Mike Brennan

I'm shooting a "dig" of what archeologists are calling "Hobbitts" in Georgia (Europe). More fun than NAB!

Last year was worth it for me to see RED in the flesh, this year I expect some interesting 35mm format camera mockups/prototypes from the main players.

Had a look at a RED today for aerial applications it was aptly named after a well know aviator.

Mike Brennan

Now that REDs are becoming available for rent worldwide, Im wondering what we think its major plus and minus points are in its CURRENT state of development, including associated problems.

Ill start off with my impressions not gleaned from personal experience just tlak and web chat...

RED Plus

35mm DoF

High frame rates

Adaptable formats

Price

Plenty of feedback on REDuser as to bugs...

Good viewfinder

RED Minus

Apparently Flakey post performance

Some critical menue operations time consuming.

Lack of coherent opinion as to its dynamic range

Numerous minor bugs with recording camera software and post sofware.

No live 4k or 1920x1080 output.

Mike Brennan

I'm glad to see some big projects using the Origin. Ever since the ASC comparison test last year, I've thought the Origin had the best overall image of the digital cinema cameras. Being able to grade RAW is now unlocking the full potential in post, too.

Here here!

14bit A/D from the Origin CCD is superior to 12bit A/D

16bit A/D to match 16bit Origin output would be better!

Mike Brennan

walter G wrote

" a secret! Want to hurt Red? Don't post. In marketing we say any PR is good PR. "

Normally I would agree with this sentiment Walter but I am finding in Europe that producers are telling me of the current pitfalls of using RED (normally it is me telling them about the traps of new kit!)

It is interesting to compare this level of producer consciousness and technical awareness to the introduction of HD.

In 2001 the anti HD feeling, from those with a vested interest in film, was far more severe than RED is facing from pro HD and digital guys.

Although the anti film lobby has lost some of its steam in the intervening years, in its hey day Kodak and Arri and others had sizeable budgets and used the best and worst corporate tactics and behaviour to keep one and all away from shooting digital.

Also there were few HD owner operators and rental companies were not willing to spill the beans about problems with HD (namely the f900) Also in the early years no coherent balanced forum or site existed.

In comparison RED owners have it easy!. REDuser is a useful source of red hot spilled beans direct from owner operators.

That it is run by RED is neither here nor there, I'm yet to hear of a substantiated claim of pressure being born on RED owner operators not to report a fault to the group.

However I seem to recall that one early adopter decided not to report a couple of faults the group as he was using beta software or hardware.

It appears to me that Euro producers are coming to sound decisions about use of RED. On balance I'd say they are being conservative, not so much from shying away due to camera faults or performance as the workflow issues, an aspect of production which has bitten many a producer in the past 5 years.

There appears to be least 500? RED owner operators, some with limited funds and high expectations that RED will bring art/love/money/success.

As the honeymoon comes to an end and cameras hit the streets, they could easily turn into REDs worse critics if life deals them a bad hand.

We seem to be twice as vocal when things are going wrong!

Jim said it was unexpected that RED would take off in the way it has for pros, but now that it shows promise REDuser will become an even more multifaceted community.

Mike Brennan

the sooner you pay your fixers the less youhave problems yen you don't master them. simple law of production!

That simple law doesn't work with all fixers. In rare cases, the more money you pay the more they want.

Mike Brennan

I'm torn between feeling that Argentinean civil law is bent as a corkscrew, and being aware that there are several times in my career where I'd have loved to have nonpaying clients dragged back to town in shackles.

 

Needless to say this is inapplicable here as the lying toerags have been paid, so all I can do is wish your poor unfortunate PM a comfortable stay until it's all resolved - and I'm sure if there's anything anyone can do from this end you wouldn't hesitate to mention it.

 

We're left, I feel, with a severe disinclination to go within a thousand miles of Argentina.

 

P

Thanks Phil,

Our experience was in the towns mentioned in the original post, not sure about the rest of the country which we were deprived of shooting in and thus finishing the project....

Seems to be the largest number of film crew held for the longest time.

Mike

Did I miss something in your original post? Theft of what? Or is that the local term for, "Failure to bribe a local official."? Will charges be filed against the Tourist official?

The charge appears to be Theft. Technically leaving town when the money hadn't arrived.

By the time we were arrested (following day) the money had arrived by wire.

But the charge had been made and local law says if a charge is subsequently dropped the accuser is bought before the judge for wasting time or making false statements or whatever.

The real problem was the person named 12 members of crew on the charge so they had to return to the town. After 3 days the judge agreed that only a couple of people should have to answer. But as he made this judgement two more charges of theft were made against the 12, this took another few days to deal with....

The rest of the crew had to remain as witnesses.

The PM is still under detention in Arg today.

Mike B

Im safely back in Europe now and the rest have the crew have returned to France Switzerland and Australia.

It was quite odd to go across the Arg/Brazil border and then to check in for the flight with a lawyer as bodyguard.

But sadly the French production manager has to remain, supported by translator and a lawyer, until the court case is heard or unless enough pressure can be applied to the judge to drop the case.

Yesterday there were 5 lawyers and the French consul himself in the hotel.

Thanks for the on and off list support, in this instance going public via local media and web, whilst still under detention was deemed a good approach and probably did help.

Mike

Wow, Mike. This is really bad luck. I hope your trouble will be over soon. In the meantime all the best!

 

Cheers, Dave

Thanks Dave,

yes bad luck it is. Short of carrying at least US$150k in cash I don't see anything that a foreign production company could have done to prevent this farcical tale.

And if the locals realise film crews have substantial amounts cash this could create a new danger.

Mike, from the poolside prison.

12 members of a French/Australian crew I am part of have been held under house arrest (in a hotel) for the last 5 days in Northern Argentina.

Our shoot has been grounded as the rest of the team must remain as witnesses.

The charge against the 12 named was theft from the local Tourist official in the town of Iguazu, who was our fixer and provider of transport ect. Here problem was she wanted cash payment of the balance on the day of our departure. The PM had apparently agreed a wire transfer which arrived to the company account the day after we left.

In the meantime she had called in the police and did not call them off even when she had received payment.

She named as many mebers of crew as possible even though the "miscommunication" was between her and the production manager.

Locals tell us that a Belgium and Japanese crew were also victim to this behaviour by the very same person.

Under local law the accused must return to the town where the alleged offense was committed.

Unfortunatly for us we had travelled 4 hours to the south by the time the local judge issued an order for our return.

This involed us being arrested at an airport, held at a poice station then driven at 3 am under armed guard back to our unfriendly fixer.

It took a few days for the judge to deliberate.

On Friday he declared that most of the 12 had no case to answer.

But also on Friday two other locals, employed by the fixer made same charge against the same list of crew! So the judge has to sit these cases today (Monday)

Despite intervention from French Prime Minister, three lawyers and it being front page news in the capital the Argentine President (who called in the Federal police to investigate) has not been able to make it possible for us to either leave or carry on filming.

In the meantime a $US1m value in HD rental kit remains idle, including a cineflex and three HD cameras.

Latest news is that the "victim" is said to have fled the country and is now in Cambodia!

Mike

HI everybody,

someone said the image shooted by Panasonic P2 can catch up with sony F900, is ture?

if not ,then P2 can get 2K? or 4K? thanks!

The latest camera from Panasonic the 3000 uses P2 to store AVC Intra 100 codec.

In general the 3000 surpasses f900 by having

14bit AD

10bit 4:2:2 recording of all 1920x1080 pixels

However AVC Intra artefacts are yet to be fully explored but at first glance looks like a better balance of useable data vs compression than HDCAM.

Mike Brennan

Odd thing is that I would have expected that pattern in the F23, being a prism block camera -- I get that pattern all the time on the F900, so I figured it was an artifact of the prism block design (though in the case of the F900, it's worse because you get red, green, and blue patterns of dots in a grid pattern.) So now I'm going to take a guess and say it's an artifact of the low-pass filter? Because I can't explain the pattern of red circles in the grid pattern otherwise.

 

You also have to figure that the two cameras couldn't have been using the same lens, not that that explains the pattern, but it might explain other aspects of the flaring differences.

David the most obvious grid pattern internal reflection on early f900s and solved in f900s mark 3 1500s and f23 with what Sony called an improved optical assembly, is this is what you saw in Claudio's footage?

If it is you may be right in that a common factor with both cameras is the low pass filter.

If RED have solved it by moving the filter forward, Im curious if this is what Sony has done, anyone with a mark 1 and mark 3 f900 could measure distance of filter to mount? I had assumed that either the low pass was set at an angle or the prism was at a slight angle to reflect the problem out of frame.

I've had the effect on f900s with many lenses including a 101x shot of jet taking off with landing lights on

Knowing how some problems are solved is useful!

Mike Brennan

From Claudios web site,

http://www.claudiomiranda.com/redvsf23/redvsf23.html

"This has become too political for my tastes. What I wanted was both parties to benefit and grow from these tests. Again both cameras offer a tremendous value to the industry. There are many shots the Red has an advantage over the F23 and visa versa. These tests are my own and it is my process on how I judge and how I choose to implement them in future jobs. This test only needs to satisfy myself. Everyone has their own tastes and needs to judge what is important for their task at hand.

Claudio"

Mike Brennan

A new film handling wall chart for projectionists made by a team lead by Brian Guckian is dedicated to the late John Pytlak.

The chart will be launched on Dec 17th in London.

More details at BKSTS

Mike Brennan

Commercial concerns to one side I think keeping 50 to 100 cameras in the market place for much longer before nailing down the hardware and software spec would have been better than just a few months.

The latest generation RED is being beta tested now for just a few months, then +2000 cameras will hit the streets.

A longer period would allow long form projects to be completed and further mechanical or hardware changes made.

Ideally lets see a TV series and few features completed and distributed.

No matter how long the development phase, there is nothing like a completely new camera and post software being used in the field for a healthy period of time.

The concept/promise by RED of free or low cost upgrades is new in the pro electronic camera industry and so I stand willing to be to be corrected that it can be done!

And if after 18 months RED can upgrade/ software or hardware of +2000 cameras at a price that is commercially acceptable it indeed will be a revolution.

Mike Brennan

OK, if I'm understanding this correctly, cosmic rays are not the only thing that can put a tranient defect in the image, but the most energetic of them are the only thing with enough oompf to potentially do permanent damage.

 

 

 

 

-- J.S.

John and Keith,

Johns statement above may be the most precise description of the effects of cosmic rays.

But I note that gamma rays used to get the bad press from manufacturers for knocking out pixels permanently.

Now its cosmic rays that is said to be doing the damage.

Are the terms interchangeable?

Mike Brennan

Thanks Mitch. What do you think about Panasonic's CAC feature that comes with the HPX3000? According to the Panasonic brochure, using a Fujinon or Canon Lens that's CAC enabled will result in images virtually the same as using good quality primes. If you like that CAC feature, would the Canon HJ22ex7.6B work as a jack of all trades lens? Sorry to keep grilling you -- but do you have a ballpark price for that lens?

 

Finally, I'll be phoning you soon. I know your company handles rentals and sales -- do you work both departments?

 

Thanks again,

 

Tim

Tim,

There is a trick to make the CAC work after it has loaded the lens CAC file. Apparently one has to zoom and rotate focus to their end stops to activate it.

Only found this out after I handed the camera back last week.

In UK Fuji lens fles were loaded in the camera but no Canons.

Mitch have you been able to see CAC working?

Does it reduce prism fringing evident with the use of wide lenses or is it purely counteracting lens induced aberrations.

Tim, the 3000 is an excellent package and in its class the highest quality 2/3inch EFP style camcorder on the market.

With a small caveat to wait and see how AVC I 100 flows through the post production pipe.

Mike Brennan

Couldn't agree more - the only workflow to come close to what I suggested is S2, but it's frankly so heinously unreliable that I would never, ever use it for anything serious.

 

I would prefer to create a situation whereby we were able to do everything on set without having to involve the post house all the way through principal photography. Transferring stuff to two verified mag tapes and creating proxies is something that can be made pushbutton easy and is clearly very much more reliable than couriering unprocessed 35mm neg around and dunking it in caustic chemicals.

 

The real cost saving of it would come from keeping it self-contained. Of course, you'd have to have a preproduction discussion with the post people anyway, but you could avoid having to have them spend time on dealing with dailies work.

 

Phil

Thanks for the replys, both very valid.

Its interesting to predict how workflow and technology will develop.

The Panasonic P2 player with LCD screen is a neat device.

Given small footprint of CF hardware perhaps duplicate, simultaneous in-camera recording could be the most insurance friendly?

Mike Brennan

Cosmic Rays are the only type of naturally-occuring radiation where a single particle has enough energy to produce a visible effect on an imaging device.

 

They are not really "radiation" in the normal sense, they are actually single atoms travelling at enormous speeds through space, ........

I still think the biggest "radiation" danger would be from someone pointing the camera at the sun with the iris wide open. Although it's hard to "cook" the silicon sensor itself, the Bayer Mask dyes might not be so forgiving!

Thanks for a very informative post Keith!

Does make me realise how dependant we are on our relatively thin atmosphere...

Mike Brennan

What price do CF cards need to drop before production gives up on the idea of dumping data on location and instead, buys and owns cards as it does tape?

What is a data wrangler with support equipment, expenses hotel transport ect going to cost? Min US$1k a day?

For docs and low budget there will be pressure for existing crew to do the transfers.

In which case it is just cost of the cloning kit and a six pack of Red Bull.

I reckon 60 minutes for $150 would consign data transfer (as a neccessity) to the history books.

It really would be beter if we didn't go down this road at all... erasing rushes before final edit :o

Mike Brennan

Cosmic radiation can be one possible cause. But not the only cause.

 

Going deep under ground would substantially reduce the number of particles of cosmic origin you encounter. But it could introduce loads more radiation from surrounding minerals. Here on the surface, there are places where radon gas comes up from the ground and can reach significant concentrations in buildings. Red clay bricks can contain elements that decay and give off radiation. Perhaps people who live in brick houses shouldn't shoot HD? ;-)

 

And radiation isn't the only source of tranient hits or noise.

 

 

 

 

 

-- J.S.

John,

what are the other possible causes of transient hits?

Mike Brennan

A bunch of guys starring at a big monitor in the dark is not the way you measure soft errors in semiconductor devices. What you are seeing is noise. First you have monitor noise, line noise, power supply noise, cable inductive noise, thermal induced noise along various leakage mechanisms which change bias or back bias and other levels affecting data lines. And that's just the monitor and interface circuits. This doesn't even take in to account all the noise accumulated from the sensor to the input of the monitor cable. Okay, so you caught it on tape. Yeah. Do you here what your saying? A tape, a tape? Now there's a way of insuring there is no noise introduced into your video. Let's see, a magnetic induction device, mechanical friction of mylar over glass/metal sensors, and mylar has an elasticity, humm . . . coefficient of what? And how many mechanical connection circuits, and how many sources in the path are susceptible to noise? Your joking right?

 

How can you discern all the noise from all these sources, from a single bit soft hit which lasts only 1/24 or 1/30 second, or what ever the refresh rate of the acquisition was?

 

Such would be akin to putting your head to the ground in Poughkeepsie listening for the biorhythms of a California Condor on Palomar Mountain.

 

The only credible way to test is in the lab, and testing the device directly under very controlled environments. I could probably think of another dozen or two things to do, but to begin with you need to know what your testing, so you would first probe at wafer before passivation and gather the parametric of the lot. Then you'd put the glass on, scribe, break, package, etc. Then you'd run the packaged devices and test and bin-out. Send them to burn-in and at least get out the infant mortality fallout from contamination, etc. Retest and bin-out again. Then set up the test racks with enough device to give some statical relevance for your observation. I used to run several thousand devices. You would isolate by lots and bin-out parametric. Probably would eliminate the bayer mask to eliminate any anomalies induced by the mask for the test. Then you'd test using different patterns of write charge levels, read, write deplete, read, varying then number of writes at varying writes which will leave varying residual charges in the cells so that you can measure cell sensitivity/susceptibility to particle impingement as well as the soft error. All this has to be done in a very controlled environment for any kind of quantifiable and qualified measurement.

 

My guess is that, if you put a bucket under your camera to catch all those cosmic particles and come back in an hour or two, you'll find it full of red herrings.

 

As far as the Red's debayering algorithm making an error "twice as big" . . . huh? Do the math.

Well at the time we weren't trying to measure soft errors we were looking out to initially identify that white specs were in the camera.

Yes by catching it on tape it proved that it was something originating in the camera and not a monitor issue.

How can we see them amid the noise? well the hits are much brighter than the camera noise and are clean, they are quite easy to see!

I don't understand all of the detail of what you wrote, but have you seen the effect as presented on a HD camera output?

In respect to buckets of herrings you could be right, but nothing less than an examination of a HD camcorder using your test method is likely to replace the cosmic/gamma radiation story we have been fed from NHK, Sony JVC and NASA.

Here is a link to a report on the NHK HD camera in the International Space station that NHK say suffered 150 cosmic hits that were strong enough to keep the pixel hot. http://idb.exst.jaxa.jp/edata/02497/pdf/STJ85-14.pdf

Also here is an freely available extract, available from the technical reports server of a NASA study saying HD cameras seem to be more prone to cosmic hits than SD cameras. Full report here;

http://ntrs.nasa.gov/search.jsp?R=672212&a...294967190%2B269

"Recent experience using high definition video on the International Space Station reveals camera pixel degradation due to particle radiation to be a much more significant problem with high definition cameras than with standard definition video. Although it may at first appear that increased pixel density on the imager is the logical explanation for this, the ISS implementations of high definition suggest a more complex causal and mediating factor mix. The degree of damage seems to vary from one type of camera to another, and this variation prompts a reconsideration of the possible factors in pixel loss, such as imager size, number of pixels, pixel aperture ratio, imager type (CCD or CMOS), method of error correction concealment, and the method of compression used for recording or transmission. The problem of imager pixel loss due to particle radiation is not limited to out-of-atmosphere applications. Since particle radiation increases with altitude, it is not surprising to find anecdotal evidence that video cameras subject to many hours of airline travel show an increased incidence of pixel loss. This is even evident in some standard definition video applications, and pixel loss due to particle radiation only stands to become a more salient issue considering the continued diffusion of high definition video cameras in the marketplace."

So most techies agree that cosmic radiation causes stuck or dead pixels. Do you agree?

To be clear I am referring to white specs where the level of energy isn't enough to make them "dead" and stay white.

Rather it is a transitory hit and discharged by the next frame.

At the time I was lead to believe that Sony link cosmic radiation to both dead pixels and flashing pixels.

A simple way of establishing if the white transient hits are not cosmic rays would be to use the camera deep underground. Would you concur?

Anyone on the list shooting in a salt/coal/gold mine?

In respect of asking the question about a pixel going hot in a bayer pattern sensor I figured that if one green pixel took a hit it would then be used as incorrect scene information to build colour for surrounding pixels, thus making the blemish larger.

Perhaps a Bayer could do the opposite and disguise it if it was a red or blue pixel that was hit, but the NASA report suggests that in-camera processing could also amplify it and they say varies with different sensors and between CMOS and CCDs.

Mike Brennan

Do a complete test shoot at a local night club.

Take the pictures through an offline and online edit.

Try and make the test as realistic as possible.

It sounds like you may work to a tight deadline so make sure time pressure is included in the test, a very difficult thing for the crew to take seriously unless they are being paid:)

Data wrangling under pressure on location will be a challenge if logistics are not fully worked out so include data wrangling in the test.

Try and include some kind of meaningful picture quality spot checks whilst you are shooting, including audio if you plan to record in camera.

Get to know local RED owners (there is a list on the RED forum) in case you need their brains/experience or to supply additional REDs.

Budget to customise the computer kit including practical flight cases and battery power for laptops ect.

Factor in schedule of the band which may not leave much time to charge batteries or do running repairs.

Look forward to hearing how you do it!

Mike Brennan

It would indeed be quite a phenomenon to see that rate in that part of the universe. In all the time we studied the alpha particle problem in my lab at Mostek I never observed such a rate. We began observing alpha-particle induced soft error mechanism in NMos devices as we were scaling dynamic memory using 5 microns processes and below. The alpha particle would create electron hole pairs as it passed through the memory cell thus altering the charge sufficiently enough that the sense amp would inaccurately determine the state. OFF being the state susceptible. I designed special test devices (both NMOS and CMOS) to evaluate the ceramic materials we used for industrial and mil spec devices and never saw rates like 4 hits per minute. We had customers like Borroughs, CDC, DEC, LTV Vought and IBM that occasionally made such reports, be each time we evaluated such claims, the data errors were caused by cross coupling of data to data or data to control signals though reactive coupling created by signal patterns. On occasions some solders used in manufacturing had metal whose decay had elevated alpha particle emission.

 

The rate you are reporting falls well into that of noise not particle induced soft errors in the storage cell. After the read and equilibration the next write (photo detection) cycle the cell would be recovered, after all it's a soft error. The probability of that same cell being hit again, even if the radiation was being emitted from the decaying materials of ceramic packaging is quite small.

 

I suspect between the deBayering, codec manipulation of the data and frame rate, such soft errors are imperceptible. It is much like what happens in film which integrates defects or rather variances in the oxide crystals, from frame to frame.

 

There much more important issues to be concerned about than single bit soft errors in the digital acquisition realm of cinematography.

Very interesting but your explanation contradicts what I believe to be Sonys explanation.

Im not taking sides:) here is a little more detail on my experience just to be clear.

Random hits.

There are three ccds of high density so three times the likely hood of cosmic hits?

It happens on cameras from different manufacturers.

The effect was obvious enough (when you really look for it!) of an issue for Sony UK to change the CCD block in my f900 a few years back.

This was a big expensive deal. But then, the same day with new CCD block the hits were still there! Im told Sony Japan then said that it was cosmic radiation, something new to me and the UK engineers! We knew about gamma but cosmic was out of this world.

Then we checked a SD camera under same viewing conditions and bingo we could see it there too. We recorded and played back to identify, so it isn't a monitor issue it goes to tape.

Needs a few dbs of gain a big monitor, dark room black cloth and don't blink and you can see "them".

Just checked the Panasonic 3000 HD camera that I have on test and the rate is two a minute at 1730GMT.

Will be very interested to know of manufacturing or inherent design issues that could cause these sparkles if the cosmic theory is in doubt.

I think it is important to learn about random white hits, in two cases that I've heard about in the UK the lab was blamed for them! When I first saw them during a shoot I considered informing the producer that the camera had developed a minor fault!

More reading here.

http://www.jai.com/SiteCollectionDocuments...-CosmicRays.pdf

In REDs case Bayer may well disguise the hit or could make it twice as big?

Mike Brennan

I don't understand your question.

Will you do a comparative test between Dalsa and RED side by side in the same place at the same time, ie a shoot out or just test one camera?

Mike Brennan