HSBS/SBS vs 3d Bluray

Depth wise, it's the same. Format conversion like that has no effect on the 3D itself.

Quality wise is another matter.. compressing the file into a single image can result in the two sides not matching exactly, so a spot of lighting or something might be off.. brighter in one eye than in the other, and it can make things look a bit weird.

I disagree with Jedi partially. H-SBS maintains two images, but reduces horizontal resolution somewhat so that the resulting file is 1920 x 1080 with each frame of the 3D image becoming 960 x1080. This allows it to be compatible with MKV files on smart TVs. H-OU becomes two images of 1920 x 540.

If encoded right there should be no difference in 3D unless the compression rate is really high and the image quality overall drops too much on the SBS and OU. Then it will be a pretty dramatic overall loss of image quality across the board.

That said half OU and HSBS are both assumedly half the number of pixels in the original BR version and so will suffer some level of image quality loss over BR.

That said even if encoded wrong I don't think Jedi is correct... if it's not split properly down the middle the whole image will be off. This will cause some serious playback issues but it would effect an entire side of the image (or both sides realistically) so it's not like you will see a sparkle floating off on it's own somewhere, the image will overall just be hard to resolve between both eyes due to how it's spaced.

So short answer: No. A good HSBS or HOU version will have the same 3D effects as the BR version.

Most of the ones I've seen have been pretty decent, I haven't come across any that are out of sync or not split properly. But they're still no match for a hardcopy on Blu-ray. They tend to suffer more from compression than anything else. It seems to have a greater impact on 3D than 2D.

Absolutely..

In fact a compression free half ou video should theoretically be indistinguishable from a frame packed BR an a Passive TV (as all the lines thrown out to make a half ou are thrown out by the TV anyway when it comes time to display the frame packed).

But definitely compression artifacts (or any artifacts that differ from eye to eye) show up much more in 3D than they did in 2D because your eye has something to directly compare against.

An analogy migth be one of those "spot the differences" puzzles where two pictures have several slight differences.

If you look at one and then the other it can be hard to spot the differences, but if you lay one over the other, the differences will stand out quite easily.

In 3D your eyes are meshing two images over each other and expecting any differences to be resolveable in terms of parallax offset/3D information.

When a compression artifact comes across in a half split 3D video file one eye gets an unresolvable difference from the other and this just feels wrong and so stands out very strongly.

Anytime you squeeze an image and stretch it back out again as in SHS half you risk generating an artifact in one eye's image and not the other. But in practice, I have never noticed this as being disruptive to the 3D or image quality. If your TV can handle it ( most can't) SBS full or OU full should have no difference in artifacting vs. frame packed.

It depends very much on the codec used and gets worse with lower bitrate but there is a very solid reason you could get different compression artifacts in both eyes: Both eyes images are different.

If the images are different (which they will be in any scene with depthy) even the same algorthym will compress both sides sliglty diferently. This can result in artifacts however at higher bitrates and resolutions these aretifacts should be too small to be easily perceptable.

Agreed and this is key to the overall look of the video. It is always important to use your best codecs if quality is important and then those artifacts that cause left - right differences will not affect overall visual in 3D.

However,
I once experimented with an example of left right image differences that were major. Here I did two tests. You may wish to experiment in a similar way.
I intentionally took a right image and modified it's color , contrast, and brightness then paired them and observed for what difference was in 3D.

The color did cause a bit of problem as colors blended to a new hue except where the images diverged and here the ghosting of double image outlines surfaced. In contrast and brightness differences, the same ghosting happened but the objects all just looked additive in contrast and brightness. e.g. a pitch black became a dull gray. Note in my test I made obvious radical changes that were well outside the normal artifact range.

Where this issue did become critical had to do with the left right image differences that were actually captured during recording. For example, I tested with glare elimination using circular polarizing filter. In this case a different level of filtration was recorded for left and right lens because the filter was in a different setting due to each camera lens being off center to the single filter rotation on each lens. Glare was never completely eliminated but worse the glare was recorded with different depth than the object generating the glare. Like a windshield, the glare was no longer on the glass but either under it or above it. This was weird and I found no easy solution so I chose to avoid using any filter that could not function equally for each off center lens. UV and ND filters work great, Gradient filters can work in horizontal rotation alignment but not if rotated at an angle. Polarizing filters don't work at all for 3D where one filter is used for both lenses. It works fine when each lens has it's own filter adjusted individually.

Finally, anytime any object is present in one image and not the other they will combine when paired and the result will display in 2D at the screen plane. This can be a problem, or not with occlusion in the paired images.