Using "clean" HDMI out from A5100/A6000

M. Gilden said:
Seriously. I feel like Sony and others have sort of dismissed it, saying "just press record in camera, problem solved".

Feel free to click "me too" on my request to get this fixed on Sony's website:
http://community.sony.com/t5/Join-t...oothing-so-it-doesn-t-show-up-in/idi-p/538535
Click to expand...

Yes, I'll do it!
If at least the output in manual mode was accurate in terms of aspect ratio and good image quality, without moire' and so forth, that would be already something.
But, man, not even that. They have to fix this.
 
dsp_418 said:
Yes, I'll do it!
If at least the output in manual mode was accurate in terms of aspect ratio and good image quality, without moire' and so forth, that would be already something.
But, man, not even that. They have to fix this.
Click to expand...

Thanks! Spread the word if you know anyone who might be effected. I want Sony to take notice!
 
Looks like using clear zoom takes away the skin smoothing. Whaddya know.

Stinks because we're adding a minimum 1.1x crop to the already 1.5x crop... that makes it an effective 1.65x of full frame. Considering a Canon APSC camera is 1.6x, this isn't the worst option, but it still shouldn't be necessary! But for anyone with this setup that is stuck banging their heads on a wall, this could be a solution.
 
That sounds like a pretty effective solution, actually. It sucks that you have to trick it this way instead of just telling it how to work, but hey, problem solved!
 
Samuel H said:
That sounds like a pretty effective solution, actually. It sucks that you have to trick it this way instead of just telling it how to work, but hey, problem solved!
Click to expand...
It sucks because I shouldn't have to negatively effect the image elsewhere (further crop) just to fix it here. But yeah, its definitely something.

Here's the thing- I notice that using the rocker to enter clear image zoom changes some other things on the display- metering and focus method icons change. I'm wondering, maybe its not the clear image zoom that is disabling it, but something else that the camera is doing as a result of clear image being active?

I have to do some more research into this... does anyone know what clear image zoom actually does to Sony cameras when enabled?
 
Ok, another quick update (if anyone is still following this):
Apparently the reason clear image zoom works, is that face detection is unavailable when using any kind of zoom other than optical. From the manual on the A5100 (http://docs.esupport.sony.com/dvimag/2014/ILCE-5100_guide/en/contents/TP0000301886.html):
You cannot use the Face Detection function with the following functions:
  • When using zoom functions other than the optical zoom
  • [Sweep Panorama]
  • [Picture Effect] is set to [Posterization].
  • When using the [Focus Magnifier] function.
  • [Scene Selection] is set to [Landscape], [Night Scene] or [Sunset].
Click to expand...

Sweep panorama doesn't help me, neither does posterization or landscape/night scene/sunset. BUT... Focus Magnifier sounds promising! If I'm using an Atomos Ninja as my monitor, I can just have the focus magnifier turned on the entire time to disable the skin bug. The screen on the camera will actually double as a constant zoomed-in focus aid, while the monitor will be for composure and framing! This won't work well for streaming boxes at live events or other HDMI uses, but at least if you have an external monitor, it could work?

So far, I touched the screen on my A5100 which is supposed to enable focus magnificiation, but the bug still exists. Even in the magnified view! So much for that one. Maybe I'm doing it wrong, but so far the only thing that actually makes a difference in movie mode is to touch Clear Image Zoom.
 
So stupid. This crap is on par with the ridiculously strong noise reduction being forced on in video mode. Absolutely destroys image quality at higher ISOs. Also, the motion compression is absolutely awful in this camera even in XAVCS. Just wave your hand really fast and look at one of the frames. Terrible. Then there's the ridiculous compression in the shadows which gives you macro blocking city if you try to push them up. Bleh, Im frustrated with my a6000.
 
JoeThePro said:
So stupid. This crap is on par with the ridiculously strong noise reduction being forced on in video mode. Absolutely destroys image quality at higher ISOs. Also, the motion compression is absolutely awful in this camera even in XAVCS. Just wave your hand really fast and look at one of the frames. Terrible. Then there's the ridiculous compression in the shadows which gives you macro blocking city if you try to push them up. Bleh, Im frustrated with my a6000.
Click to expand...

I'll be honest, I've seen complaints like this elsewhere on the forum (possibly also from you?) but I haven't noticed anything to get concerned about. It didn't seem any worse than any other camera in this class, to be honest.
I may be out of my element here, but I'll propose a possible reason for what you are seeing: The a6k and a5100 don't do a full sensor readout, I think. It's pretty respectably clean for pixel binning, but no where near the A7S in sharpness and IQ from movie mode. And that's to be expected since we are trying to shove a ridiculous amount of pixels into 1080p in real time. So aside from the expected loss in resolution comparing the detail from a photo vs a still from a movie clip, the movie clip is going to also appear to have some sort of smoothing of detail and less noise... but not because of strong noise reduction, rather because the pixels are being resampled. This will be true of any moderately high MP sensor binning down to HD video. Heck, 4K video scaled down to 1080p will appear to have less noise for the same reason. It doesn't mean there is noise reduction being applied.
As for motion compression... Aside from judder which is again sadly par-for-the-course, I didn't notice anything like that with my A7S in XAVCS. My A5100 has been mostly recording externally and bypassing the codec, so I can't really vouch for it there, but the XAVCS didn't seem to have much problem there either. I haven't compared the A6k's XAVCS since it got it in that update, I don't have any reason to think its any worse, but maybe it isn't the same implementation? Does it do it in AVCHD too?

To be perfectly honest, I'm actually really impressed with the A5100 as a little APS-C video camera. The quality of the video, even with the internal XAVCS codec, still overall looks better to me than my Canons or old GH2s I used to use as a B-cam. The only caveats I've found are that the body overheats if recording internally for too long (fixed with an external recorder), and that this #@$&^ing SKIN SMOOTHING BUG WON'T GO AWAY. Seriously, if they fix this one issue, I'd have no qualms continuing to use and recommend this camera.
 
M. Gilden said:
I'll be honest, I've seen complaints like this elsewhere on the forum (possibly also from you?) but I haven't noticed anything to get concerned about. It didn't seem any worse than any other camera in this class, to be honest.
I may be out of my element here, but I'll propose a possible reason for what you are seeing: The a6k and a5100 don't do a full sensor readout, I think. It's pretty respectably clean for pixel binning, but no where near the A7S in sharpness and IQ from movie mode. And that's to be expected since we are trying to shove a ridiculous amount of pixels into 1080p in real time. So aside from the expected loss in resolution comparing the detail from a photo vs a still from a movie clip, the movie clip is going to also appear to have some sort of smoothing of detail and less noise... but not because of strong noise reduction, rather because the pixels are being resampled. This will be true of any moderately high MP sensor binning down to HD video. Heck, 4K video scaled down to 1080p will appear to have less noise for the same reason. It doesn't mean there is noise reduction being applied.
As for motion compression... Aside from judder which is again sadly par-for-the-course, I didn't notice anything like that with my A7S in XAVCS. My A5100 has been mostly recording externally and bypassing the codec, so I can't really vouch for it there, but the XAVCS didn't seem to have much problem there either. I haven't compared the A6k's XAVCS since it got it in that update, I don't have any reason to think its any worse, but maybe it isn't the same implementation? Does it do it in AVCHD too?

To be perfectly honest, I'm actually really impressed with the A5100 as a little APS-C video camera. The quality of the video, even with the internal XAVCS codec, still overall looks better to me than my Canons or old GH2s I used to use as a B-cam. The only caveats I've found are that the body overheats if recording internally for too long (fixed with an external recorder), and that this #@$&^ing SKIN SMOOTHING BUG WON'T GO AWAY. Seriously, if they fix this one issue, I'd have no qualms continuing to use and recommend this camera.
Click to expand...

Yeah it was probably one of my comments (or multiple... :D ) about it that you are thinking of. Im just so frustrated with this stuff because I see the potential this cam had for video. If I compare directly to my GH3, which had NOISE REDUCTION CONTROLS (AAHHAHHDSHFSKFHUIB SONYYYYYY) IN VIDEO MODE, the mid and shadow detail is much better. It doesnt smear it in the shadows at mid-high ISO like the a6k does, and the motion compression is much better. It IS the codec, for sure. I remember on personal-view.com I think Nick Driftwood even mentioned that the B frames in the IPB compression are a good bit smaller in size on the a6k than they are on the GH3. This alone will kill motion detail.

The smoothing Im seeing on the a6k is absolutely not due to scaling, it is plainly noise reduction. Just shoot ISO 3200 in a dimly lit area and look at the gloopy clumpy crawlies everywhere. Its nearly impossible to clean this up well with Neat Video, whereas on the GH3 you at least had grain (and tons of it) to work with in post. Since the a6k is a better noise performer than the GH3, it kills me that they force this NR on us when it could be so good without it...

BUT... this is a very inexpensive camera, so I guess I shouldnt expect it all. Hopefully the a7000 (when the hell will this be announced!?!??!) fixes some of these issues. Oh wait no, they just focused on getting 30+ useless MP in there. Great....
 
JoeThePro said:
Yeah it was probably one of my comments (or multiple... :D ) about it that you are thinking of. Im just so frustrated with this stuff because I see the potential this cam had for video. If I compare directly to my GH3, which had NOISE REDUCTION CONTROLS (AAHHAHHDSHFSKFHUIB SONYYYYYY) IN VIDEO MODE, the mid and shadow detail is much better. It doesnt smear it in the shadows at mid-high ISO like the a6k does, and the motion compression is much better. It IS the codec, for sure. I remember on personal-view.com I think Nick Driftwood even mentioned that the B frames in the IPB compression are a good bit smaller in size on the a6k than they are on the GH3. This alone will kill motion detail.

The smoothing Im seeing on the a6k is absolutely not due to scaling, it is plainly noise reduction. Just shoot ISO 3200 in a dimly lit area and look at the gloopy clumpy crawlies everywhere. Its nearly impossible to clean this up well with Neat Video, whereas on the GH3 you at least had grain (and tons of it) to work with in post. Since the a6k is a better noise performer than the GH3, it kills me that they force this NR on us when it could be so good without it...

BUT... this is a very inexpensive camera, so I guess I shouldnt expect it all. Hopefully the a7000 (when the hell will this be announced!?!??!) fixes some of these issues. Oh wait no, they just focused on getting 30+ useless MP in there. Great....
Click to expand...

I hear you, but I'm still not convinced you're seeing noise reduction. The GH3 has less megapixels and a different resample method, I think. My GH2's image was definitely sharper than the A5100's, that thing just had a magical sensor readout. Better than the GH3, some have said. But the Sony image is far more pleasing to look at (DR and color), so I actually just sold all my Panasonic gear in favor of a Sony-centered setup. Plugging this into my Atomos, I'm getting a decent image aside from the waxy skin thing. Is it a poor codec implementation you're seeing? Maybe your picture profile is losing the detail? I don't know what to tell you man, I don't think its an opinion shared by other Sony shooters.

Meanwhile, what I'm discussing in this thread is clearly a bug.
 
M. Gilden said:
I hear you, but I'm still not convinced you're seeing noise reduction. The GH3 has less megapixels and a different resample method, I think. My GH2's image was definitely sharper than the A5100's, that thing just had a magical sensor readout. Better than the GH3, some have said. But the Sony image is far more pleasing to look at (DR and color), so I actually just sold all my Panasonic gear in favor of a Sony-centered setup. Plugging this into my Atomos, I'm getting a decent image aside from the waxy skin thing. Is it a poor codec implementation you're seeing? Maybe your picture profile is losing the detail? I don't know what to tell you man, I don't think its an opinion shared by other Sony shooters.

Meanwhile, what I'm discussing in this thread is clearly a bug.
Click to expand...

This is just a case of catching me at a moment of frustration. I think Im making a bit much of it. I was just playing with some iso 400 footage and the shadows retained a fair bit of detail when I brought them up. Its definitely usable, just not to the extent the GH3 was I guess. I still think it is NR that Im seeing making it extra soft as the chunks are simply too big to be attributed to scaling averaging, but thats just my thoughts. It looks similar to high ISO jpeg NR to me. Again, this camera is a fairly decent noise performer, nothing special, but not bad, so its usable. Just look at some low light shots from Nikon's D5300 though, that camera is CLEAN.
 
I also saw some of those codec issues back in my NEX-5N times. I never thought they were a lot worse than the Canon ones at the time, but they were there (they were just different: Canon had worse macroblocking in skin, Sony was better overall but could go crazy bad at some point). Once I moved to XAVC-S, they went away.

And yes, NR at high ISO creates ugly big noise and loss of detail, but that's always been like that on all the cameras I have had, so...
(I never liked any Panasonic cameras: sharpness and detail are impressive but DR is always too low)
 
Samuel H said:
....(I never liked any Panasonic cameras: sharpness and detail are impressive but DR is always too low)
Click to expand...

It does seem a tiny bit lower, but being able to underexpose for the highlights and then fairly cleanly boost the shadows almost 2 stops definitely helped!
 
There's still quite a big difference: GH4 a7S (shots taken at the shutter speed that makes the black-box-to-table boundary on the shadow barely visible, so, changing it so I have the same headroom in the shadows in both cases; the a7S shows a lot more range in the highlights)
 
Samuel H said:
There's still quite a big difference: GH4 a7S (shots taken at the shutter speed that makes the black-box-to-table boundary on the shadow barely visible, so, changing it so I have the same headroom in the shadows in both cases; the a7S shows a lot more range in the highlights)
Click to expand...

Interesting comparison. I assume this is Rec 709 vs Rec 709, not log, right? I do like the GH4 colors better though, they seem less "video-ish".
 
Samuel H said:
There's still quite a big difference: GH4 a7S (shots taken at the shutter speed that makes the black-box-to-table boundary on the shadow barely visible, so, changing it so I have the same headroom in the shadows in both cases; the a7S shows a lot more range in the highlights)
Click to expand...

My experience has been the same as Samuel H's. However, I'm not just comparing to the A7s... the A5100 (and a6000) are not as a sharp and don't offer the same image quality as the A7s, but they are close enough to be a B-cam especially for multicam event coverage. And while my GH2 videos were definitely sharper (especially with the hack), I still find the Sony footage more pleasing to look at, without being as soft or M&A ridden as Sony footage.
 
You must log in or register to reply here.
Back
Top