A7Riv - what use 61MP? Or 240MP pixel-shifted?

Beatsy · Post by **Beatsy** » Thu Aug 29, 2019 8:43 am

I have an A7Riv arriving tomorrow and a whole battery of wide and telephoto lens tests waiting for it. That will easily fill my spare time through Saturday. Then a footie (soccer) cup match Sunday to test AF and tracking.

But by next week I'll get onto macro-related tests. The main thing I'm currently interested in is how much detail can be captured with a low-ish mag objective (probably Minolta DImage 5400). Application is for capturing archive/record shots of microscope slides without stitching (though shallow stacking may be involved). The 4-shot "true colour" mode certainly looks useful (61MP) but I'm intrigued to see if the 16-shot 204MP mode can pull out any extra/useful detail.

Any comments or suggestions before I get to it...?

ray_parkhurst · Post by **ray_parkhurst** » Thu Aug 29, 2019 9:04 am

I'm interested in the A7Rm4 but as I am not an early adopter (I got burned on the m3) I am waiting for a few questions related to macro work to be answered before purchasing...in no particular order:

- Can the m4 be used with USB-tethered Live View (with Imaging Edge or any other software) while simultaneously allowing hardwire (USB or other) triggering for stacks. I understand that the m4 has more connectivity than the m3 so am hoping this problem has been solved

- Of course comparing the 4-shot mode with 16 shot mode and with 2x downsized 16-shot mode is of prime concern.

- Does Imaging Edge do the same unrealistic-looking oversharpening to the 4-shot and/or 16-shot images that it did to the m3 composites

- If not oversharpened, then how realistic do the 4-shot and 16-shot images look compared with similar FOV taken with higher magnification optics. A microscope objective of ~2x the magnification used for the composite mode, or even a good 2x teleconverter, would likely answer this question long as the basic optics are high enough resolution, which they would need to be to make sense using them in composite modes anyway.

- What workflow is required to do a stack.

I'm sure some of this is already on your plate but I thought I would put in the added requests before you start work with the new camera.

zed · Post by **zed** » Thu Aug 29, 2019 9:11 am

Personally I am very interested in this camera and would love to hear your thoughts on a few things:

1. I would be really interested to see what you find in 5X range with the pixel shift images compared to single frames at 10x and 20x.

2. What your thoughts are on the tethering solutions Sony provides - I understand the USB-C port eliminates the need for a separate AC power adapter but also supports wireless tethering. I would like to know how stable this solution is.

Thanks!

ray_parkhurst · Post by **ray_parkhurst** » Wed Sep 04, 2019 4:02 pm

Beatsy...any results so far?

Lou Jost · Post by **Lou Jost** » Wed Sep 04, 2019 4:27 pm

I've been enjoying the pixel shifting of my S1R, which seems artifact-free. Of course it will always be better to double resolution by stitching four images at double the magnification, and it may even be faster to do it that way! But pixel-shifting is a painless and simple way to greatly improve the resolution of images. I have found that even older legacy film-era macro lenses benefit from pixel-shifting at m=0.5.

ray_parkhurst · Post by **ray_parkhurst** » Wed Sep 04, 2019 4:42 pm

Lou Jost wrote:I've been enjoying the pixel shifting of my S1R, which seems artifact-free. Of course it will always be better to double resolution by stitching four images at double the magnification, and it may even be faster to do it that way! But pixel-shifting is a painless and simple way to greatly improve the resolution of images. I have found that even older legacy film-era macro lenses benefit from pixel-shifting at m=0.5.

I'm still hoping to see some comparison shots Lou, similar to what I requested of Beatsy, to show how realistic the pixel-shift images are versus an image taken with optical magnification.

My expectation is that the results will be somewhat unbelievable, but that the extra detail and color information would give an excellent image when downsized 2x back to the original size. I saw this phenomenon when shooting SR. With SR you can take a lot of pics, and as long as you don't overlap exactly the results can be really nice, but the larger image is just not quite "right". However, when that image is downsized 2x, it gives a really nice boost in color resolution and overall sharpness vs a single image.

Even if the A7Rm4 is unable to do a great job at 240MP, I still expect it would do well with 60MP when downsized 2x from 240, and that still keeps it interesting for me.

Lou Jost · Post by **Lou Jost** » Wed Sep 04, 2019 4:53 pm

I have never seen any problems with my pixel-shifted images, though of course they aren't as sharp at 100% as a properly-stitched image at twice the magnification. They become very sharp at 100% after downsizing to 66% of original size.

There was no false detail in my example posted here:

http://www.photomacrography.net/forum/v ... 1r+feather

On the contrary, it was the unshifted image that showed false detail, because of strong moire.

ray_parkhurst · Post by **ray_parkhurst** » Wed Sep 04, 2019 7:01 pm

Lou Jost wrote:I have never seen any problems with my pixel-shifted images, though of course they aren't as sharp at 100% as a properly-stitched image at twice the magnification. They become very sharp at 100% after downsizing to 66% of original size.

There was no false detail in my example posted here:

http://www.photomacrography.net/forum/v ... 1r+feather

On the contrary, it was the unshifted image that showed false detail, because of strong moire.

I do remember that post, but was still not convinced. I'm not really thinking of false detail, but perhaps that's what you could call it. When I shot with the A7Rm3, the details were just "wrong", ie the picture had an unnatural look to it vs the 2x optical image. I'm still curious if this is the case with other pixel shifting cameras. A simple comparison of the image taken with a good 2x teleconverter should show how well the pixel shifting works to create a realistic-looking image.

Lou Jost · Post by **Lou Jost** » Wed Sep 04, 2019 7:18 pm

My 2x converter is 40 years old, not sure it is good enough. But I have seen nothing unnatural about any of my shifted photos, at any magnification, neither on my MFT cameras nor on my Pentax K-1 nor on the S1R, so that kind of test has slipped somewhat in my priority list lately. I apologize for that, I know I had offered to do such tests, but life has been unexpectedly very busy lately and I have had to prioritize harshly.

ray_parkhurst · Post by **ray_parkhurst** » Wed Sep 04, 2019 7:55 pm

Lou Jost wrote:My 2x converter is 40 years old, not sure it is good enough. But I have seen nothing unnatural about any of my shifted photos, at any magnification, neither on my MFT cameras nor on my Pentax K-1 nor on the S1R, so that kind of test has slipped somewhat in my priority list lately. I apologize for that, I know I had offered to do such tests, but life has been unexpectedly very busy lately and I have had to prioritize harshly.

It's OK Lou. If I end up with another pixel-shifting camera I'll do the tests. I just hope the results are better than I saw before. The result you showed before looked "good" but I am not sure how the 100% image would look with the converter instead of pixel-shifted, so it's tough to judge. Ultimately my backup is a downsized image which likely will look "natural" assuming the compositing algos are not completly off.

Lou Jost · Post by **Lou Jost** » Wed Sep 04, 2019 7:59 pm

Dropping down the shifted image size to 50% does result in an excellent image, much improved over the unshifted version.

ray_parkhurst · Post by **ray_parkhurst** » Wed Sep 04, 2019 8:01 pm

Lou Jost wrote:Dropping down the shifted image size to 50% does result in an excellent image, much improved over the unshifted version.

I see a similar thing in the SR images. So I guess SR is my backup to downsized pixel-shifting which is backup to full-size pixel-shifting. It's nice to have options.

Beatsy · Post by **Beatsy** » Thu Sep 05, 2019 12:19 am

ray_parkhurst wrote:Beatsy...any results so far?

Still buried in "ordinary" photography stuff (shooting and testing). Should get to studio macro tests at the weekend.

ray_parkhurst · Post by **ray_parkhurst** » Tue Sep 24, 2019 10:29 am

Beatsy wrote:
ray_parkhurst wrote:Beatsy...any results so far?
Still buried in "ordinary" photography stuff (shooting and testing). Should get to studio macro tests at the weekend.

Not to badger you, but I am hoping you've been able to do some macro work with the m4. If not, then no worries, I'm just anxious to determine if it will be workable in my system.

chris_ma · Post by **chris_ma** » Tue Sep 24, 2019 11:53 am

dpreview just posted their studio scene for comparison. the page with their methology:
https://www.dpreview.com/articles/65114 ... fic_source

or here a comparison of all the pixel shift cameras: CLICK TO VISIT LONG URL

their online widget is nifty, but I found I get more meaningful comparisons by downloading the RAW or JPEG files since the scaling in the widget blurs things. direct links to the A7 IV and S1R:

https://www.dpreview.com/reviews/image- ... 44.acr.jpg

https://www.dpreview.com/reviews/image- ... e8.acr.jpg

In my eyes the sony still clearly has a lot of micro artefacts and I much prefer the output of the S1R even if it's somewhat soft, but reacts well to sharpening (I could be biased of course because I bought one and it would really hurt to admit I wasted money

. workflow seems to be a pain for the sound at the moment too and speed is not that great either.

chris

Admin edit RJL: formatting (hide long URL)