AC130: Video footage of the Auto Focus in action

[ajcourtney]

Get real Rod. I've got nothing to gain by advocating the use of ANY camera or ANY equipment. I'm just presenting an alternative viewpoint that ABSOLUTELY is on point.

Anyone looking to offload their 130's or 160's because of defective autofocus, please PM me.

 

[Rod Pike]

My comment was to attempt to keep the topic in focus (ha! how appropriate) . It was not my intention to offend you in any way, but please let's not get personal or patronising, this is not the place, or the spirit of this excellent forum.- I am very real and we are engaging in an extremely fruitful discussion on an issue that concerns us. I don't believe I suggested that you had anything to gain in any way.

We all have our (sometimes) differing points of view on this subject. My point was that people who actually own these cameras and work with them have identified a concern- and not in an isolated way.

It is being very actively discussed here, sometimes from different angles and viewpoints . But these users are also putting personal effort in to develop factual data and trialling, rather than opinions and we are all gaining from that (I certainly am anyway!) . The topic is the camera and this function which concerns many. I was only concerned that we might digress.

 

[Barry_Green]

we are in substantial agreement except for some physics nit picking.

I see exactly where you're coming from now. Yes, of course the physics of the issue could be solved by engineering, but then we have to keep in mind the cost and size of the resulting solution. So there's the theoretical "what could we do" and there's the practical "what could we do" and those aren't the same thing. What I was referencing is that the task of focusing the AC's lens is a lot more difficult than the task of focusing the HMC150's lens, because of the size, the range, the shallower DOF, and the higher res of the chips making the task that much more demanding... unquestionably a system could be designed that could execute the focus task absolutely identically to the HMC150, but -- at what cost, and at what size? Would it be practical at that point?

Our only difference is that I believe focus performance is actually limited by engineering and economics rather than physics.

Agreed at this point, the physics of the lens have made the engineering more difficult and the economics more pricey, and therefore there's a limit to how much they can or will do.

I just switched to manual focus because I could see that it would not be fair for any camera to focus on a set of dancing grey cards.

Exactly -- that's just plain reasonable. There are some things that you cannot expect an automatic function to handle, and you chose wisely there!

Thanks for participating!

 

[ELN614]

More videos

More videos

First video:show you what I shoot. Typical moving horse at typical speed and my typical framing. It takes concentration to follow and keep it framed just right. Try to fiddle and move your hand and you can easily lose your shot.

.
.
.
.
(another childish post) Next is for ajcourtney with 31 years of experience suggests

... Regarding iris pulls, IMO there is no need for that unless you are transitioning from massively changing lighting conditions, like from interior to exterior.I have no problem seeing temporary reductions in ambient lighting conditions due to passing cloud cover. That is exactly what happens with our own eyes, so it's not an unnatural occurrence. What i find absolutely maddening is fluctuations in background exposure due to the autoexposure being fooled by changes in framing composition. And the only way I know how to prevent that is to lock off your exposure with a grey card, which is pretty much what I do for every shot, even run and gun...

well below is what happens when you let the exposure to stay put- suggested by ajcourtney. (shot specifically for aj)

I think everyone saw this coming a mile away, except aj. Just a simple cloud cover can ruin your image when the sun decides to pop out.
.
.
.
NEW VIDEO BELOW
Last clip is the main reason for this whole thread. Panasonic's flaky Auto Exposure. I have three clips in this video
Disclaimer: I didn't add any sharpen (sorry I didn't want to alter the original) Plus I just uploaded the file and let youtube make the flash file.
1) Is what I normally get 90% of the time decent acceptable focus well within the physics of lens auto focus
2) Is the unpredictable focus going wacky.
3) Focus is in, then out, and then it goes back in.

If it is physics -- then I would never get any shots in focus. And you clearly see shots in focus. I have 60 hours to prove it. = Proof that it is not physics. It is a flaw.

 

[Rod Pike]

ELN614 Those focus lock shots are grim, and exactly the random event I am seeing. Only yours are in perfect conditions with high contrast high lux and moving object in a repeated position ( I checked the horse position with an on screen square). My examples are with lower lux (but well within limits) lower contrast and static subject- but same result.

We have now seen this repeated in all light conditions and subject matter. What we are showing is a camera that cannot duplicate consistently, identical cycles. I agree with ELN614 and just cannot see how this can be physics, dynamics or anything else on that principle, as they are laws and are constant, given common parameters.

What is clearly demonstrated is that the camera, occasionally in its search for focus, gets stuck. I observed near subjects getting stuck on infinity but other duplicate cycles were fine.

Many here are also professionals and we know how to work round these issues: But this camera has an auto mode that quite frankly cannot perform consistently at the basic level- see above example. Way back in the early part of this debate Karl produced some work which showed that going from a near object to a far one (the church/ anchor scene) the camera could not find infinity without a jiggle-again in high lux ,high contrast daylight. This is not the King's new clothes: we, the users can see it doesn't work from time to time and that is not what we paid for.

 

[ajcourtney]

(another childish post) Next is for ajcourtney with 31 years of experience suggests
well below is what happens when you let the exposure to stay put- suggested by ajcourtney. (shot specifically for aj)

Thanks for the education ELN614! What is that - at most a stop difference. Yep completely ruined. Here's an idea, open up scopes in your NLE, split your scene, apply a levels plug to the overexposed part, and drop your output and gamma to match the original exposure. Should take all of 2-3 minutes of your time.

Let me know when you want to sell your defective 130.

 

[ChrisHarding]

Hi ELN614

It certainly looks like you have an issue here much the same as Rod ..What you do is now up to you!! If Panasonic come forward and offer to fix it (which I probably doubt as Jan has already told us why it happens and it's something you might have to live with) then that's great. If Terry kindly does the same sort of test with his HPX250 and it doesn't show up the problem then you might have to consider selling the 130 and getting an HPX and living with DVCPRO format, but that might not suit you.
Maybe the 130 is just not the right camera for the job so then you need to find the right camera. From what I have seen it appears that a standard 12X zoom just won't be enough for you so the burning question here IS there another camera, with a 22X zoom that will autofocus the way you want it. I wouldn't beat myself up over the camera not doing what you want it to. The 130 and 160 ARE good cameras with lots of satisfied users but maybe your application is not suited to it. I shoot weddings solo and I had no option but to on-sell my 130's because they clearly were not designed for the task I set them...images were nice but they just were not the right tool for the job so I sold them and went back to 1/4" chips and 12X zoom (I don't need 22X) with another pair of HMC80's ...they work perfectly for me and (just my opinion) I think the footage has more WOW factor.

What did you use before the 130??? Is there something else you can use...even a brilliant camera like the 130 can easily be the totally wrong tool for the job and you have to cut your losses (I lost $2000 on my sale of both) and find the right tool. At least you will then have peace of mind and great footage and won't have to stress big time about fuzzy video.

Chris

 

[Mike Warren]

If it is physics -- then I would never get any shots in focus. And you clearly see shots in focus.

That sample B clip certainly looks like a problem. Again, I feel the algorithm is delaying any action because it's waiting for more stability in the shot. It's slightly less stable than the proceeding shot. They do have to balance several issues and it looks to me like they've gone too far on the anti-hunting part of the algorithm.

But Panasonic are aware of all the complaints and it seems they are unable to make it better. They even did what they could with the revision A to improve the workaround options available. That, to me says that if they could make it better they would.

So I don't really know what you can do if you need reliable auto focus except what Chris did, move to something else.

But I've never found an auto focus system to be even close to 100% reliable and therefore wouldn't trust it for any shot I couldn't redo.

 

[ChrisHarding]

Hi Mike

If you are talking smallish chips and wide angle and plenty of light I think AF is pretty reliable..at wedding photoshoots I have no option but to run the cam on full auto since it's on my stedicam sled..then again, that's at full wide, nice light and 1/4" chips so it wouldn't have to work very hard. Is there a camera that has reasonably good autofocus that would work at 22X zoom??? or are we clutching at straws here ...my HMC82's are pretty consistent at 12X zoom which is max and quite repeatable too BUT I would think that 1/3rd full HD sensors and 22X would be a fairly stern taskmaster. I wonder if anyone else here has a big zoom camera with 1/3rd chips that can honestly say that AF is perfect at full zoom? I'm sure that if Pansonic could do anything about it they certainly would have by now.... Having shot Panny for the last 23 years I have a lot of confidence in the company (still love my cams to bits too) but this might just be a physics issue that no camera can achieve.

Chris

 

[ELN614]

Chris, last year I used my SD camera DVC30 16x lens (clients were begging for HD and I told them as soon as a Panasonic comes out with a 18x lens or better - we'll make the switch.. I bought the AC130 at the end of Oct 2011- when it came out.)
The issue or problem boils down to the need to own an extra camera. I hate the idea of my company's shelves looking like a camera store. I already have 10, I don't need 11. heck I sold off more than 6 cameras in the last few years to have space in my place.
I use the AC130 for Depo work which works fine. I run everything on manual. Actually, pretty much all other work I run on manual except for a few odd jobs. I would sell the 130 for the new JVC now coming out but it does not do time date stamp. The plan is to purchase the new JVC and hope the older JVC, listed below, sells. I will be meeting with my JVC Rep and plan to send a very compelling letter to JVC to explain how they are missing out on a lot of sales just for the fact their cameras no longer have the time date stamp. I hope they will revamp the firmware in the camera and have it out by NAB2013.

Even though Karl's tests and others -- deal with going from far object to near object. (based on my first video) I am in fact staying on my object all the time. Only variable is zoom and distance the object is from the lens --- which is the whole point of auto focus -- to focus on an object which is continually changing distance.

And a footnote: Barry I have not forgotten about you. You requested SD footage. I need to sift through my SD footage from the 130 for focus issue footage and post it. (I hate the idea of spends several hours pulling clips to beat the dead horse. no pun intended.)

 

[ChrisHarding]

Fair enough!! I see your point entirely

However based on the fact that the DVC30 (I never got that far as I had two DVC20's) has something like a 460K x 3 effective pixels compared to over 2000K x 3 on the 130 and your zoom factor has gone from 16X to 22X ..I can see why the camera struggles with focus based on what Jan and Barry say (I said I see but I didn't say I like!!!) Now I bet the little AC7 with 16.8X zoom has a much better autofocus at full zoom cos it only has one 1/4" chip. It's not your fault at all and your tests are in fact correct...Just my opinion but Panasonic obviously didn't do autofocus tests at full zoom at the pre-production stage ..if they had done them and found the issue they would have at the very least put a disclaimer in the manual saying something like "Autofocus over 12 X on this camera is not advised due to the high zoom settings and large pixel count" They often put warnings in manuals ...sorta like "You may get bars on the footage under flourescent lighting in this mode...This is not a fault" During field tests this was either no noticed or they figured surely no-one will try to autofocus at full zoom. As you already know the 130 never worked for me either...it was hopeless in a solo dual camera shoot ..and for other reasons too.
I would still like to know if other manufacturers have a 22X zoom camera with the same pixel count that does autofocus correctly...my feeble brain says no cos the image has just become too magnified so there is a focus problem. Then again I can leave my HMC80's in full auto and they do not hunt on a stationary scenario (I don't of course and they are in manual at weddings) but you have a situation with a horse thundering down a narrow paddock and you want it to be in focus 100% of the time and I understand that. I have much the same situation at weddings ..the bride comes down the aisle and she has to be continually in focus from maybe 50' away right up to almost on top of me however I am not at 22X zoom!!
I would hate to have to leave Panasonic after such a long time for JVC as I love the cameras...Are you sure that the new camera (I'm assuming the new 600 series) will not do the same thing in auto???

Chris

 

[ELN614]

I can see why the camera struggles with focus based on what Jan and Barry say

It doesn't struggle 90-95% of the time. It is in focus!!! I said in this thread it works MOST of the time!
Just perhaps 2% of the time is just plain wacks out. Like a broken camera with an intermittent problem. If I send this into Panasonic.. they will do what car dealerships do. Test it once and see that it works and say "nothing wrong with it" and send it back. Intermittent problems are hard to show when they are just that-- intermittent. You notice how if car dealerships can't duplicate the problem they won't fix it. Same thing would happen with Panny. Heck, they already came out and said this is normal and it is suppose to lose focus while in auto focus. (granted they said that in a very convoluted / beat-around-the-bush kinda way.)

Just my opinion but Panasonic obviously didn't do autofocus tests at full zoom at the pre-production stage ..if they had done them and found the issue they would have at the very least put a disclaimer in the manual saying something like "Autofocus over 12 X on this camera is not advised due to the high zoom settings and large pixel count"

Look at the video clip again --- "B" goes out of focus before I even get halfway through the zoom range. I can't even get to 12x before it goes out of focus.

And the first clip of this video too.

...they would have at the very least put a disclaimer in the manual saying something like "Autofocus over 12 X on this camera is not advised due to the high zoom settings and large pixel count"

Which is one kind way of saying: "disclaimer: don't expect auto focus to work past 12x point."
An analogy: you just bought a 8 cylinder Ford Mustang -- "disclaimer in owner's manual: don't step on the gas pedal -- the car will stall."

AGAIN this does not happen every time .. very few times. But still this should not happen as consistently as it does.

 

[kwkeirstead]

Very very interesting.. I guess more testing is needed on moving objects, particularly ones going straight out where you have to adjust zoom on the fly (and iris?).

After all of the testing I did on static objects over several months, the conclusion is if you can approach a near object via a path where you slowly arrive at the target at an edge that has high contrast, you will get consistent results and fast lock in of focus. I can run a test where this is repeated ten times to hopefully demonstrate this.

Now, the big problem is, how do you identify, other than via trial and error, which edge, at what height of a target, has high contrast?

I have noticed that a no contrast scene gives a flat line on the waveform screen. Can waveform help to pre-identify how an object should be approached?

Is there such a thing as a contrast meter?

Or is framing, zoom, iris and smoothness of approach enough to worry about such that having to travel on an arc that takes you through a high contrast area close to the target is just one extra thing too many to be concerned with?

I don't see how a person can pan to an object but identify in advance how that object should be approached to achieve rapid focus lock-in.

On the horse videos, is AREA ON in these?

If AREA is ON and you manage to keep the area on the horse as it moves, does this decrease the chance that focus will start wandering.?

 

[ELN614]

I tried the area on a few times and really felt uneasy about it. I was afraid it was going to mess up and I only get one shot of getting it right.
Iris is typically auto unless I am at a specific location where I know manual is better.

Interest question: What if auto iris had the problem instead of auto focus ---- what would the stir would be? Would Panny feel more obligated to find a solution or just say it is physics?

 

[Rod Pike]

It doesn't struggle 90-95% of the time. It is in focus!!! I said in this thread it works MOST of the time!
Just perhaps 2% of the time is just plain wacks out. Like a broken camera with an intermittent problem. If I send this into Panasonic.. they will do what car dealerships do. Test it once and see that it works and say "nothing wrong with it" and send it back. Intermittent problems are hard to show when they are just that-- intermittent. You notice how if car dealerships can't duplicate the problem they won't fix it. Same thing would happen with Panny. Heck, they already came out and said this is normal and it is suppose to lose focus while in auto focus. (granted they said that in a very convoluted / beat-around-the-bush kinda way.)


Which is one kind way of saying: "disclaimer: don't expect auto focus to work past 12x point."
An analogy: you just bought a 8 cylinder Ford Mustang -- "disclaimer in owner's manual: don't step on the gas pedal -- the car will stall."

AGAIN this does not happen every time .. very few times. But still this should not happen as consistently as it does.

Exactly my fear ELN614. I have submitted this issue to Panasonic and hope they don't reply in that mode!!. In subsequent tests I have repeated this failure in the range of 10-30%. .
If this camera were used on a remote unmanned crane or tripod as a "B" roll for long zoom follow focus shot it would clearly fail some of the time. (and please don't tell me that the op would manage it, some of us work alone with two cameras).

I think the nub of the matter is that it doesn't comply with Panasonic's own product description or specification as laid out in their literature, manual and advertising material.

Page 11 of the manual: para 2:2 "In accordance with setting of items on the setup menu (AUTO SW) screen, focus, gain, iris and white balance are adjusted automatically."

There are no caveats written,assumed or implied regarding focus lock up under certain conditions. If there were, then we would have been better placed on a purchasing decision in the first instance. This is all the info we are given.

This is about this camera, not our various skills and judgements when discovering it doesn't perform as specified.

 

[Mike Warren]

Of course, once the JVC HM600/650 are out we will see if it is in fact possible to get better AF results in general. They are very close to the same specs. If that also has similar issues, I'd conclude that it's currently not possible to get good AF performance on such a large lens at this price point.

Actually, the HM 600 is another "almost right for me" camera. A requirement for me is 1080p50/60. Without that, I'm not even going to consider it.

 

[ChrisHarding]

Hi ELN614

Your interest question is the EXACT reason why I dropped the AC-130 pair I had...I seemed to be able to handle the focus side. With the HMC80 in auto iris the camera changes not only the iris and auto ND filter but also changes the shutter speed between 1/50 and 1/2000th based on the light..This way you can get a seamless exposure from heavy shade to bright sunlight. Now the AC-130 only changes iris..shutter remains fixed and the ND filters are manual...so you expose for a scene (either manually or in auto initially) and because you are a solo shooter, you leave the camera running doing a wedding ceremony in autoiris. The sun goes behind a cloud and the ND at 1/64th is now too much and the camera prompts for an ND change to 1/4th ...you have to stop your second cam and rush over to the camera and change the ND (you also know that darn sun will pop out again!!!) Every time you flick the ND switch it creates a flash in the footage and we are talking about a wedding ceremony here so it could be at a crucial point too.
That was my major reason for changing, autofocus was a close second followed by the awful ALC on the XLR's and poor default scene files. However I was warned before buying too..another forum had already come up with the autofocus issue but I "poo pooed" it and bought the cameras anyway. For those that still have the 130 or 160 the randomness of the AF focus lockup would be scary for me!! What happens if it's simply a non repeatable shot??? I would rather have it not work at all then at least I know where I stand! Plenty of us also work solo ..I'm often up to 20 metres away from my main camera at weddings so it HAS to be able to look after itself!!

It will be interesting to see what response your dealer will come back with Rod??? Were you given any time frame ?

Chris

 

[Rod Pike]

Hi ELN614

It will be interesting to see what response your dealer will come back with Rod??? Were you given any time frame ?

Chris

No I was not, but the Panasonic service centre here in NZ (Provideo) has an enviable reputation I understand, and having spoken to the MD on several occasions found them to be patient, helpful and gave unbiased advice to me on other issues, so I am confident that they will not let this slide. They are of course now governed by Panasonic Japan's response. So watch this space.

 

[Rod Pike]

Of course, once the JVC HM600/650 are out we will see if it is in fact possible to get better AF results in general. They are very close to the same specs. If that also has similar issues, I'd conclude that it's currently not possible to get good AF performance on such a large lens at this price point.

.

Early adopters please report back here immediately! (this time I won't be one of them)

 

[Mike Warren]

Now the AC-130 only changes iris..shutter remains fixed and the ND filters are manual...so you expose for a scene (either manually or in auto initially) and because you are a solo shooter, you leave the camera running doing a wedding ceremony in autoiris.

I understand completely why you need that functionality, but my guess is that Panasonic considered the feature non-professional because of the change in look as the shutter speed changes. But considering it's a pure software function and would have only cost them once to have included it. It would have been different if there was a larger usable iris range.

As chip size changes, the way various functions work and interact changes, but there's no magical formula to say exactly at what point something should be done a different way. If Panasonic had not included the manual NDs but had instead used auto variable NDs, everyone would be screaming about how unprofessional the camera is. I also suspect that by the time you get to 1/3" chips, variable ND may start to show problems with uneven graduation across the frame.

I'm working on a product design at the moment where I have a budget and a set of wanted features and performance. That just isn't going to happen so I need to firstly try and come of with some clever ways of meeting some product specs at a lower cost to allow more wiggle room in other areas. Ultimately, some wanted features may need to be left off.

And I just know people using the final product will say "I would have paid an extra few dollars for feature X. Why was it left out?". But then someone else will say the same for feature Y and another for feature Z.

Bah! Who would want to design electronics for a living? Anyone want to hire a video camera operator?

Edit:

it's a pure software function and would have only cost them once to have included it

Actually, it's possible (but unlikely) that the program memory is so full that they couldn't squeeze in anything else. I remember a design of mine about 10 years ago that even after all the assembly language optimization I could do, the memory was only 3 words short of being full. That was a close one. If I couldn't have done it, it would have meant moving to a more expensive chip and blowing the budget.