Lens profile corrections

[Susi Gata]

It appears that all lenses newer than about 16 months have no profile corrections available. I need the corrections for the Sigma 24-70mm f2.8 DG DN and the Sony 70-350mm f4.5-6.3 g oss. Both lenses are popular and were released before Dec 2019. 

When will these profiles be available?

Edited March 18, 2021 by Susi Gata

[NotMyFault]

This might help

 

[Susi Gata]

Thank-you. I have been checking the Github lensfun database. It looks like it has not been updated since 2019-02-16. I'm still missing the profiles for the Sigma 24-70mm f2.8 DG DN and the Tamron 70-180mm f2.8 Di III VXD. A comment from Affinity would be nice. The Sigma 24-70mm f2.8 DG DN is a very popular lens. Waiting more than a year for the profile correction is too long.

I just noticed that the profile for my Sony 24-105mm f4 was included with version 1.9 of Affinity Photo. 

Edited March 19, 2021 by Susi Gata

[Xasi]

Hello

I'm using a Tamron 17-28 mm F2.8 Di III RXD Sony E. This lens correction is not included in the newest lensfun database. When will you add it?

Thank you

[Lee D]

I'm afraid we don't have control over which lens profiles are added to the LensFun database that Affinity Photo uses and when they release their updates. When they do we pull it and add it to our own code, ready for when we release an update to our apps or as part of a beta testing cycle.

[eobet]

On 7/6/2021 at 5:56 PM, Lee D said:

I'm afraid we don't have control over which lens profiles are added to the LensFun database that Affinity Photo uses and when they release their updates. When they do we pull it and add it to our own code, ready for when we release an update to our apps or as part of a beta testing cycle.

The creator of that database has said that he does not have time to maintain it anymore. You are a commercial entity using that database, and apparently it is open source as well. Why don't you put some effort back into the community and update it yourselves?

[walt.farrell]

3 hours ago, eobet said:

Why don't you put some effort back into the community and update it yourselves?

Serif would need to own all those cameras and lenses to do that. I wouldn't expect them to do that.

Users who own a camera/lens can also perform a calibration, using the instructions and programs linked from the Lensfun site, and then upload the information to Lensfun so it's available to others.

[somethingp]

Can they not at least provide a warning before purchase that many recently released lenses are not supported (even if the lenses are very popular), and will not be supported unless they are added by a community run database over which Serif has no control? I purchased the software without even knowing this could possibly be a problem (coming from lightroom where I've never encountered a missing lens profile). And, no, serif wouldn't need to purchase every camera and lens. If reviewers can get free loaners, I'm sure Serif could build partnerships with manufacturers to borrow lenses and cameras at no cost.

[Ron P.]

8 hours ago, somethingp said:

If reviewers can get free loaners

😄, Not technically free. Sure the Reviewers may get the use of them to provide a review, but the reviewers are providing advertisement for the manufacturers. Ever notice how those reviews are somewhat bias toward the manufacturer? Think the reviewer will be asked to review another lens/camera if they find a lot of faults, dislikes?

Take a look at this list. Do really think that Serif can get their hands on that many and more lenses and cameras?

 

[R C-R]

15 hours ago, somethingp said:

If reviewers can get free loaners, I'm sure Serif could build partnerships with manufacturers to borrow lenses and cameras at no cost.

Even if they could do that, making the calibrated lens profiles is not a quick & easy process. It requires many steps & specialized equipment. Considering how many new cameras & lenses are released just in a single year, it should be obvious why Serif relies on a third party for this instead of trying to do it all in-house (& pass the added expense on you its customers).

[v_kyr]

On 10/24/2021 at 8:54 PM, walt.farrell said:

Serif would need to own all those cameras and lenses to do that. I wouldn't expect them to do that.

Not necessarily, since it highly depends on where the lens data stems from, further if you are able to reuse the manufactors data out of RAW files data or not! - In case of Affinity they relay on third party services here (aka the LensFun DB), others (like the big players) are also able to reuse just the manufactor lens profile data out of RAWs for lens corrections, as far as those are there.

See and read for example:

• Lens correction in Capture One: The Essentials   (read there about Manufacturer Profiles and their advantages)

 

[walt.farrell]

50 minutes ago, R C-R said:

making the calibrated lens profiles is not a quick & easy process. It requires many steps & specialized equipment.

I've never tried it, but from reading about the steps needed to make profiles for LensFun (and thus for Affinity Photo) it seems relatively straightforward.

[R C-R]

15 minutes ago, v_kyr said:

Not necessarily, since it highly depends on where the lens data stems from, further if you are able to reuse the manufactors data out of RAW files data or not!

Where would Serif get the RAW files with this data embedded in it? If they could get that, would the info in one file be enough or would they need to consider the data from many different files?

I guess what I'm asking is if it that easy to get the necessary data, why would anyone need to use the LensFun DB?

[v_kyr]

7 minutes ago, walt.farrell said:

I've never tried it, but from reading about the steps needed to make profiles for LensFun (and thus for Affinity Photo) it seems relatively straightforward.

Yes I think so too. - Even for Adobe's LCP files it looks to be relatively straightforward, here is an older description of how to use their Lens Profile Creator tool.

[R C-R]

2 minutes ago, walt.farrell said:

I've never tried it, but from reading about the steps needed to make profiles for LensFun (and thus for Affinity Photo) it seems relatively straightforward.

Straightforward or not, wouldn't it take considerable time to do it for each lens & camera combination?

4 minutes ago, v_kyr said:

Yes I think so too. - Even for Adobe's LCP files it looks to be relatively straightforward, here is an older description of how to use their Lens Profile Creator tool.

From that, it does look like the procedure has to be done for each lens & camera combination.

[v_kyr]

20 minutes ago, R C-R said:

Where would Serif get the RAW files with this data embedded in it?

From the cam manufactors related shown cam sample RAW images and from all the huge foto cam specific sites (canon r3 raw samples) etc. etc. - Aka the same way I get specific cam manufactor/models RAW files from. Further all the cam manufactors always test their whole by a cam supported lens equipment, thus they have all the needed RAWs and data.

20 minutes ago, R C-R said:

If they could get that, would the info in one file be enough or would they need to consider the data from many different files?

Depends first of all on their knowledge to parse right the cam related Exif maker notes, second on the lens data format and db they relay on in their own software and if they have own accurate help tools for these profiling tasks. - Further it never hurts to make some more secure checks, in order to see if your parsing is accurate among several RAW files then.

[walt.farrell]

4 minutes ago, R C-R said:

wouldn't it take considerable time to do it for each lens & camera combination?

It would take some amount of time, and it would require having each camera and lens, yes. 

"Considerable", to me, feels like hours per lens/camera combination, and I don't think it's that much. Mainly, I think, I was reacting to your comment about needing special equipment, and I don't think that's needed.

[R C-R]

3 minutes ago, walt.farrell said:

"Considerable", to me, feels like hours per lens/camera combination...

Even if it was just one hour per combination, or just 30 minutes, consider how many combinations that would be & how much time it would take to do them all.

But I keep coming back to the same question: if it was not so labor intensive to do this, why are there services like the LensFun DB to begin with?

[v_kyr]

3 minutes ago, walt.farrell said:

It would take some amount of time, and it would require having each camera and lens, yes.

If you do it always with own calibration sheets and don't or can't reuse the cam/lens manufactors already available data. Further as I said above, every cam manufactor tests before their whole by a new cam supported own brand lens equipment etc., thus they (the cam manufactor) have all the needed data anyway. - Then if you have a close collaboration with those (as C1, Adobe and maybe DxO too have), you will get that data sooner or later.

[R C-R]

9 minutes ago, v_kyr said:

Depends first of all on their knowledge to parse right the cam related Exif maker notes, second on the lens data format and db they relay on in their own software and if they have own accurate help tools for these profiling tasks. - Further it never hurts to make some more secure checks, in order to see if your parsing is accurate among several RAW files then.

But does this still have to be done individually for each camera & lens combination, or is there some way to avoid doing that?

[R C-R]

3 minutes ago, v_kyr said:

Further as I said above, every cam manufactor tests before their whole by a new cam supported own brand lens equipment etc., thus they (the cam manufactor) have all the needed data anyway.

Do camera makers like Cannon test with 3rd party lenses or just with the ones they make?

5 minutes ago, v_kyr said:

Then if you have a close collaboration with those (as C1, Adobe and maybe DxO too have), you will get that data sooner or later.

I have to wonder if Cannon, Nikon, et al would have any interest in a close collaboration with other companies like Serif, or if there would be any strings attached, like fees.

[v_kyr]

1 minute ago, R C-R said:

But does this still have to be done individually for each camera & lens combination, or is there some way to avoid doing that?

Good question. - Related to the above referenced C1 lens correction essentials, they can do parse it out of the RAWs and those are mostly cam specific often slightly different. So even among one and the same cam manufactors, different cam models those with the same used sensor types, might not always have the exact same data here. Obviously if they use then also different sensor formats (full frame, APS-C etc.).

[R C-R]

4 minutes ago, v_kyr said:

Good question. - Related to the above referenced C1 lens correction essentials, they can do parse it out of the RAWs and those are mostly cam specific often slightly different.

So then if we also add 3rd party lenses to the mix (like a Sigma lens on a Cannon body) aren't we talking about a huge number of combinations & a ton of labor to keep all the data up-to-date?

[v_kyr]

3 minutes ago, R C-R said:

Do camera makers like Cannon test with 3rd party lenses or just with the ones they make?

Sony for sure not only tests with their own lenses but also with those they have a close collaboration with, like Zeiss. They also have specified their lens mounts, so third party lens maker (Zeiss, Sigma, Tamron, Tokina ...) can make use of that. - Nikon and Canon do have a long history in manufacturing huge own lens lines, they also have specs about their lens mounts which other third party lens makers use to adapt to their cams. But I doubt that they will test every third party lens here.

[v_kyr]

2 minutes ago, R C-R said:

So then if we also add 3rd party lenses to the mix (like a Sigma lens on a Cannon body) aren't we talking about a huge number of combinations & a ton of labor to keep all the data up-to-date?

If you look into LensFun db you can see how many lenses that has actually as entries and if they do share similar entries with other lenses of that same kind. Though the overall procedure of LenFun is different from taking the lens datas out of RAW files here, they use target based calibrations and some tools for reusing Adobe LCP profiles data too.