Hi All
Just thought I’d check this. So I have taken a photo the ISOs only 320 but there’s a bit of grain.
So I ran this RAW through PL7 with DeepPRIME XD and then through DeepPRIME XD3 X-Trans in PL9 - all the adjustments were exactly the same.
The PL7 jpeg was bigger 11.9mb, to 8.6mb. I was somewhat surprised. What didn’t surprise me was the end result. If anything the RAW run through PL7 was sharper.
Maybe that’s why the file is bigger, it has more detail?
I’m surprised at the significant difference in file size, images of the same resolution will have slightly different JPG sizes due to the amount of color variation (it’s the color variation that shows details).
In your testing did you use exactly the same levels of JPG compression/quality in the Export panel?
I think that we’re comparing fruit and vegetables here.
The X-Trans denoising is quite different from DxO’s generic de-noiser and would therefore need different settings to obtain similar results.
Moreover, DxO’s algorithms are reworked and updated and might therefore deliver different results even if the names were identical.
I propose to export to TIFF for such tests. This removes the risk of drawing conclusions that are mainly influenced by JPEG compression.
I’m not surprised as I’ve already seen that before when choosing different noise reduction.
And for me it’s quite logic as when choose an option with less noise reduction this means keeping more noise which could be considered as « details » which will be less compressed in the jpeg.
And if the image have great uniform zones the difference in file size will be greater.
@Gareth Once again where is your proof, just stirring up the mud to see what comes to the top!?
I just exported 23 X-Trans images with PL7(XD) and PL9(XD2s i.e. XD for X-Trans) and PL9(XD3).
I need to go to the dentist for an emergency appointment but here are three images to be going on with
and I have another 22 x 3 to look at when I get back.
So the times for export on the 5900X-5060TI(16GB) were
and the sizes of the 100% JPG exports are
There are differences in size are present and a comparison at 300% shows this
The image, from PL7 XD and PL9 XD2s (XD when used with X-Trans images) are sharper that the XD3 image but that happens when comparing XD exports to XD2s, in my experience.
XD2s sought to resolve issues with XD and it is generally better on high ISO images, at removing noise and (arguably) not adding unwanted artifacts.
But I have always complained to DxO that it is not as sharp as XD was and, unfortunately DxO removed it from PL8.
So no PL8 user can use the additional facilities of PL8 while preserving what they may have considered a better noise reduction for some images, excluding X-Trans users who only had XD, labelled as XD/XD2s available to them!
The problem is that I cannot complete my testing because I would need a Bayer and an X-Trans image taken of the same scene at the same time with the same settings, but then it would still be a different camera with a different lens!?
Your comments showed none of what you saw in your tests and in my opinion are worthless until you start including your tests results.
You make derogatory remarks about DxO, as do I, but with nothing to back them up, whereas I bore other forum users with excruciating levels of detail.
To further support your claim the XD3 isn’t as good with detail as XD, at 500% (fairly extreme pixel peeping) on another image we have this
But I had to search hard for another image which showed that level of difference on a 4K screen.
As a PC user you have access to FastStone Image Viewer for free, from which the comparisons come, but the snapshots come from FastStone capture and that costs money.
Exported at 100% for both
To stay on the same level of rudeness, if you didn’t understand @Pathal reply, the same applies to you.
BTW, I wonder how long DxO will tolerate this amount of useless posts that @Gareth has created.
Usually I love you sarcasm 
. But yeah, may a bit strong in some cases 
@andras.csore and @Wlodek Thank you for pointing out that it was a little underhand but its fine.
@Gareth Sorry, hadn’t I kept you up to date, that was last week when I picked up my new glasses at the opticians. The sting will come at the end of this week when I need to pay for them.
However, it would have been impossible, new glasses or no glasses, to comment on your image(s), unless I was a clairvoyant, because you hadn’t published them then, one of the principal reasons for my concern.
I did review X-Trans images I had exported and also stated that I have made observations to DxO in the past that I believed that XD2s was “soft” compared to XD, but it does tend to do a better job on high ISO images.
However, my biggest objection was, and still is, that DxO removed the XD option entirely, for non X-Trans users, thereby denying PL8 and now PL9 users from the option of choosing one or the other on an image by image basis,.
Given that XD was still in the product (and still is), for X-Trans users, how hard was it to provide access to that facility for all users, one extra item on the Noise Reduction menu which had suddenly increased in size on PL8 anyway.
Sorry, this is your topic but
With respect to your images, for some reason they don’t they don’t align and you happen to have given them both the same name!? You did suggest that the PL7 export was bigger than the PL9 export but when coupled with the alignment issue it is all a little confusing
It is possible to align them but why the discrepancy?
Using FastRawViewer one is definitely sharper than the other, and I don’t have to rely on my aging eyes, the program does that for me!?
or
With the image I included in my post we have
So you are correct in your analysis that there are differences so submit a support request asking for XD to be made available to all users and reference this topic.
So was I but it fell on deaf ears.
I will dig out the XD2s versus XD images and “bore” you with them
It’s amazing how people claim they support free speech, until they disagree with it.
perhaps better to know when to stay silent
I copied the RAW file into 2 separate folders it has the same RAW file name. I then opened the first in PL7 and did a little post processing. Then saved the jpeg to the same folder. Then I opened the second RAW in PL and did the same, using the exaclty the same settings. So each has the same number.
As you can see each is slightly different because each was processed and cropped separately. The difference in file size though cannot be attributed to that.
As you say, it clearly can be seen that the PL7 file is sharper and I check at 100%, if you go further (though I never do), well there is a great deal of difference. Whether this extra detail is the cause of the larger file size, who knows?
I am glad you confirmed my findings.
Strangely, the irony that the DeepPrime that is designed specifically for X-Trans sensors is not as sharp as the DeepPrime XD helps me with the decision to keep using PL7.
But it was ok for him to say “!Once again where is your proof, just stirring up the mud to see what comes to the top!?”
Instead of checking one of his own files and now he has done so. Yes I am right.
I don’t hear anyone saying, thanks for pointing this out? By now people should be aware that because a new version comes out, it doesn’t make it better, just newer.
What amazes me, in all this time. All these people who I am sure process way more photos than I do, haven’t noticed.
You were simply rude and my reply had nothing to do with free speech (or free lying).
Let me repeat once more:
German: Wenn Sie nichts zu tun haben, tun sie es nicht hier!
English: If you have nothing to do, don’t do it here.
French: Si vous n’avez rien à faire, ne le faites pas ici.
Italian: Se non hai niente da fare, non farlo qui.
Spanish: Si no tienes nada que hacer no lo hagas aquí.
Hindi: अगर आपके पास करने के लिए कुछ नहीं है, तो यहां मत कीजिए।
Georgian: თუ საქმე არ გაქვს, აქ ნუ გააკეთებ.
Understood?
Out of your recent 100+ posts, zero were useful to the community. You earned GarethZero nickname, feel great. Congrats, over.
If you have nothing to do, don’t do it here. - If you have nothing to do, how can I do it here? You obviously did not read his post first. Try harder.
Hi.
Before this topic gets closed for vitriol -
Could it be that the differences in sharpness aren’t between XD3 and X-Trans but are between different versions of Lens Sharpness Optimization? Using the same adjustment values in PL7 and PL9 probably won’t produce the same results. Can we rule this out?
To find out, I did the following:
Note that I left all settings in PL and LrC at their default values in order to exclude differences caused by different customising.
Lessons learned
Whether one prefers a certain denoising method over an other is a matter of taste and the subject at hand. While less detail might be preferable for images of cars, images of furry creatures want more detail…specially when seen from up close.
I have just re-run with no adjustments. JPEGS come out 7.18mb and 8.02mb. It appears that the difference between DeepPRIME XD3 X-Trans vs DeepPRIME XD remains. Perhaps the reduction in noise in DeepPRIME XD3 X-Trans removes more detail, so at ISO 320 the result is a softer photo?
I’m about to go through a load of my old X-Trans files soon so I’ll be looking at this closer now when I do
Admittedly I have nothing to compare against as I am only using PL9 now (can no longer access my last version PL5) but recent exports of two test files (XT-4 with 55-200) I dug out whilst evaluating PL9 were both perfectly sharp (one of a Jay the other a Squirrel) to my eyes and gave no reason to complain.
Maybe they would have been better under PL7 though hence your findings.
Anyway. I don’t know if you look in the PureRaw section on here but prior to going with PL again I toyed with getting PR and sticking with LR. In the end I never trialled PR but I wonder if it’s related to the experience/findings some are seeing and the ‘work arounds’ they have employed may help.
Worth a look if you haven’t seen it already.
