.e. 6K 4:3 open gate footage on a 16:9 UHD timeline), and I've filled the frame with the original shot (so there's clipped media outside the frame), an Adjustment Clip that zooms out or pans down will not reveal the clipped media -- it only acts on what's already visible.

Visually you can work around this by setting the scaling to Fit so the entire 4:3 clip is visible (with pillarboxes), and use Adjustment Clips to do ALL the scaling (so literally every shot has to have an adjustment clip over it), however then Resolve is scaling up the already scaled down footage -- you are not looking at "restored" original footage. I find this wholly unacceptable, and a massive shortcoming to Adjustment Clips. (EDIT: I may be wrong about it not restoring the source… I've tested multiple ways with multiple results. Curious what the collective says about this. Basically, if I scale a clip on the timeline to 50% then scale it with an Adjustment Clip to 200%, restoring it to an original appearance of 100%, am I looking at original 1:1 pixels, or pixels that have been scaled down and back up… different tests I've run show conflicting results)

All I'm left with is copy/pasting adjustments from one clip to another, which is really tedious, especially when you have to scroll around the timeline to find the last time you used a particular framing that you want to reuse. I've read workarounds of storing the Transform settings to an Adjustment Clip which you then add to a Power Bin, or even to another video clip in a Power Bin, but you still can't copy the settings from the clip in the Power Bin without first adding it to your timeline… so you have to add to the timeline, copy, paste settings, delete the dragged clip. Ugh.

So the original question… what do you do to work around this limitation? I know this has been a feature request for years… maybe worth bringing up again.

I'm curious what folks do to work around the lack of transform presets. Below is a specific use case, but this can apply anywhere that Adjustment Clips don't work -- or simply aren't desired.

If I need such a thing, I built it in fusion. If I need to apply it to other clips, I use fusion reference composition. Sometimes a macro. But generally reference compositions offer more dynamic approach since unlike a macro its easier to change. But macro can be used in fusion and reference composition in the edit page, for easier set up and lot of flexibility.

If I need to apply it to other clips, I use fusion reference composition.

Googles "fusion reference composition"…

looking into it now, thanks!

fusion reference composition

WOW. This is an amazing feature I'd totally missed. Here's a great tutorial for anyone else looking. Thanks man, this is a super helpful tip!

You bring up a good talking point, but it's important to understand there are really only 2 ways for Resolve to scale native pixels, without filtering and getting soft - (higher than timeline resolution) pixels that I'm aware of.

1 is the edit inspector zoom control.
1 is fusion single clip comps - when you have a clip on the timeline and click the fusion page (or reference with the bottom layer pulling in the highest resolution)

All other ways, including fusion zoom presets, adjustment clips, and compound clips, all depend on the timeline resolution and raster to that moment.

Hey

As your question was kind of asked as a poll, I either saved a .comp (which here is a behavior I should change* in favor of fusion reference comp), or - and this is very likely - I make a macro reachable in the edit page (meaning I save a .setting in the Edit folder of the templates)

*I was aware fusion referenced comp as soon as it was released but realized I really prefer keeping everything for a long time, which supposes I also put theses fusion referenceable compositions in the power bin... which I don't want to.

As a conclusion, I hardly use them just because I won't put them in a power bin, but in the meantime I'd like to use them later in another project

Sam just a heads up - if you make a .setting macro - you are limited to the timeline resolution.

This is fine if you have a 4k timeline, and 4k footage - it has to do some filtering to blow up an image.

However if you have a fancy new 6k camera on a 4k timeline - you will want to continue to use your comp file with XF transforms to blow it up. Or use the edit page inspector zoom.

You bring up a good talking point, but it's important to understand there are really only 2 ways for Resolve to scale native pixels, without filtering and getting soft - (higher than timeline resolution) pixels that I'm aware of.

1 is the edit inspector zoom control.
1 is fusion single clip comps - when you have a clip on the timeline and click the fusion page (or reference with the bottom layer pulling in the highest resolution)

All other ways, including fusion zoom presets, adjustment clips, and compound clips, all depend on the timeline resolution and raster to that moment.

Hey buddy. See that’s what I thought but yesterday I ran tests with contradicting results. I guess that’s on my list today then — redo the tests, verify output and publish them here!

You bring up a good talking point, but it's important to understand there are really only 2 ways for Resolve to scale native pixels, without filtering and getting soft - (higher than timeline resolution) pixels that I'm aware of.

1 is the edit inspector zoom control.
1 is fusion single clip comps - when you have a clip on the timeline and click the fusion page (or reference with the bottom layer pulling in the highest resolution)

All other ways, including fusion zoom presets, adjustment clips, and compound clips, all depend on the timeline resolution and raster to that moment.

OK, time to run some tests and hopefully challenge this statement. Which, was also my understanding, but now I'm second guessing that, so let's prove it right or wrong.

Here's my test methodology.

Timeline: UHD 16:9 3840x2160
Source media: Open Gate 4:3 5760 × 4320

We already know that if you place the OG media on the UHD timeline and fill the frame (doesn't matter how; [1] Scaling to Fill… [2] Scaling to Crop then Transform Scale (Zoom) down… [3] Scaling to Fit then Transform Zoom up) such the top and bottom of the OG media is clipped, then add an Adjustment Clip above that and Zoom it down, the clipped media does not return. This should be sufficient evidence to confirm what we all know and Chadwick said; Resolve is not scaling native (source) pixels.

In my quest to have an easy way to consistently scale clips on the timeline and not have to copy/paste Transform settings from clip to clip, the one way I came up with to do this with Adjustment Clips (since they can be easily saved, option-dragged around to duplicate, etc.) was to set the OG footage on the timeline with Scaling to Fit, so you see the full 4:3 source with pillarboxes, then zoom up using Adjustment Clips. The math there is that a 1.331 zoom brings brings it up to the equivalent of "Fill", and a 2.0 zoom brings it up to 100%.

So, the test is easy. Put one OG clip on the timeline with Scaling to Crop and Transform Zoom to 1.0, and another at Scaling to Fit and Transform Zoom to 2.0. Both clips are now positioned exactly the same; at 100% source size.

BTW my Project Scaling settings are Resize filter: Custom and Method: Mitchell.

The method I've used for 30+ years to compare two clips is to stack them in Photoshop and apply Difference to the top layer. If the result is anything more than solid black, then there's a difference. However to ensure that there aren't differences that escape the naked eye, there's an extreme test. Flatten those layers, then apply Auto-Levels, which will bring even the faintest shade of grey up to pure white, which will reveal even the slightest differences between the two clips.

So, that's what I did first. Set the two clips (same frame, obviously) as explained above, and exported two TIF files (Export >Current Frame as Still > TIF and there are no compression options listed, so one assumes this is an uncompressed TIF. I also tested this in Cineon format and PNG format). Stacked in Photoshop and differenced, pre-auto-levels, this is what it looks like. This is a crop on my beard and shirt, so lots of details that are easy to spot (it's hard to see against the bright forum page, but right-click and open in a new tab, and you can see it):

TIF export, stacked, differenced

1 difference.jpeg (69.88 KiB) Viewed 3240 times

Let's be fair; that's pretty good. As far as scaling goes, that's impressive. But let's force the issue and see what's really hiding there, by merging the two layers and apply auto-levels, and this is what you see:

TIF export, stacked, differenced, merged, auto-leveled

2 difference + autolevel.jpeg (509.99 KiB) Viewed 3240 times

So, that answers that question, right? Well, maybe not.

I decided to run the same test entirely in Resolve, without exports. Here's how I did that:

1. Make a Compound clip of the "Scaling to Fit and Transform Zoom to 2.0" clip (so the Transform doesn't affect the other test clip)
2. Place that Compound on top of the Scaling to Crop and Transform Zoom to 1.0 clip.
3. Set the Compound to Difference

So far… I see pure black. Nothing visible on screen, and nothing in the scopes that would indicate anything above pure black anywhere on screen. Curious; there should be something on the scopes.

4. Make a Compound Clip of these two, open in Color page, and lift the Gain all the way up (to 16.00). Still pure black.
5. To second-check it, I exported a TIF of that comp, opened that in Photoshop, and applied auto-levels — still, pure black.

I am the first to admit this makes no sense. Did I miss something in my tests? Are my tests flawed? Nothing would make me happier than to know that this works and doesn't sacrifice quality. But, the two tests conflict. I'd argue the all-Resolve test is the accurate one, but TIF exports should be uncompressed so blaming compression there doesn't track.

What are your thoughts??
-Joseph

Resolve and fusion are resolution independent, which is not the most accurate term, I prefer resolution agnostic, but what it really means is that you can keep the original source when you start and than it depends on the operations you do what happens to it. But source will always be available in untouched state. Depending on your workflow you can keep it untouched.... or not by the time you output.

First thing to consider it the timeline settings, edit, input and output sizing and mismatch resolution.

If you want to keep source resolution, use appropriate input scaling option in the timeline settings.

Input Scaling

Contains one setting, Mismatched resolution files, that lets you choose how clips that don’t match the current project resolution are handled. There are four options:

— Center crop with no resizing: Clips of differing resolution are not scaled at all. Clips that are smaller than the current frame size are surrounded by blanking, and clips that are larger than the current frame size are cropped.

— Scale full frame with crop: Clips of differing resolution are scaled so that the clip’s shortest dimension is fit to match the frame. Excess pixels are cropped.

— Scale entire image to fit: The default setting. Clips of differing resolution are scaled so that the clip’s longest dimension is fit to match the frame. The shorter dimension has blanking inserted (letterboxing or pillarboxing).

— Stretch frame to all corners: Useful for projects using anamorphic media. Clips of differing resolutions are squished or stretched to match the frame size in all dimensions. This way, anamorphic media can be stretched to match full raster, or full raster media can be squished to fit into an anamorphic frame. An added benefit of this setting is that it makes it easy to mix anamorphic and non-anamorphic clips in the same project.

.........................

I chose the Center crop with no resizing and the same in the output tab of timeline settings.

I made timeline 1080p while the source image is from ARRI website at 4K.

ARRI_Helen_John_ALEXA_Mini_LF_AWG3_LogC3_to_Rec709g24.tif

https://www.arri.com/en/learn-help/lear ... rence-imag

.........................

I make sure there is no scaling changes in the retime and scaling section of the inspector panel in edit page. And I chose Linear for scaling method. So there is minimal change.

Linear: This uses a simplistic filter, which produces relatively clean and fast results.

.........................

I than use zoom controls in the inspector of the edit page to scale the footage to fit the timeline. in my case it was 0.440 zoom. That way I am keeping the original pixels of the source media, and only temporary zooming in and I can than later zoom out and not lose resolution. Similar to using Smart Objects in Photoshop except there is no precomping.

I export the frame as still in Tiff format and than in fusion page, I apply crop to 1080p to mimic timeline resolution and I used transform node before it to fit the frame at same zoom 0.440 with linear filter for scaling.

In fusion transform and merge tools concatenate, meaning they add their scaling and pass it along to another tool without changing source resolution. Just instruction how to scale it in a chain type of way , so each transform tool will be able to scale it back up and not lose quality. Unless you use checkbox "flatten transformation"

Tool in fusion that actually change resolution are Crop, Letterbox, Scale and Resize tools. I used crop similar to timeline setting Center crop with no resizing and much like zoom options in the inspector I replicated the same with transform tool.

Once all that is done, I merge on top the exported still, so we are now comparing pixel for pixel the original source footage from media pool, scaled and transformed to match the same operations as in edit page, expect done in fusion this time, and than we merge exported still on top with apply or blend mode set to difference. I expected black screen with no visible changes and that is what we get.

Since processing is done in 32 bit float there is a way to force miniature mathematical differences using auto-gain tool in fusion, but that is not visible by naked eye. Photoshop it mainly restricted to integer 8 or 16 bit and tiff is I think 16 or 8 bit interger as well. You could export EXR in 32 bit float and try that, but it shouldn't matter. To replicate the photoshop processing in interger and do the same in fusion, you could add change depth tool after original MediaIn and change ti to 16 bit or 8 bit integer. The image compared with difference mode should still be the same .

sshot-839.jpg (184.78 KiB) Viewed 3201 times

.................................

About Adjustment clips and other generator type clips. I think they are restricted to timeline resolution, and don't support transparency, unless you apply fusion effects inside of it. Otherwise they act on the composite image in the viewer, and don't have access to information outside of it.

Here is 0.440 zoom applied to clip that is larger than timeline resolution, set to Center crop with no resizing and as you can see I can still get all the pixels outside of visible frame.

sshot-840.jpg (226.41 KiB) Viewed 3201 times

But If I apply the same 0.440 zoom applied to clip that is larger than timeline resolution, set to Center crop with no resizing using adjustment clip, I get black since it has no access to original file. It works on the composite image of all the files on timeline. Similar how in fusion page Media In set to "background" as media source, works. It shows the composite image of what is bellow it, similar to adjustment clip and at the timeline resolution. it does not have access to original clip and hence it doesn't have access to pixels outside the viewer. It just works on the composite image, like taking a screenshot of what is in the viewer an trying to work with that.

sshot-841.jpg (193.95 KiB) Viewed 3201 times

.................................

Compound clips and fusion clips also can be useful in some workflow, but one needs to be clear about what they do and what they don't do and why.

If you were to take a clip from edit page timeline and open it in fusion page, fusion will ignore all that was done on the color page and edit page except lens distortion settings. Everything else, like zoom and transform, timeline resolution settings and colorgrading, open FX applied etc, will be ignored and fusion page will only use what is on the timeline as reference to know what file to open and where to set render points, but will source the clips from media pool, ungraded and at source resolution. Giving you maximum flexibility in what you do in fusion while also independent of the edit page for the most part, means you can crop or color grade on top and via the resolve image processing pipeline, the VFX you apply to the clip will be intact, so you don't have to repeat it, while everything else will be added on top.

Normally this is the way to keep flexibility and choose what you want in fusion.

However, if you wanted to conform clips first to whatever is in the timeline to be the same in fusion page, you would than use compound clip or fusion clip first. Once you select some clips in the timeline and convert them into a special kind of nested timeline we call compound clips or fusion clips, they than are automatically also created as copied in the media pool for fusion to access.

Since fusion will always skip edit page and go to source in the media pool. and if the clips you open in fusion are are compound or fusion clips, they will be source again from media pool as duplicate version of the ones in the edit page. Follow the same process as with original clips. Except this time the container that wraps the effects conforms it all to what is in the edit and color pages. and that is what will be open in fusion page.

The difference between compound clips and fusion clips is in the way they communicate with fusion page.

Compound clips flatten all tracks but keep the lenght at timeline resolution. So if you have bunch of clips in edit page of the differnt resolutions, set in both sequential mode, one after another and all a are set to fit and you want the same in fusion page along with color grading already applied, than you by converting them to compound clip and than opening the compound clip in fusion you get one long clip as MediaIn. Along with all the color grading, and at timeline resolution. Useful if you want to apply something in fusion that is exclusive to fusion, like particular kind of blur or something.

compound clips can be decomposed back to original clips, but when you choose decompose option it will discard all fusion and other settings and bring back original clips, so keep that in mind.

Fusion clips are similar to compound clips but are more suited for compositing in fusion, since they will take multiple video tracks in the edit page, stacked on top of it each other and convert them into timeline resolution set of clips, but connected as MediaIn node sin fusion, allowing you to manipulate layers as nodes.

Both compound clips and fusion clips are nested timelines, which you can open by right clicking on them and open in timeline. There the original files are preserved and you can still make changes which will be reflected in the timeline containing the clip and in fusion as well.

When you use fusion clip, all the MediaIn nodes in fusion, corresponding to tracks or layers in fusion clip, have layer numbers. Layer 0,1,2 etc, If you change the number of the layer than that media in node will source a differnt layer in the fusion clip, but at timeline resolution. Most of the time this is not what you need, but it can be useful if you want to reorder fusion clip after the fact or add more tracks to it later. I don't think fusion clips support decompose option.

..........................................

Reference Fusion Composition unlike fusion clips or compound clips acts similar to opening a clip in fusion, meaning it retains original resolution of the source media. And of course being a reference can be used as single point of contact, a command center that can be used to sync up multiple clips in the same timeline, or differnt timelines and even projects if you use power bins or dynamic project switching and copy it manually to a new project.

The dilemma about weather or not to put clips in compound clips, fusion clips or not, depends on what you are doing. if you need all the clips to be the same resolution as timeline because you are building some kind of macro for edit page or something, than you might want to use them, because if you don't all the clips will be at source resolution of media, which may or may not be what you need.

It is possible to extract timeline resolution information in fusion automatically, but it requires extra set up of the composition and some tweaks. When just dealing with differnt resolution but same aspect ratio, this is generally speaking not a problem. But if source clips are differnt resolution and aspect ratio and you want both flexibility and conforming to timeline resolution than you need to apply some more steps in fusion to set it all up.

Like I said, most of the time its not a problem. So if you are not doing that but for some reason require conformity at the expense of flexibility you would use fusion or compound clips. Depending on weather you have files in a sequance or as layered video tracks.



Generally speaking if you know what is happening, you don't have to conform anything at the expense of flexibility, because unlike many adobe products which are layer based and require precompositions, fusion and color page are node based and they work with the rest of resolve, as long as you understand the image processing pipeline and dos and don'ts of the process.

I hope all that helps to clear out some of the ways in which resolution treated in resolve.

Also here is a useful diagram about image processing I post often which helps to clear out order of operations.

DaVinci Resolve_19_page3200_image (2).jpg (124.64 KiB) Viewed 3217 times

sshot-485.jpg (175.07 KiB) Viewed 3217 times

sshot-1666.jpg (58.45 KiB) Viewed 3217 times

One quick thing to add - if you use Mr. Alex Tech latest Magic Zoom preset - you CAN get native source pixels with the plugin, if after you apply the preset - you also drag to the inspector filename section the clip from the media pool. This forces the plugin to reference the source resolution instead of the timeline. It's an extra step but can but useful - and his tool is handy!

1) if it is the same value (like predetermined scale) to a bunch of clips, in media page, clip attributes you can apply a preset that you saved from the color page scaling.
2) do the transform you want to the clip you need, then drop an adjustment layer, copy paste from clip to layer, pit the layer back in the media page bin and store it. When you need it again, drop it in the timeline, copy from there the values, delete it, then you can select as many clips you want and paste the values

your entire post…

Thanks for going through all of that; that’s quite a bit of work! You’ve explained the Fusion process very well, and that’s what I would have expected to see as a result. Since my original objective is to find an easy way to repeat movements, using Fusion isn’t on the list. And I did know that Fusion operated from the full resolution file, but it was nice to see that confirmed.

I guess I’m not sure from your results which one of my demonstrations was meant to be right (if that’s even what you were trying to show)… it does not seem to matter which scaling method you choose; once you match the visible size, it’s all the same. So I’m still unsure how to interpret my test results.

your entire post…

Good reminder on the flow of things, thanks. I don’t meant to be obtuse, but is this meant to explain the discrepancy that I’m seeing? In this part of my thread viewtopic.php?f=21&t=216874&p=1125890#p1125858 where I question why I’m seeing the apparent results of scaling vs not, I’m only working on the edit page (yes using adjustment clips but those are supposedly only acting on what is already visible underneath them), not going into Fusion or Color.

Essentially I don’t know what Resolve is doing; pulling back to the source (the OG clip’s full resolution), or not.

One quick thing to add - if you use Mr. Alex Tech latest Magic Zoom preset - you CAN get native source pixels with the plugin, if after you apply the preset - you also drag to the inspector filename section the clip from the media pool. This forces the plugin to reference the source resolution instead of the timeline. It's an extra step but can but useful - and his tool is handy!

Yep, but it doesn’t work well with video. We discussed it quite a bit when he first released it. Really great tool though for sure.

1) if it is the same value (like predetermined scale) to a bunch of clips, in media page, clip attributes you can apply a preset that you saved from the color page scaling.
2) do the transform you want to the clip you need, then drop an adjustment layer, copy paste from clip to layer, pit the layer back in the media page bin and store it. When you need it again, drop it in the timeline, copy from there the values, delete it, then you can select as many clips you want and paste the values

Thanks. This is just further proving that Resolve needs an “adjustment presets” tool (like FCP has; I know people hate it when I say that, but it’s bloody true). Personally I like the idea of being able to assign adjustment presets to keyboard shortcuts. Would make so many things so much faster.

Thanks for sharing your workflow.

Essentially I don’t know what Resolve is doing; pulling back to the source (the OG clip’s full resolution), or not.

Adjustment clips work with composite image so they don't really work with source resolution. if access to source resolution is what you need, adjustment clips are not the tool to use.

Understanding input and output scaling and how mismatch resolution works , is critical to understanding how clips are treated on the timeline, and if you understand how it works, you can get anything you need. So its worth putting the effort to understand it. Which is why I tried to explain much of it, but not all. The rest can be found in the manual.

if you need anything more advance or want to build custom presets, fusion is generally best for scaling and transformations since it has tools you would need for just about anything and it works with source resolution by default. if you are looking for presets for transforming clips, fusion is the logical place to go, but you need to understand how it works to take advantage of it. Which is why I explained various methods in which fusion and edit page communicate to each other.

There is also color page which deals with output scaling and blanking etc. Depending on what you need it may be important aspect of transforming clips as well. You can also find more about it in the manual.

Generally speaking there are quite a few places and methods by which you can transform and scale clips in resolve. They each have their place and are quite powerful if you know how they work.

Essentially I don’t know what Resolve is doing; pulling back to the source (the OG clip’s full resolution), or not.

Adjustment clips work with composite image so they don't really work with source resolution. if access to source resolution is what you need, adjustment clips are not the tool to use.

Understanding input and output scaling and how mismatch resolution works , is critical to understanding how clips are treated on the timeline, and if you understand how it works, you can get anything you need. So its worth putting the effort to understand it. Which is why I tried to explain much of it, but not all. The rest can be found in the manual.

if you need anything more advance or want to build custom presets, fusion is generally best for scaling and transformations since it has tools you would need for just about anything and it works with source resolution by default. if you are looking for presets for transforming clips, fusion is the logical place to go, but you need to understand how it works to take advantage of it. Which is why I explained various methods in which fusion and edit page communicate to each other.

There is also color page which deals with output scaling and blanking etc. Depending on what you need it may be important aspect of transforming clips as well. You can also find more about it in the manual.

Generally speaking there are quite a few places and methods by which you can transform and scale clips in resolve. They each have their place and are quite powerful if you know how they work.

Hey again. I do understand that — but did you see my comparison post? It makes no sense. The image appears to be scaling from the source when using an adjustment clip — but that should not be the case.

Hey again. I do understand that — but did you see my comparison post? It makes no sense. The image appears to be scaling from the source when using an adjustment clip — but that should not be the case.

Hey again yourself. If you read all I wrote, I'm sure you will find the answer. I don't have ready made files you have to replicate exactly what you did, but I assure you if you understand what I wrote, and how resolve works with resolution, you should have no problem getting results you want. I don't. Since you are the one that has the problem, assume you are missing something.

Hey Joseph - I haven't kept fully up on this post - but one thing I think you might be missing is about compound clips.

They concatenate if there is only 1 layer inside of it. The moment you stack another clip on another track inside the compound clip - it is now baked down to the compound clip resolution and you can't access those native pixels anymore.

So - for 6k, 8k, 12k blows ups on a 4k or 1080 timeline you really should do them only in the edit zoom inspector or a fusion transform.

Hey Joseph - I haven't kept fully up on this post - but one thing I think you might be missing is about compound clips.

They concatenate if there is only 1 layer inside of it. The moment you stack another clip on another track inside the compound clip - it is now baked down to the compound clip resolution and you can't access those native pixels anymore.

So - for 6k, 8k, 12k blows ups on a 4k or 1080 timeline you really should do them only in the edit zoom inspector or a fusion transform.

Hey Chadwick. Yep, I understand that. You should check out my lengthy post above about my test results; I think it's quite interesting. KrunoSmithy wrote something detailed that I need to dig into but haven't had the time to come back to it yet, but I will once I wrap this current edit.

Hi - I see your tiff export tests. For a 2x blow-up, that's not going to be bad.

It always depends on the content, too, regarding the filter method (you used Mitchell).

I would like you to place a 6k or a higher image on a 720p timeline to test a 10 times blow up. The software doesn't change how it works on this, and it will be much more apparent on why you should scale in either:

1 - edit transform controls
or
2 - fusion transform

That being said, this is all relative to speed and workflow. Presets are nice and I use them a lot myself. But those are the two options to keep the quality of very high-resolution material without filtering.

One more thing to note - Resolve is working in a 32-bit float space. Photoshop is integer and rounds off code values (it is probably float under the hood, but not to the user). This is one of the reasons one will give a different result.

OK, I'm finally back to this.

Essentially I don’t know what Resolve is doing; pulling back to the source (the OG clip’s full resolution), or not.

Understanding input and output scaling and how mismatch resolution works , is critical to understanding how clips are treated on the timeline, and if you understand how it works, you can get anything you need. So its worth putting the effort to understand it. Which is why I tried to explain much of it, but not all. The rest can be found in the manual.

I hope you understand -- my comment that "I don't know what Resolve is doing" was not a blanket "I don't know", but rather a comment after my tests, where I do know what Resolve is supposed to be doing, which you confirmed with your extremely extensive post (which I genuinely appreciate), what I know (and what you confirmed) did not match up with my tests.

That said, I've reviewed my detailed test post, and it seems that I left something out of my test explanation that explains the confusion. My test was accurate; my explanation of my test was not. If you want to ban me from the forums for this boneheaded mistake, I'd understand (please don't).

Where I wrote:

So, the test is easy. Put one OG clip on the timeline with Scaling to Crop and Transform Zoom to 1.0, and another at Scaling to Fit and Transform Zoom to 2.0. Both clips are now positioned exactly the same; at 100% source size.

I left off the exceedingly important point that in second test, where Transform Zoom was set to 2.0 — that that was on an adjustment layer.

The whole point of my test was to show that scaling the image entirely in the Transform controls vs scaling it using an Adjustment Layer was different. And the Photoshop export proved that. However the internal test in Resolve disproved it — according to the Resolve-only test, the pixels were the same.

I feel like such an idiot for leaving that out. So I'm going to follow something Chadwick said and take this farther — put 6K footage on a 720p timeline. The difference will be much more dramatic.

And it is… and oh my god I figured it out. Maybe this is what some of you were explaining… but I think with my omission, no one was. But anywhere, here we go.

It is in fact as expected, which becomes very obvious when going from 6K to 720p. Scaling using an Adjustment Layer does scale from the timeline, not from the source. Here we see (now accurately written to include the Adjustment Layer component):

720p timeline, 5760×4320 Open Gate source
(1) OG clip set to Crop, Transform 1.0, so we see it at 100% source scale
(2) OG clip set to Fit, Adjustment Clip on top with Transform 6.0, so we see it at 100% source scale

Comparing Crop to Fit + Adjustment Clip

CleanShot 2025-03-06 at 11.22.02.jpg (245.97 KiB) Viewed 1164 times

OK, this is the result we expected. It's super blurry because it's been scaled down then back up 6x. This is now super obvious to the naked eye, where before scaling just from 4K to 6K, it wasn't so obvious. In the previous test, at this point I exported a TIF of each and stacked/differenced them in Photoshop. Now there's clearly no need to do this as we can plainly see they are very different.

Then, in the original test, I stacked the two in Resolve and differenced them. However, if I just moved the (2) Clip+Adjustment Clip on top of the (1) Clip, then the Transform Clip from (2) would affect the underlying clip in (1). So to avoid this, I made a composite of (2), and placed that comp on top of (1), then differenced them in Resolve, and there was no difference between them. Very confusing.

However… what do you know… THIS is the stage that it gets "fixed".

Look at these two side by side… (2) is the unchanged, and (3) is simply (2) added to a Compound Clip.

Comparing Adjustment Clip, and same thing in a compound clip

CleanShot 2025-03-06 at 11.30.46.jpg (244.45 KiB) Viewed 1164 times

By placing the stack in Compound Clip, it is now going back to the source media!!

Perhaps a Compound Clip is essentially a Fusion Composition, so this makes sense? And if this is what y'all were trying to explain earlier… then my apologies for missing it. But this is HUGE to me.

So now for my workflow of using 6K OG clips on any size timeline means I can set the timeline to Fit, use an Adjustment Layer on top of every clip to bring it to size, which makes duplicating scale/positions super easy to do, then just put all of those into a single compound clip before rendering to ensure I have the cleanest pixels.

I don't know about you guys, but to me, this is huge.

(I'd still rather have adjustment presets, but hey, this is a VERY good solution for now).