MattH

There's no chance that the GH5 promos will look like upscaled 720p.

Exactly. Thats the response to people who say 4k isnt everying. I actually agree with that, but its only true as long as the full HD is actually full HD quality. I think we both know that that isn't going to be the case. To be truthful I never expected this to have an increase in quality over other canon apsc stills cams anyway. I can see someone messing up sharpness and messing up the dynamic range. But I think the perceptive resolution pretty much is what it is.

Lackluster HD when viewed on a HD monitor. In 2016. Come on man. They've got dual pixel going for it, but for me its so close yet so far.

1/50th will give you a smoother look. People tend to blurt out the 180 degree shutter rule, but that rule was invented for use when the playback speed was always between 24 and 30 fps. Really it should be called the 1/48th to 1/60th rule. If you shoot 1/50th at 50p you are still going to get the same amount of motion blur per frame as you do shooting 1/50th at 25p just with more frames. I have to say I haven't done a lot of research on this myself, but basically any 50 or 60p video I see I HATE it because of its staccato look. Which seems counter productive because you would think it would be smoother. Perhaps it is just the look of high frame rate, but I suspect it is strongly related to the fast shutter speeds people are using when they follow a silly catch phrase 'rule' like lemmings. I would do a few tests. Shoot the same scene, maybe a person walking or a car going past, using both 1/50th ant 1/100th and see which one looks smoothest.

If this can deliver good very good 1080 (Better than the 80d which is better than normal but still looks blury compared to downscalled 4k) and it has the HDR mode that the 5D mk4 then it will have my interest. If not then I don't see the point. I like that there are some manual dials now. As for looks. I think the body looks good, though the hump doesnt fit. It should be more angular and sleeker like the top plate of the body. Though the fact that it looks more like a DSLR will mean that it sells better in the west now I think.

*It's on a go-slow

Amount of rolling shutter, quality of interframe compression (or lack of ideally). They effect the perception of motion in video.

Depends on how you calculate it. As you are aware, the GH2 has a multi aspect sensor. So it maintains a diagonal crop factor of 2 at every aspect ratio. So speaking in terms of stills, no, with the speedbooster XL you will get 1.28 crop factor at every aspect ratio in comparison to 3:2 ratio full frame (135 format). If, however, you are comparing with respect to a 16:9 crop of full frame then yes the video mode of the GH2 has a 1.91 crop factor, and thus with the speedbooster XL the crop factor will be 1.225 Whether this makes the GH2 more appealing if you don't already have one is another issue altogether.

Not from what I've seen of the issue. Highlight aliasing, this problem, is when you get black outlining against clipped edges. Magenta coloured edges is generally a lens issue. Though there could be another sensor/processing issue other than this one. Whether you call either one highlight aliasing Im not really bothered. I think 'highlight outlining' might be a better term for this issue anyway.

Why? Why would you not want it all in the GH5 for GH money?

I'm glad you have come to this realisation as this is what I was saying for years. (However there is big exception that I have come to realise). As far as comparing field of view what sense did it ever make to compare a 16:9 video mode with the 3:2 full frame. By that reckoning even full frame video has a crop factor over full frame. Much easier just to use the horizontal. That way you know that a video mode with a 2 times crop factor with a 25mm lens will give you the same field of view as a 50mm on a full frame camera in video mode. If you used the diagonals of the sensor this could be out. One slight exception to this is DCI 4k as technically it is a different aspect video mode. By the book you should use a diagonal when comparing a 16:9 video mode to a 17.66:9 video mode. But its a lot of work for such a small difference. And as long as you say 'horizontal crop factor’ it’s technically accurate anyway. This is the big exception: One purpose that you need to use diagonals for is lens coverage. Because the diagonal will always bisect the centre of the image circle: the diagonal is the diameter of coverage that is needed for that image size. So when trying to work out whether a speed booster and lens combination will work on a particular system the diagonal crop factor should be used. Case in point: Full frame lenses are designed at least to cater for the diagonal of 3:2 full frame which is 43.26mm. The diagonal of full frame cropped to 16:9 is 41.3mm. This is where full frame video's crop factor of 1.047 over full frame actually matters. It means that you could use a full frame DSLR lens with a 0.955* speed booster on a full frame mirrorless and it would still fall within the full frame image circle when shooting video (but not 3:2 stills). Failing to use diagonals when doing this could make one over cautious or over zealous depending on what ones intentions are. (buy the way 0.955 is the reciprocal of 1.047. If you ever work out one of these and you want the other, just press the 1/x button on the calculator)

That would be the biggest buzzkill ever. Like handing a kid an untied balloon. I'd rather it not have 10 bit than that. At least you wouldnt be constantly torn. Unless the HD was downscaled from full sensor readout. I wouldn't mind that then.

Exactly, you could get wide angle with a sigma 8-16 or a tokina 11-16, but then you are still going to have to buy full frame lenses to cover that focal range. All that money for all that hassle. Doesn't draw me to it.

To be fair, it's less of a crop than four-thirds. But also to be fair, you're paying full frame prices for it. (edit: Written before reading the article that already makes this point)

It's a well known artifact with sony cameras. Referred to mostly as 'highlight aliasing'. Google 'highlight aliasing' and you will find many hits. It is a lot more egregious in HD than it is in 4k but its clearly still there. Aparently It was even apparent on the f5 and f55 but was addressed and removed on those cameras. On other cameras like the fs700 it wasn't addressed. However, people observed with the fs700 that it could be bypassed with Raw recording. Go and look at videos shot with sony cameras. Now that you have seen it you will notice it everywhere! Even in cases were you can't see it obviously it still alters the overal look of the image as a whole, making it seem over sharp and less organic. Its a big contributor to sonys reputation as giving a video look. The good news for you is that it is smaller in 4k and in most cases people wont notice it.

Dont expect free speech from Dpreview. I was temporarily banned from posting in the forum. No reasoning given. Presumably because I posted a comment that insulted the week mind of one of the mods.

Do you have any examples of the image with this converter installed. I cant seem to find any online.

The C line of cameras have log because they are 'cinema' cameras. The stills cameras dont. It's just that simple unfortunately. The people who implement video features will probably have nothing to do with still cameras. If they bring out a 1DC mk ii, that will have log.

Just so it is borne in mind: The shroud is cosmetic on the 17-55, but on other lenses it is not. If you removed it on the 10-18 there is a very high chance the rear element will smash into the optics of certain speed boosters when zooming towards the widest setting. As for the sensor stack, that is certainly something to think about. Something doesn’t seem to add up though: If the bmmcc has a thinner sensor stack and will therefore not work optimally with a speed booster designed for a regular mft sensor, wouldn’t the bmmcc therefore not work optimally with native mft lenses either?

Good question. I think it is a relevant question because it is a real choice you have. You only have one lot of $3000 so cant get both and you don't have a camera so you have to get something and you are free to get anything. I cant really say much extra about the NX1. The fact that you are considering it shows that you don't care that Samsung is pulling out of the camera business. That is something that may bother some, including myself, but it depends on how you look at it. I suppose you would have to think of it as getting a fixed lens camera as you know you wont be able to use native lenses on anything else in the future. You cant with fixed lens cameras either but people buy them. The bmmcc part of the question is really a subject in itself. A few things I can think of that may be of note. You say the 24mm 2.8 is the widest lens with IS. Its the widest prime with is, and its actually the joint widest EF lens with IS. Its not, however the widest ef-s lens with IS You've got the 18-55,18-135, 15-85 and the widest the 10-18. And that appears to be one big advantage with the Aperture DEC LensRegain over the metabones XL, that it seems to have enough room for the protruding rear element protector of EF-S lenses. (Although I haven’t been able to find confirmation of this yet) If this is the case then the 10-18 f4.5-5.6 would become a stabilized 7.5 - 13.5mm f3.2-f4 Giving 21.6 - 38.9 equiv on the bmmcc. And 15-27mm equiv on a full mft sensor. That's one big advantage over the bmpcc specific metabones as well, you can use it for still photography and any other mfr camera in the future. Its true that having the bm view assist would be overkill in the sense that the bmmcc already has a recorder, but the good thing about it is that it would enable you to easily switch to the micro studio camera if you needed resolution/even lower rs over dynamic range in any shots. And of course, any other camera with 4k output. Back to the overall question. I think you have to remember that the bmmcc gives 10bit all - i frame prores with true HD resolution. If it weren’t for 4k being all the rage right now we could appreciate how amazing that is. All in all the NX1 would give you a self contained system which doesn’t require many extra purchases but cant be upgraded in the future. And The bmmcc route would give you a modular system with a lot of choices, flexibility and future possibilities beyond the bmmcc.

For anyone who is interested in rolling shutter performance I have done multiple measurements on the panning shots in Gordon Laing’s video and determined that the rolling shutter skew is about 65% which means a readout speed of around 27ms. Of course its not an exact science because there is no vertical object that goes from the top of the screen to the bottom and there could be acceleration or deceleration from one frame to the next, but all of the results were coming up around this area. I’m sure Samuel H will eventually come out with an ‘official’ figure after doing the exact same thing and proclaiming a number that is for all intents and purposes the same. So people are free to wait for that, but for anyone who wants to cut to the chase: 25 to 30 ms is your ballpark figure.

Wow! So only 1 frame a second is a real sample? That's bad! And just think, your NLE doesn't know or care which 1 frame is the I frame, so when you render out it will mostly chose P frames to use as I frames in the output file. That would probably account for all the macroblocking and mush we can see in the footage, especially on the foliage and tree trunk in the forest shots. No wonder they are trying to force videographers to record externally. So basically, as far as I'm concerned, as well as adding on the cost of the battery grip as an essential we should also add the cost of a 4k recorder as an essential if we don't already have one.

Don't be so defensive then.

Sorry, you brought me to this. All I can say is: DUMB ASS! I answered the question you quoted.

It's funny how forcefully you declare that you don't get something. All in all I agree with getting this back on topic.