[BigMackCam]

A7 is a nice camera and a huge update from K5 (except IBIS).

... and lossy compressed raw files that can result in high contrast edge artefacts and banding in large areas of gradual colour change (twilight skies, etc.). The MkII models onwards can be firmware updated to offer uncompressed raw, or else have the capability already. But the original A7 series can't. My own Sony A99-based Hasselblad HV has this same limitation, and whilst it doesn't affect a large number of photos, I *have* experienced the negative impact that lossy compression can have. So it's worth bearing in mind depending on the types of photography the OP is involved with...

[Rondec]

The A7 is a really old but cheap camera. Sony keeps it around to get people into their full frame cameras, probably realizing that many will be dissatisfied after a little while and end up purchasing something more functional. As others have said, if you are going with Sony, I would buy a newer camera -- A7 III or A7r II, both of which can be had for a decent amount. The other option would be to find a used K-1 which would allow you to use your K mount lenses without an adapter.

[yusuf]

... and lossy compressed raw files that can result in high contrast edge artefacts and banding in large areas of gradual colour change (twilight skies, etc.). .

It would be interesting to see visible evidence that A7x lossy compression files perform inferior to K-5. that's what OP is comparing.

[BigMackCam]

It would be interesting to see visible evidence that A7x lossy compression files perform inferior to K-5. that's what OP is comparing.

I can't show you A7 -> K-5 direct comparisons. But good evidence of the impact (or, rather, one potential impact) of Sony's lossy compression is in the following DPR article (and we all know how "pro" Sony the folks at DPR can be):

The Raw and the cooked: pulling apart Sony's Raw compression: Digital Photography Review

Click on that first photo, wait 'til it loads, then click on it again to zoom full size... then move around until you find the moon in the lower left of the image. It looks like this:

Look at those artefacts along the edge of the moon... That's caused by the lossy compression. I can tell you now, the K-5 would not have these artefacts because the raw files use lossless compression.

I wouldn't want to overstate the impact of Sony's lossy approach... In most photos, depending on what you shoot, it will never be obvious. But where you have high contrast edges against a smooth dark background, it can rear its head (long exposure star trails against a dark blue or black sky, for instance). I get the same effect (though I've never seen it quite this bad) every now and then on my Sony A99-based Hasselblad HV, which uses the same sensor and lossy compression as the original A7. Is it a huge issue? No, not for me at least. But it doesn't happen with Pentax DSLRs because compression of raw files is lossless.

---

Last edited by BigMackCam; 10-18-2019 at 03:44 PM.

[yusuf]

I can't show you A7 -> K-5 direct comparisons. But good evidence of the impact (or, rather, one potential impact) of Sony's lossy compression is in the following DPR article (and we all know how "pro" Sony the folks at DPR can be):

...

A7/A7R lossy compression is well debated and I am aware of that. My point was that does it make it worse than K-5?

A7 and A7R files, even with their lossy compression, carry more lossless data than K-5. Lossless data in K-5 will be 16x14bit, while it will be 24x11 for A7 and 36x11 for A7r, not to forget there are still more not-so-lossy data there. Also, the larger pixel and resolution will results in a less quantization error. So I doubt K-5 can perform better than A7/A7r despite later having lossy compression but would be interesting to see otherwise.

Being a previous K-5 owner, I certainly find A7/A7r files treat to eyes.

[BigMackCam]

A7/A7R lossy compression is well debated and I am aware of that. My point was that does it make it worse than K-5?

A7 and A7R files, even with their lossy compression, carry more lossless data than K-5. Lossless data in K-5 will be 16x14bit, while it will be 24x11 for A7 and 36x11 for A7r, not to forget there are still more not-so-lossy data there. Also, the larger pixel and resolution will results in a less quantization error. So I doubt K-5 can perform better than A7/A7r despite later having lossy compression but would be interesting to see otherwise.

Being a previous K-5 owner, I certainly find A7/A7r files treat to eyes.

Short of going through all of my A99 / HV files (which use the same lossy compression) to find my affected photos - there aren't many, but I have them - and then running a set of side by side comparisons taking photos on both my HV and K-5, I won't be able to convince you, and sadly I don't have time for such a time-consuming exercise at present.

All I can tell you is that I've experienced both high contrast edge artefacts (of which my previous post linked to one example, albeit quite an extreme one, in DPR's article) and banding / posterisation in large areas of smooth, gradient colour... enough that in one shot I recall having to use dithering to address the visual impact. Additionally, I've found shadow recovery in my HV's raw files to be noisy and lacking in colour information. That's just my perception, but it has made me wary of under-exposure to the point where I'm more likely to shoot multiple exposures and carry out an HDR merge. I've never had these issues with files from the K-5.

As you say, the lossy compression issue has been well debated. And there is plenty of photographic evidence online demonstrating the artefacts and posterisation issues.

When all's said and done, all that matters is you're happy with the A7's output. Indeed, I'm very happy with the output from my HV too - it takes wonderful photos. And I'm not saying the K-5's output is better or worse. All I'm saying is, Sony's lossy compression results result in irrecoverable loss of data that can have a visual impact in certain circumstances. Pentax's lossless compression does not.

[Rondec]

A7/A7R lossy compression is well debated and I am aware of that. My point was that does it make it worse than K-5?

A7 and A7R files, even with their lossy compression, carry more lossless data than K-5. Lossless data in K-5 will be 16x14bit, while it will be 24x11 for A7 and 36x11 for A7r, not to forget there are still more not-so-lossy data there. Also, the larger pixel and resolution will results in a less quantization error. So I doubt K-5 can perform better than A7/A7r despite later having lossy compression but would be interesting to see otherwise.

Being a previous K-5 owner, I certainly find A7/A7r files treat to eyes.

I think the question isn't really which is better, a K5 or an A7, but rather today, at the end of 2019 is the A7 a good camera to invest in. My strong feeling is that while you can work around an A7's deficiencies there are significantly better cameras out there for only slightly more money. I mentioned in my post a used K1, an A7 III or and A7r II, all of which I think would have less limitations than the A7 if someone is interested in moving from a K5 to a full frame camera. Yes, they are a bit more, but I think it would be worth saving to get something with fewer limitations.

It isn't surprising is it? This is a camera that was released in October of 2013 and for some reason is still being sold as new. Six years is a huge amount of time in the camera market and things have developed a long ways from then.

[yusuf]

Short of going through all of my A99 / HV files (which use the same lossy compression) to find my affected photos - there aren't many, but I have them - and then running a set of side by side comparisons taking photos on both my HV and K-5, I won't be able to convince you, and sadly I don't have time for such a time-consuming exercise at present.

All I can tell you is that I've experienced both high contrast edge artefacts (of which my previous post linked to one example, albeit quite an extreme one, in DPR's article) and banding / posterisation in large areas of smooth, gradient colour... enough that in one shot I recall having to use dithering to address the visual impact. Additionally, I've found shadow recovery in my HV's raw files to be noisy and lacking in colour information. That's just my perception, but it has made me wary of under-exposure to the point where I'm more likely to shoot multiple exposures and carry out an HDR merge. I've never had these issues with files from the K-5.

As you say, the lossy compression issue has been well debated. And there is plenty of photographic evidence online demonstrating the artefacts and posterisation issues.

When all's said and done, all that matters is you're happy with the A7's output. Indeed, I'm very happy with the output from my HV too - it takes wonderful photos. And I'm not saying the K-5's output is better or worse. All I'm saying is, Sony's lossy compression results result in irrecoverable loss of data that can have a visual impact in certain circumstances. Pentax's lossless compression does not.

You are ignoring the math - refer my previous post. It is technically NOT possible to have k-5 file with less info to have less artifacts than A7/a7r files having more info but I'll be happy to learn if you can prove it technically.

Again, the following is a simple calculation

K5 - 16MP*14 = 224 lossless Mbits
A7 - 24MP*11 + 24MP*3 = 264 lossless Mbits + 72Mbits lossy bits
A7r - 36MP*11 + 36MP*3 = 396 lossless Mbits + 108Mbits lossy bits

It is a fact that A7/a7r lossy compression having artifacts in some cases but it's worse only when compared against comparable cameras (FF, 24MP and above) but not against k-5.

[BigMackCam]

You are ignoring the math - refer my previous post. It is technically NOT possible to have k-5 file with less info to have less artifacts than A7/a7r files having more info but I'll be happy to learn if you can prove it technically.

Again, the following is a simple calculation

K5 - 16MP*14 = 224 lossless Mbits
A7 - 24MP*11 + 24MP*3 = 264 lossless Mbits + 72Mbits lossy bits
A7r - 36MP*11 + 36MP*3 = 396 lossless Mbits + 108Mbits lossy bits

It is a fact that A7/a7r lossy compression having artifacts in some cases but it's worse only when compared against comparable cameras (FF, 24MP and above) but not against k-5.

I'm not ignoring the math, and I appreciate you providing me with a simple calculation I can understand, though I'll admit I'm a little confused as to why total quantity of data - which is in Sony's favour due to the resolution of sensors - should be considered better than quality of data, which is lossless in the K-5. The Sonys are recording less lossless bits per pixel, or area, than the K-5 - no?

Let's put the K-5 aside for a moment and replace it with the K-3, which has a 24MP sensor. Using your simple calculation, the K-3 would record 24MP*14 = 336 lossless Mbits. The A7 would, as you've shown, record 264 lossless Mbits and 72 lossy Mbits. So for the 24MP both cameras are working with, the A7 is providing less data due to lossy compression - the same lossy compression that can result in artefacts, posterisation, and some loss of information in shadows.

But this is just simple math. Let's bring the K-5 back into the picture and compare the image quality to that of the K-3 (I own both). The K-3 wins on resolution and the fact that it has no AA filter. But otherwise - and this is borne out by many owners in these forums - the K-5's image quality holds up remarkably well by comparison. In fact, some owners here prefer the K-5 series to the K-3, especially where dynamic range is concerned.

I can't address your request to prove anything technically - that's not where my limited skills lie, sadly. You can choose to accept - or not, if you prefer - my claim that I have images where artefacts and posterisation have appeared, and shown unexpectedly poor results in shadow recovery, due to Sony's lossy compressed raw files... issues that would not have occurred with lossless compression on the K-5 (or, indeed, uncompressed raw on my A7 MkII). I'm not saying that makes the K-5 better, I'm not saying it makes the A7 worse - I'm just stating what I've observed in my own files.

It's worth noting that I wasn't even aware of the lossy compression - I'd never researched it; never even heard of it - until I noticed such artefacts in a limited number of my HV's photos, and started looking for reasons online. So it's not like I read about the issue then started imagining the effect. I saw it in my own photos first - just a few, but a few is more than none. And those artefacts would not be there without the lossy compression.

I will go back to the DPR example I posted, since that's all I have readily available. With or without the math, how would you explain the artefacts on the right edge of the moon? And would you expect to see those on any raw image from a camera that does not use lossy compression? I propose that you would not. Does that make the overall image worse than you could expect from the K-5? Not necessarily... it just means those artefacts wouldn't be present with the K-5 (or the K-3, or the A7 MkII set to uncompressed raw).

I'm going to leave the discussion there, as we can't move forward unless I (a) trawl through my photos for examples, then (b) set up a side by side shoot with my HV and K-5. With the greatest of respect, I don't have time to do that in order to prove a discussion point. I'm happy to accept that you're unconvinced.