And presumably they're not inundated with viewers writing in to complain about their picture quality!
I think that may be the key, at a certain point something is just "good enough" meaning that if you increase the quality you can measure the difference and post specs and side by side comparisons but in it's final form there is no difference. It's very much the same with stills, the difference in image quality from a £300 consumer DSLR and a £5000 pro one is not really significant on the web, in magazines, in 10x8 prints. It's only when you do massive enlargements that it shows. Ditto with moving picture I guess, you'll notice some difference on the big screen - they did when turning Planet Earth into Earth for cinema release, the producers' first criticism of Varicam, that it didn't hold up so well as the 1080 cams. The biggest problem though AFAIK when making Earth was stabilising shots as the long lens work showed lots of wobble when maginified - hence from then on we've all been urged to get bigger tripods (Ronford Atlas, O'Connor 2060 etc.)!
Steve

Steve, the push is not just for 1080, but for 1080p at 50 or 60fps. Such a standard would make things easy because as David pointed out it can be downconverted very well to 720p/50/60 as well as 1080p/25/30 very easily.

Sony have already gone some way towards this with chips that support this. I believe SR can handle these framerates already, and most new TV's with HDMI can also display it already. Sanyo have already kickstarted it at a consumer level with their new Xacti cameras. All that is needed is for more professional cameras to start supporting it. Using LongGOP compression as per digital broadcasting the datarate does not need to be hugely higher than it is already either.

But also the other thing you missed is that although broadcasters might be showing 720p now, many of them want to be secure for the future. 1080p at higher framerates would mean that there wouldn't be the current spaghetti mess of which HD format to shoot in any more. Regarding the wildlife, with the rarity of many of the things being filmed it would certainly make sense to use a camera that would be capable of 1080p/50/60 as soon as possible.

I'm sure you're right Simon that 1080 is the goal, but I think you missed my point re image quality - at some point good enough is good enough. I don't know anyone who's complained about the quality of Planet Earth for instance either on HD broadcast or BluRay. For TVs even upto 50" while there's no harm in having it 1080 there's probably no desperate need from a viewer perspective.
You raise an interesting point re 1080/50 or 60P. If a programme is transmitted at 50P that'll mean that for even half speed slomo in camera of course we'll need a 1080/120P camera. We'll all have to get Phantoms!
Steve

I think you missed my point re image quality - at some point good enough is good enough
In principle you're absolutely right. But what's good enough on a 19" screen is not good enough on a 100" screen, given the same viewing distances!

OK, that's a bit flippant, but a lot of work was done a few years ago to try to discover exactly what "good enough" was. The conclusion was in favour of 720p - but unfortunately the research became out of date almost before it was published. Not through any fault of the researchers, but because screen technology and sizes improved way faster than anybody thought.

For UK homes, the average viewing distance is about 2.7m, and that's remained pretty constant for a long time. The research was assuming about 37" as the "norm" screen size for the foreseeable future, and at the time the best resolution on domestic screens was typically about 1350x768. That's nowhere near true anymore, as a trip to any High Street proves. I've got a 1920x1080 42" Panasonic, but 46" and even 50" are quite common now, and it's hard not to get 1920x1080. The same research indicated that 720p wasn't "good enough" for a very significant percentage of viewers for those screen sizes.

Hence the drive to 1920x1080. True, the full advantage may not be seen on all screens, but that's no argument for not trying as hard as is practical, certainly for top end production. Neither 720p nor 1080i are seen as totally satisfactory for transmission, but they are currently the best that's realistically practical. Transmission to home of 1080p may still be a way off, but that's no excuse for not producing as such unless there are overriding factors - and the importance of varispeed for wildlife filming is a good example of that at the moment.

Regarding the wildlife, with the rarity of many of the things being filmed it would certainly make sense to use a camera that would be capable of 1080p/50/60 as soon as possible.
Only for slow motion - the whole point of 720p in this context is it assumes 720p/25 as the viewing format, for overcranking to give slow motion. If you're transmitting 720p/50, such as the 2700 won't give slow motion.

Which is why what's really needed is a camera with 1920x1080 chips, able to give 1080p/25 recording for normal speed shooting, but be switchable to 720p/25 when overcranking is needed. Accept the resolution drop just for off-speed work. It's really not surprising that the people Jeff earlier referred to got caught out by assuming the 3700 did just that. Why the 720 mode was missed off it is totally beyond me.

If you're transmitting 720p/50, such as the 2700 won't give slow motion.

That's true. Which is why we need 1080@120 :-)

Which is why what's really needed is a camera with 1920x1080 chips, able to give 1080p/25 recording for normal speed shooting, but be switchable to 720p/25 when overcranking is needed. Accept the resolution drop just for off-speed work.

The trouble with that is that it is a pain to do. The EX can do this, but I hate having to switch resolution modes. It means going into a menu, which takes time, and sometimes the event I want to capture goes by. So if I know I might need to do off speed shooting a fair bit I will just shoot 720p. That way I can just press a button on the side, dial in my framerate and press record. Or if my framerate is set I just press the button followed by record.

So it would be better to have a 1080/60 capable camera. There's nothing stopping the main recording being 25p while the offspeed is up to 60. Just as now you shoot 720/25 with 60fps overcrank instead of just using 720/60.

Though really I would argue that shooting at 25p so that slow motion can be seen is sort of a fake effect. 25p is such a slow framerate that surely a camera that shoots 120fps for slow mo and 50 or 60p as normal is far more desirable? Decent slow motion in 50 or 60p playback. Ultra slow motion for 25p shoots.

Well, we know there are large sensor RAW cameras that can shoot higher frame rates, although many producers don't know that the RED One shrinks the sensor size the higher the frame rate selected, losing resolution.

The HPX3700 consumes 38 watts, 10 watts more than the 2700 due to the full raster CCD's. The Sony SRW9000 consumes 57 watts in power save mode without any option cards. Hopefully this will go down once the large sensor CMOS option and SSD recording options appear in 2011 or so.

1080/60P with CCD's is expensive and power hungry, wasn't practical for Panasonic to do with the 3700. Certainly the P2 cards have the throughput to deal with bit rates in the 600Mbps range. Why they couldn't provide 720P at any frame rate for the 3700 or 3000, I don't understand.

I love the convenience and frame rate flexibility that shooting in 720P provides. With my 2700, I can choose two different frame rates on user buttons. I can ramp from any frame rate to any other within 1-60 fps while recording and back, if I want. Not to mention being able to record 160 minutes on a single 64Gb P2 card in 720/24PN vs. 80 minutes in 1080/24PN.

My screening room has me seated 10' from a 100" screen. I can tell the difference between 480P and 576P compared to 720P, but I cannot see the difference between 720P and 1080/60i. Indeed, on the high end projector forums, the projector enthusiasts, who are actually pretty critical in their viewing, can't tell the difference between 720P and 1080P projectors with 120" screens or smaller. I can tell looking at pixel structure in white titles, but not on normal live action at normal viewing distances.

Jeff Regan
Shooting Star Video
Shooting Star Video (http://www.ssv.com)

Jeff, I know that BBC did viewings when Varicam first came out, on projectors and big monitors and looked at pictures from both Sony F900/750 and Varicam and many decided subjectively that the Varicam just looked better. Perhaps not due to sharpness alone, as has been said it has a very nice quality to the image apart from sharpness.
Simon, I actually don't find it too bad changing mode on the EX3, it was the PDW700 where it was a nightmare - 2 problems over the way it's done on the EX3, first you have to power down for the change to take effect, second you have to change the bloody dsc as they won't do 1080 and 720 on the same one!
1080/60 (let alone 1080/120) does seem to be a struggle for CCD cameras, and with CMOS you've always got skew issues, and if you're after slomo the chances are your subject is moving fast and hence likely to induce skew. It really does seem like a good solution is a pretty long way off - but you never know with the way technology advances. I've never seen skew when shooting with the Phantom HD, but perhaps it becomes less and less evident when shooting really high frame rates?
Steve

I've never seen skew when shooting with the Phantom HD, but perhaps it becomes less and less evident when shooting really high frame rates?

Depends what kind of CMOS sensor they use. Skew isn't necessarily something that happens on CMOS regardless. CMOS chips can be made with zero skew, but is expensive and difficult, and I believe also reduces the effective light gathering area due to more jiggery pokery having to be crammed onto it. Someone else other than I will know the specific reasons.

The Phantom may not be showing skew due to the sheer speed of it. At 500-140,000fps that sensor is going to have to be clearing at a phenomenal rate.

I think it's not actually CMOS as such it's whether it has a global or rolling shutter. Global shutters of course have no skew, but that's the bit that's expensive I believe.
Pretty sure the Phantom has a rolling shutter actually, as does the RED of course. There is another highspeed camera called the Weisscam which I believe has a global shutter.
Steve

All XDCAM HD cameras can now record multiple formats and frame rates on the same disc, so no need to change disc anymore.

Nat Geo specify delivery on HDCAM SR at 1920x1080.

I can clearly see the drop in resolution when channels switch from 1920x1080 to 720P on my 42" TV.

Alister,

That's odd because Nat Geo's commercial submission specs call for 720/59.94P on D-5 or SR, according to a document I have dated 9-1-08. I wonder why they would have two sets of HD standards for submissions?

At normal viewing distances I cannot see a difference between a well upconverted 720P image and a 1080/60i to 60P converted signal on a 42" 1080P LCD display.

Jeff Regan
Shooting Star Video
Shooting Star Video (http://www.ssv.com)

A couple of notes. I spent a fair amount of time at Abel Cine in NY (Thanks Charlie and Andy) with the 350 they now have on the showroom floor. The packaged Fuji is a useable lens not spectacular. All my comments are based on this lens not any other. Also the camera settings have not been optimized so the colors were nice but not seductive. typical Sony out of the box look. On the showroom floor was a Panasonic 300 next to it and a 700 or 800 (can't remember which). In truth the 700 looked a little sweeter than the 350 but it did have a good piece of glass on it and had been set up where the 350 was just out of the box.. The differences between the 300 and 350 on the monitors was not night and day. Depth of field and sensitivity a nod to 350. Color and size went to the 300. The 350 body is lighter than the 700 but the same height and length. It is thinner than the 700.
Sony 350 was better on motion blur and roll from Cmos than 300. Not sure you could shoot any differently with it but Sony does have some advantage in that respect. Camera was very clean so you could see very little extra noise in the first few gain positions. This may turn out to be a godsend in the field but at the showroom it was not as important.
Viewfinder is pretty nice and seems sharp enough to tell focus while showing Color. Of course everything looks rough edgy with the peaking in the viewfinder compared to the actually pictures coming out of the monitor.
Playback controls were a little odd as you can get clip view and or use the play rewind controls like the disc and tape cameras on top. once you selected a clip you still hit play on the top controls. Also you had to hit stop to get back to camera it didn't go back to clip view at the end but stayed on the last frame still. May be software settable.
The lens is similar to the Fuji that comes with the 300. What I really noticed was the 16X ends up being only slightly wider than the one on the 300 17X but is definitely not as long as the 300 lens. Personally I don't like the Auto Manual switch on either lens as it goes forward and back instead of side to side but it is a minor complaint. Andy Shipsides was noticing the zoom was a little sticky at slow speeds and seemed to be difficult to feather starting and stopping. Not sure if this is intrinsic or not. The focus had no breathing which was nice if you want to do a rack.
Overall it is a much better package than the EX-3 if you have the dough and don't mind the bigger size (which is easier to use). Not sure how much of a premium people will pay for it in rental but it certainly would be a good choice for the run and gun doc news crowd as well as medium to high end industrial work.

All XDCAM HD cameras can now record multiple formats and frame rates on the same disc, so no need to change disc anymore.

Nat Geo specify delivery on HDCAM SR at 1920x1080.

I can clearly see the drop in resolution when channels switch from 1920x1080 to 720P on my 42" TV.

Hey Alistair,
Good point on the disc info.

As far as delivery of a Master to Nat Geo that doesn't mean the spec for production precludes using something else as most are not shooting on HDCAM SR in the field. Of course over time I think production people will prefer the higher pixel number formats but the end viewer will not necessarily care.

As far as seeing a resolution drop between 720 and 1080 on your TV I think there are many possible explanations.
In my experience monitors tend to look better when receiving the signal they are meant to display compared to the transcoding done to display the other formats they might be receiving so a 1080 monitor has to scale a 720 signal so you may be seeing the difference caused by the monitor electronics or the cable box not necessarily what the signal would look like on a 42 inch 720 display. Fair side by side tests are hard to come by

I agree with Alister, the difference between 720 and 1080 is noticeable, I'm viewing on a 37" Plasma and I can see the resolution difference between many of the aerial shots on "Planet Earth" shot at 1080 and the 720 Varicam shots, I'm also not at all impressed with most of Nat Geo HD's output, If I switch to Rush HD which has much more 1080 material the difference is obvious, even my wife can spot it.

Steve,
While I agree with your observations about HD outputs and what looks better on your monitor I still think the argument as to why you are seeing what you are seeing is not as easy an answer as you think. Everything is being transcoded and recoded depending on the edit as well as transmission so what we end up seeing at home can be handicapped towards one side or the other. My set up at home evolved over a few days while I got the correct cables and I can tell you there was more of a difference in what the cable box was doing to the signal than almost any production issue. First off try looking at an SD signal box into an HD set. My old Sony Trinitron CRT looked better using that signal than the new HD sets. When I got the HD set top box the HD set blew away the old CRT I had. Still had some issues until I had the HDMI cable compared to the analog HD component cable. I also noticed that SD signals had improved once I was in HDMI Land on the same monitor. I did notice however that color saturation, brightness and contrast changed from looking at the SD signal to the HD signal of the same program leading me to the conclusion that there are differences in the signal path which I can only set up to completely correctly if I don't change the channel. Also watching different channels I see tremendous differences in how they approach detail. Again all of this doesn't preclude what you are saying but I attribute a bit more variance to the distribution end making absolute judgments tenuous at best. I do know that all this technical stuff has very little to do with actual content of the programs so my feeling is if you are noticing differences in a program from one shot to the next then there is a production consistency issue not only a 1080 vs 720 issue.

Daniel I've been an online editor for about 15 years and I've been on-lining HD for the last 5 years so I'm up to speed on how monitors and television vary by setup.

What I'm saying is it is possible to spot the difference during a single programme between 1080 and 720 material where there are no differences in setup

Whether the paying public can tell the difference most of the time that's another question.

Steve,
Merry Xmas. I don't disagree that it is possible to spot the difference between 1080 and 720 material in the same program at home but the certainty as to why and what you are seeing is much harder to determine as it is not like you actually know the process that was used to put the image on your home set unlike say looking at it when you are doing the online.
Again since the specs are different it is almost impossible to use side by side comparisons of the same material in both 1080 and 720 handled optimally in both and then displayed optimally in both. As you know if you online material in 1080 then any 720 material has to be converted to 1080 at some point. Same goes for 1080 converted to 720. It is also almost impossible to match the set up of cameras so they have the same visual characteristics unless they have been matched side by side which would seem unlikely if setting them up to record 720 and 1080.
What I will say is if you see differences when you do the online they may still be visible by the time it reaches somebodies home TV.
Have a nice Holiday

Spatially, there are twice as many pixels per frame in 1080 as there are in 720. Theoretically that should be visible and apparent. If that advantage cannot be observed, it owes to degradation from the down processing of the 1080 rather than some inate ability of 720 to upscale into something that looks better than it is. So the choice given, is whether to target to the lowest common denominator with 720, or to the peak potential of the HD formats with 1080. I choose the latter.

As to whether in practice it's possible to recognize all of the transmission/conversion factors at play when a signal is broadcast, it appears to me that 720p football broadcasts on ABC/ESPN always appear inferior detail-wise to 1080i images broadcast by CBS or NFL network. There must be something to this, because I have never been able to make the judgement that 720p looked better. That includes viewing on 50 inch plasma monitors both 720p native and 1080p native panels. I have both. I believe 1080i downscales to 720p at least as well as 720p displays natively on my 720p panel. On the 1080p panel, 720p loses decisively.

All XDCAM HD cameras can now record multiple formats and frame rates on the same disc, so no need to change disc anymore.

Nat Geo specify delivery on HDCAM SR at 1920x1080.

I can clearly see the drop in resolution when channels switch from 1920x1080 to 720P on my 42" TV.

Wish this had been the case before I sold my PDW700! Why can't they just make the products with all these features before they release them?
I must admit that I can see a "resolution drop" in Planet Earth Bluray going from aerials to Varicam stuff, but saying that you can see it when going from the aerials to the Emperor penguins - and that was shot on 35mm, so what does that say? The other difference I see between the 1080 and 720 is that the 720 looks smoother, more gentle and more subtle - maybe even less "video-like" than the biting sharp aerials, got to be a good thing. Is there such a thing as "too sharp" - perhaps.
Steve

Why can't they just make the products with all these features before they release them?

The team developing these firmware upgrades are pretty small and they are under a lot of pressure so they don't often put everything in from the start so they can make release deadlines while keeping it reliable. What Sony usually do is design the cameras with a lot of features in mind and built in with the capability for later when the firmwares have been tested. Sometimes though I think that they discover that they can do something that they didn't think they previously could.

When XDCAM was first released it didn't have half the features we now take for granted. It took over a year before the Standard def cameras were given the ability to delete clips in the middle of a disc for example. The cache record wasn't present until a couple of firmware revisions after release. Although I have to admit that many of these features were listed as coming along so I was aware that my camera would get them eventually.

The ability to mix formats has been long overdue. I've never been sure why it hasn't been able to do it from the start given that were are talking about a file based recording system.

Fair enough I suppose.
I remember there was a firmware update when I still had the 700 that allowed "mixed formats" all on one disc and I thought that was problem solved, but it turned out to be just the ability to mix xdcam 422 and 420 etc. rather than allowing both 1080 and 720 on the same disc. Is it the case then that now you can mix 1080 and 720?
Steve

I continue to maintain that at normal viewing distances, the resolution difference is not easily discernible on live action content.

A couple of takes on 720/60P vs. 1080/60i, first from ZD Net:

"We still believe that when you’re dealing with TVs 50 inches and smaller, the added resolution has only a very minor impact on picture quality. On a regular basis in our HDTV reviews, we put 720p (or 768p) sets next to 1080p sets, then feed them both the same source material, whether it’s 1080i or 1080p, from the highest-quality Blu-ray and HD DVD players. We typically watch both sets for a while, with eyes darting back and forth between the two, looking for differences in the most-detailed sections, such as hair, textures of fabric, and grassy plains. Bottom line: It’s almost always very difficult to see any difference–especially from farther than 8 feet away on a 50-inch TV…."

This from Wikipedia:

"When broadcast at 60[1] frames per second, 720p features the highest temporal (motion) resolution possible under the ATSC standard. Progressive scanning reduces the need to prevent flicker by filtering out fine details, so sharpness is much closer to 1080i than the number of scan lines would suggest.[2][3]"

Jeff Regan
Shooting Star Video
Shooting Star Video (http://www.ssv.com)

A couple of takes on 720/60P vs. 1080/60i,....
But it's not just 720p/60 v 1080i/30. What about 720p/24 v 1080psf/24? (Both of which are how "film-look" material is carried over 720p/60 or 1080i/30 networks, respectively.) Then it becomes a very different argument, with 1080psf/24 most definately superior to 720p/24. And since the Varicam really needs a final 24/25 fps output to have any slow motion capability, I'd argue this is where we should be looking at for the comparison.

Equally, 1080p/60 is being looked forward to as the next big step, apart from 3D TV, and that's unarguably better than 720p/60. There's a place for 720 at the moment, due to technical restrictions, but it gives little scope to take advantage of future technology.

I can't claim to have been around at the time, but in the late 50s, early 60s there was debate about what should happen with colour in the UK. There was a quite a lobby in favour of NTSC on 405 lines, the arguments being that it wouldn't mean as big an upheaval in legacy equipment and could be implemented sooner. The proponents also crucially (and reasonably accurately) argued that the technology of the day meant the viewer wouldn't see such a vast difference at home - that 405 NTSC was "good enough". The counter argument was that technology was likely to improve, and it was necessary for the underlying system to be good enough to be able to take advantage of new technology. Fortunately (IMO) that won the day, and 625 line TV was established as standard.

It's a similar argument now. If you're going to go through all the upheaval of a format and system change, it makes sense for the improvement not to be incremental, but allow scope for future improvement. 1080 offers that far more than 720.

David,

Agreed, 1080P should be the goal, if not 2K or 4K resolution. Having said that, too many of us get caught up in pixel counting, aka RED One @4K. I'd rather shoot with an F35 at 1080P for television due to the better latitude, sensitivity, lower noise, in-camera painting.

I had a line doubled NTSC image in the early '90's on an industrial Sony CRT projector, shot test footage for Faroudja Labs in 35mm and video when line doublers were being developed.

I had off-air HD in '99 via a $2000 ATSC digital receiver and a 1080P DLP projector in 2006, when they became available. I expected to see a big difference between 720P and 1080i. Never happened. I often shoot resolution charts and the same live action scenes in 1080/24P and 720/24P with my EX1, playback via HD SDI and can't see a big difference in the live action on the projector. Yes, resolution charts can show some detail difference.

Since owning an HDX900 and HPX2700, I believe 720P is a great format for most of my client's needs. Nobody has ever complained about the resolution of these cameras when shooting in 720P or 1080P. AVC-Intra 100 is full sample in 720P, so that's a step up over DVCPRO HD. I love being able to push a button for overcranking on one of the 2700's user buttons anytime. In fact, I have the VFR button set to 48 fps, and a second user button on 60 fps.

Jeff Regan
Shooting Star Video
Shooting Star Video (http://www.ssv.com)

I must admit that I can see a "resolution drop" in Planet Earth Bluray going from aerials to Varicam stuff, but saying that you can see it when going from the aerials to the Emperor penguins - and that was shot on 35mm, so what does that say?
Steve

Not sure that's what I said Steve, I actually said I could spot the difference between the 720 and 1080 material. I'd be surprised if the 35mm material was telecined to 720, however I do stand to be corrected on that.

You're right in one respect though, badly set up cameras with too much detail in 1080 don't make great pictures once they get down the transmission path.

We actually shoot a lot of 720 50 material for online projects and it can look very good. However when correctly set up, a good 1080 picture does have the edge over 720 for me.

Agreed, 1080P should be the goal.........
But it's a goal that's available now! At least if we are talking about 24/25 fps frame rates. That was the point of the previous post. Arguing the merits of 720p v 1080i is one matter, 720p v 1080p quite another. And both 720p/25 and 1080p/25 are easily achievable now, not as a future goal, and with 1080p/25 easily superior.

The question originally asked was HPX2700 or PMW350, and it may be time to address that directly once again. Technically, the main differences between the two are that the 2700 can record with a "fully approved" codec (AVC-Intra 100), whilst the 350 has "fully approved" chips (1920x1080). The obvious question to ask then is which makes the most difference?

I don't think there's any simple answer to that - any more than being asked "which would you least mind losing, an arm or a leg?" But the possibility of being able to add an external recorder to the PMW350 changes everything.

Operationally, the main feature the 2700 has in it's favour is varispeed ramping, which may be of prime importance for such as wildlife filming. In pretty well every other way, it's outdone by the PMW350 - far lower power consumption, lighter, compatible with next gen radio mics, cheaper media, ability to dub in camera (including to consumer media) etc etc.

I can only conclude that for most users that means the PMW350 is the better camera. That becomes especially true if the Panasonic offer cutting the price of the 2700 does indeed end in a few days time, and it becomes vastly dearer than the 350. If that does happen, I can't see many more 2700s getting sold.....

Quote "Operationally, the main feature the 2700 has in it's favour is varispeed ramping, which may be of prime importance for such as wildlife filming."

You're pretty deluded if you believe that David. And to answer your question, no ramping is not very important in wildlife filming. I, and I'm sure others, toyed with it briefly when we first got it but it's not a great use or interest. So with that in mind then obviously all these hugely experienced film-makers, cameramen and prodcuers have made the wrong choice then - they should have checked DvInfo first I guess.
Steve

All right - maybe that should have been better worded ""Operationally, the main feature the 2700 has in it's favour compared to the PMW350 is varispeed ramping..........", but since the whole thread is about comparing those two cameras, I thought that was taken as read. Yes, of course other operational features are obviously far more important than varispeed abilities (such as taking 2/3" lenses) - but what others can you think of that the PMW350 DOESN'T offer?

I've listed quite a few above in the PMW350s favour, and might also add better sensitivity, but I'm curious to know what makes you think the 2700 would be the better choice NOW? What you think it offers that the 350 doesn't? (Bearing in mind the 350 was only announced a month or so back, so peoples past choices couldn't have taken it into account.)

David,

In the U.S. 1080/60i is what is available for broadcast reception. The goal should be 1080/60P, but of course it requires a whole lot more bandwidth.

Regarding media for a 350 vs. 2700, 32Gb SxS cards cost between $750 and $845, 32Gb P2 cards between $595 and $650. The 32Gb E series card has a throughput of 1.2Gbps vs. 800Mbps for SxS. While XDCAM EX will allow for longer record times than DVCPRO HD or AVC-Intra 100, the 2700 also has AVC-Intra 50 as a longer record time option--160 minutes in 720/24PN, still at 10-bit, albeit 4:2:0. A 64Gb P2 card would have double that record capacity.

The XDCAM EX codec isn't on the same level as AVC-Intra 100, so an external recorder is necessary, with the smallest on-board options limited to 8-bit, the other options are too large for hand held operation. Intra allows for in-board 10-bit I-Frame recording--there is no better memory card internal recorder available in a one-piece 2/3" camera.

If native full raster 1080 is required, I would rather buy a used HPX3000 for $20K with the advantage of no CMOS artifacts, Panasonic colorimetry, Film-Rec gammas, DRS, five card record capacity, three HD-SDI outputs, and of course, a much better codec.

I don't think the 350 is going to sell well because an EX3 does much the same for less than half the money. The 2X + cost of a 350 won't translate into twice the day rate of an EX3, I'm guessing. I can rent my 2700 for twice the rate of a 350, most likely.

Again, the XDCAM EX codec is a crippling factor for those likely to be interested in a 2/3" camera, ditto the low end glass for many(although that's easy to change for more $$). The lack of HD SDI outputs would be bothersome, the ergonomic problem of not being able to put your left hand over the zoom servo due to the large viewfinder being in the way is a problem if needing servo zooms and not using a remote zoom control.

The Varicam frame rate ramping is useful for commercials, music videos and other specialized applications. The Varicam name is known for high end production around the world, the 350 is Sony's lowest cost 2/3" camera for a reason--codec and CMOS limitations.

I think the 59db SN of the 350 is fantastic, ditto the sensitivity, frame rate and gamma options, low power consumption, weight, but I don't like the codec or external recorder concept(did that with HDX900/FireStore), don't want CMOS artifact potential. I agree that a 2700 at $40K again is a non-starter, but at $20K for a few more days, I'd make the same choice again.

Jeff Regan
Shooting Star Video (http://www.ssv.com)

I've listed quite a few above in the PMW350s favour, and might also add better sensitivity, but I'm curious to know what makes you think the 2700 would be the better choice NOW? .)

I think you'll have to ask the producers rather than me. As Jeff says, the EX3 has been out ages and has more or less what the 350 has to offer (maybe a tiny bit less image quality, but on the other hand lighter, less battery hungry, more lens power, all good things for wildlife). Suggest to the guys at the NHU or Wild Horizons who are making the new Discovery mega series that they should have gone with EX3s (or 350s) and see what they say. I haven't asked them specifically but I can pretty much guarantee that it wasn't even considered, let alone a close call and they went for 2700s - they are not even discussed in the same sentences for blue chip natural history.
Steve

In the U.S. 1080/60i is what is available for broadcast reception. The goal should be 1080/60P, ...........
Jeff, are you aware of what is meant by "psf" - "progressive, segmented frame"? If not, look at Progressive segmented frame - Wikipedia, the free encyclopedia (http://en.wikipedia.org/wiki/Progressive_segmented_Frame) . In essence, "This technique allows for a progressive picture to be processed through the same electronic circuitry that is used to store, process and route interlaced video."

So although the fundamental system used in the US for 1080 work is interlaced, it can (via psf) equally well carry progressive images, though not at any frame rates higher than the fundamental system rate. So 30p is OK, so is 24p with pull down, but not 60p.

It's exactly because of this that many broadcasters have gone with 1080 transmissions rather than 720. For subjects which need good motion rendition, there may not much to choose between 1080i and 720p, but when you think of drama, film etc - where 24/25fps motion is desired - 1080psf makes a lot more sense than 720p.
If native full raster 1080 is required, I would rather buy a used HPX3000 for $20K with the advantage of no CMOS artifacts, Panasonic colorimetry, Film-Rec gammas, DRS, five card record capacity, three HD-SDI outputs, and of course, a much better codec.
But then you lose any 720p ability, and the ability to do much in the way of slow motion at all. Full-raster chips are a good thing, but the ability to give a 720p downconvert should go along with them.
I don't think the 350 is going to sell well because an EX3 does much the same for less than half the money.
I see the 350 and EX3 as very different cameras - mainly down to form factor, chip size and the ability to take 2/3" lenses. The 350 is also far more suitable for use with professional accessories such as radiomics, and the stories I'm starting to hear are that it's likely to sell very well indeed. We'll see.

I think you'll have to ask the producers rather than me. .........Suggest to the guys at the NHU or Wild Horizons who are making the new Discovery mega series that they should have gone with EX3s (or 350s) and see what they say.
When those decisions were being made, the PMW350 wasn't even on the horizon, and I'd be the first to say that the 2700 was a more appropriate choice for the programmes you mention than EX3s (largely down to 2/3" lens versatility).

So leaving all that aside, then I'm interested to hear what YOU think the 2700 offers that the 350 doesn't? Codec yes, but I'd say that is pretty well balanced out by lower res chips, and the ability to add an external recorder to a 350 anyway. I've listed some ways in which the 350 is superior, what features would you point to that you think are superior in the 2700 rather than the 350?

If the choice of camera for those programmes was being made NOW (with the 350 in the frame), why now choose it over the 2700?

Quote "If the choice of camera for those programmes was being made NOW (with the 350 in the frame), why now choose it over the 2700?"
Couldn't really say for definite, but I know that it'd still be the choice. As I said, the 2 cameras would not even be discussed in the same sentence for blue chip nat hist. The EX3 not being 2/3" as I have said is actually a plus point for wildlife. Both the standard and wide lenses on the EX3 I've found very good, so that's not a problem, and stick an HJ40 on there and you've got even more reach than on a 2/3" camera. As to whether the HJ40 and HJ18x28 would suffer due to a smaller chip I don't know, maybe, and that could be a factor and would mean that the 350 is more suitable.
Even for those series already in production like Frozen Planet, if the 350 was so good it'd be worth trading in the 2700s for them as you'd get plenty of change. Can't really see it happening or even being considered though. The new big Africa series is in pre-production at the NHU now, due for completion in 2013 I think, guess what cameras they're going to buy/use? It's not the 350, and again I doubt whether it's even been considered.
Steve

David,

I think we've gone over these points before. I hope the OP has learned enough to make their own informed choice. No one camera can be right for every project. No question that the 350 offers a lot of flexibility at any price point, much less $20K with lens.

Yes, I'm very familiar with Psf, having shot with Sony cameras, both HD and SD in 24 and 30 fps for years. Problem is what can and does happen in post or distribution or broadcast where interlace artifacts are introduced for a myriad of reasons. With 720P broadcast, no need to worry about interlace artifacts, theoretically.

I will once again go over some reasons why I would go with a 2700 again:

P2 cards are proven since 2004 in the harshest production environments, the cards are very robust, have the fastest throughput and support multiple codecs, so are scalable for the future. Work flow is well supported by Panasonic and third party hardware and software manufacturers and known well. They offer their best codecs at lower price points, not protecting their high end models like Sony does, limiting the image quality of lower cost models such as the 350. They offer a five year warranty on many models.

Panasonic is known for great colorimetry, tonality, flesh tones. Sony cameras are known to come out of the box looking more videoish, harsher.

Varicam is a known name, synonymous with high end quality, frame rate flexibility and Film-Rec gammas. The P2 Varicams improve upon this reputation, and offer a look that many DP's consider to be the most filmic of 2/3" cameras. They can do frame rate ramping while recording.

CCD's are generally preferred over CMOS sensors because of skew and flash band artifacts in some cases with the latter.

Every high-end DIT I know, as well as colorist or online editor prefers 10-bit formats, which AVC-Intra provides. Most prefer I-Frame frame structure as well, and definitely 4:2:2 vs. 4:2:0. The 350 offers none of these things using the low end version of XDCAM HD.

All XDCAM codec versions are MPEG2 based vs. MPEG4 for Intra, considered a more efficient, modern codec.

Panasonic P2 cameras of the 2000 and 3000 series hold five cards for extended, uninterrupted recording time.

Three HD SDI outputs are found on the 2700 vs. only one for the 350. This requires looping or DA'ing for multiple monitors, necessary for higher end shoots.

Some camera operators don't like color viewfinders and/or LCD, LCOS viewfinders, no choice with the 350.

Bottom line, there is much to recommend both cameras, given similar pricing. For some, native 1080 sensors are required, even if CMOS artifacts could be evident from time to time, for others, CCD's and the most flexibility with off-speed frame rates is useful.

As a rental house, I already own an EX1 and am not convinced my clients would be willing to spend twice as much or more for a 350, even though I'm a big fan of 2/3" cameras. Indeed, many don't see the need for any 2/3" camera these days.

The OP asked for opinions and he got them, no one is right or wrong in their camera choice, both will provide lovely images in the right hands and often even in the wrong ones!

Jeff Regan
Shooting Star Video (http://www.ssv.com)

XDCAM-EX 35mbps codec is exactly one of the reasons why I want this cam. Not everybody is handing over footage to a network for final editing and grading. XDCAM-EX video smart renders, and then goes straight onto Blu-ray writable media as is. I can hand over the finished product faster than AVC Intra can transcode, and yet there remains the viable option of the nanoflash, and inexpensive media. I can't remotely fathom how the low end glass is a crippling factor. It's inexpensive, purely optional, plus it's autofocus.

AVC Intra isn't going onto Blu-ray media the way it is, so it's workflow is built-in handicapped when time is money. I see lots of great live HDTV 1080 video from CBS, the NFL, the Masters, none of it is going to the edit booth, none of it is being received at 100 mbps 10 bit. It arrives at a mere 10-12 mbps, looks as good as anything broadcast on National Geographic (arguably better to me), and Blu-ray remains the premier display format for the end user.

We don't even know if CMOS per the PMW350 will have the previous artifacting problems. It certainly brings some documented benefits, lighter weight, lower power consumption, low noise and high sensitivity.

If the Panasonic colorimetry is so unique, you could question the need for AVC Intra since the out of the box look should be signature, and no one needs 100 mbps just to capture detail and artifact-free motion for anything but extreme circumstances.

XDCAM-EX speed is a 'quality' unto itself, as is the inherent suitability for Blu-ray. It is exactly one of the desired features for me that separates it from the 700 and 800 line which have the CCD, writable disk workflow, and higher bit rate preferred by many in broadcast.

Some good points Tom for sure, and certainly there are different needs for different uses.
AVC Intra is preferred because it gives extra resolution and is 10 bit.
Steve

Yes but then it goes out over broadcast at much lower bitrate and 8 bit. The advantage, if it was not used up front for grading, either evaporated or was not necessary. If the footage was going to be graded, than any high bitrate codec (i.e. nanoflash) can be pushed to achieve the Panasonic look, colorimetry in post.

In other words, XDCAM-EX codec is not mutually exclusive of higher performance codecs captured from the SDI via Nanoflash and the like, but the 2700 doesn't have a codec that ports straight to Blu-ray without time consuming downsampling.

Many people don't care a whit about Blu-ray, that is true. But I do.

XDCAM-EX 35mbps codec is exactly one of the reasons why I want this cam. Not everybody is handing over footage to a network for final editing and grading.
snip
I can't remotely fathom how the low end glass is a crippling factor. It's inexpensive, purely optional, plus it's autofocus.
snip
I see lots of great live HDTV 1080 video from CBS, the NFL, the Masters, none of it is going to the edit booth, none of it is being received at 100 mbps 10 bit. It arrives at a mere 10-12 mbps, looks as good as anything broadcast on National Geographic (arguably better to me), and Blu-ray remains the premier display format for the end user.
snip
If the Panasonic colorimetry is so unique, you could question the need for AVC Intra since the out of the box look should be signature, and no one needs 100 mbps just to capture detail and artifact-free motion for anything but extreme circumstances.


Tom,

Just so you know, DVCPRO HD and AVC-Intra 100 record at only 40Mbps in 720/24PN. With Panasonic E series P2 cards and PCD35 P2 reader into a desktop with PCIe slot, the download times are amazingly fast due to the 1.2Gbps throughput on the cards.

Low-end glass would be a non-starter for high end 2/3" HD cameras in the rental business. I have DP clients who insist upon Fujinon HA series HD lenses, no XA or ZA, even though they are also very expensive lenses. To put an HA lens or any higher end glass on a 350 would not make too much sense, unless the rental rate went way up, which would be out of line with the specs of an 8-bit 4:2:0 codec.

Please don't equate delivery bit rates with acquisition, what we see off-air or cable or satellite in HD often falls apart quickly when the low bit rate codecs are stressed. No network would normally accept this as a source file to edit with. Blu Ray looks as good as it does because it is painstakingly compressed scene by scene at great expense by compressionists, most often using MPEG4 and VC-1 codecs vs. MPEG2.

Any camera can take advantage of having 4X the gray scale steps, 8-bit at 256 shades of gray vs. 1024 shades for 10-bit can pay dividends in post and most every non-live, non-news show has some grading and color correction done if there is editing.

Once I entered scene files for my EX1, like any Sony digital camera I've owned, I was very happy with the look. I do have some clients that actually want to shoot 1080/60i and I believe this is where the XDCAM EX codec really shows its weakness.

AVC-Intra is a lot easier to deal with now with Final Cut Studio 7, Adobe CS 4, Avid 4.5, etc, but I can't speak to the Blu Ray convenience--burning a Blu Ray is just a step that happens after the final edit, no bigger hurry than any other deliverable for my clients, normally.

Jeff Regan
Shooting Star Video (http://www.ssv.com)

The 350 is also far more suitable for use with professional accessories such as radiomics.

David,

What do you mean by this and "next generation" radio mics? The Panasonic 2000 and 3000 series camera have a slot for 2-channel digital radio mic. receivers such as Lectrosonics.

Jeff Regan
Shooting Star Video (http://www.ssv.com)

Tom,

Just so you know, DVCPRO HD and AVC-Intra 100 record at only 40Mbps in 720/24PN. With Panasonic E series P2 cards and PCD35 P2 reader into a desktop with PCIe slot, the download times are amazingly fast due to the 1.2Gbps throughput on the cards.

I get this, but DVCPRO HD still has to transcode to mpeg-2 or AVC or VC-1 to go on Blu-ray. XDCAM-EX needs no transcoding, it's already mpeg-2.

Blu Ray looks as good as it does because it is painstakingly compressed scene by scene at great expense by compressionists, most often using MPEG4 and VC-1 codecs vs. MPEG2.

Agreed, yet taking transcoding off the table for Blu-ray with DVCPRO HD is not an option. The argument I made for XDCAM-EX is for speed. My client gets the finished product delivered on Blu-ray faster. If speed is not important, nothing precludes using the Nanoflash for 4:2:2 with the PMW-350, if desired.

I would not argue the PMW-350 is ideal for broadcast, or for rentals. But if you do events, or electronic news gathering run 'n gun, you might ask where else besides the kit lens you go to find 2/3 inch auto focus?

Up front, 10 bit is an advantage in gray scale acquisition agreed, just as 1920x1080 is an advantage in resolution. But the profile adopted for Blu-ray is 8 bit 4:2:0, so in the end, the best efforts from grayscale A/D get truncated to 8, but the advantage over 720p from having twice the number of luminance samples with 1080p is always maintained.

We also don't know fully what the PMW-350 is truly capable of. We have seen some stunning video from Alister Chapman shot in tough light, and we already know that 1080 XDCAM with Nanoflash is accepted by all the broadcasters. It took a few years but in the DSLR world, most manufacturers eventually settled on CMOS. At some point for video, the problem of CMOS artifacting will be solved. Some observations already seem to point to the PMW-350 representing an improvement in that regard.

For CCD, it could be the beginning of the end, perhaps not. The PMW-350 could be the end of the beginning for CMOS.

Tom,

I agree with most everything you posted. It is worth considering that feature films on Blu Ray are mastered from 10-bit HDCAM SR and D-5 formats, so there is some usefulness for 10-bit, besides the obvious grading and color correction advantages. It doesn't make sense to dumb down acquisition and mastering formats because deliverables have lower specs.

The nanoFlash recorder disadvantages are external vs. internal recording ergonomics and complexity, 8-bit, need for very high bit rates compared to AVC-Intra 100 I-Frame due to MPEG2.

Varicam tape and P2 cameras are a proven product, delivering consistently for many years in the case of the former. We just don't know yet how the 350 will perform as far as CMOS, but we are already well aware of the XDCAM EX codec's limitations.

I have no doubt the 350 would be a phenomenal events camera, an area where its codec limitations won't be an issue. High-end production is still an open question, but for many 4:2:0 and 8-bit color depth would rule it out. Similarly, there are projects where a 720P native chip set of the 2700 could rule it out, however, this same chip set has been used successfully for many years for features, commercials, episodic, docs, etc.

CMOS may be the future imager for video cameras, primarily due to cost, but the amazing Sony F35 and Panavision Genesis proves that a single large sensor CCD is possible and quite impressive--albeit expensive and power hungry at present.

Jeff Regan
Shooting Star Video (http://www.ssv.com)

I hate it when people claim that codec "A" is "the best codec" or codec "C" is the "the highest possible quality". Such claims for any codec are nonsense and misguided. All codecs other than uncompressed are compromises trying to balance quality with efficiency and speed of workflow. They all have pro's and con's. Sure you can have a favorite or preferred codec but that doesn't mean it's the "best" codec.

AVC-Intra is a terribly inefficient codec introduced as far as I can tell, as a stop gap to replace the aging DVCPPRO HD codec. As I see it Panasonic choose to use Mpeg4 because as usual they would not use the same codec, Mpeg2, as Sony, JVC, Canon etc. Then they realised that Mpeg4 (AVC) was so hard to edit natively that no one would want it for professional applications. So they took an efficient codec, Mpeg4 (which is a lot closer to Mpeg2 than most people realise), threw away the stuff that makes it work well (and thus hard to edit) and started a smear campaign on long Gop codecs to justify why they use an inefficient I frame codec. To claim AVC-I to be "vastly superior" or "the best codec" is nonsense. They have only just made it possible to edit native AVC-I in FCP while native XDCAM EX editing was possible from day one. If it really is the "best codec" then surely people would be editing with it and apple would not have had to introduce ProRes and Avid wouldn't recommend transcoding it to DNxHD etc.
I'm not saying that XDCAM EX is better, it's just the difference is not nearly as large as some like to claim or think. If you shoot progressive the difference between 4:2:0 and 4:2:2 is negligible, the difference in interlace is I agree larger.

As I have already said the EBU found XDCAM HD at 50Mb/s and AVC-I to offer similar performance and it was noted that the 10 bits of AVC-I offered no noticeable advantage for acquisition. This isn't my view or opinion but that of a panel of experts drawn from broadcasters across Europe. 10 Bit will only bring an advantage when the appropriate amount of data is used to record the extra bits, there is no such thing as a free lunch. While this is a comparison of AVC-I and 50Mb/s XDCAM and not 35Mb/s EX it does demonstrate that 10 Bit isn't everything. Yes for post work I use 10 bit, but at 220Mb/s where I really can see a difference.

The quality of any video system starts at the front end, not the rear. The lens being the first quality limiter followed by the sensor. I'm sorry but I just can't understand why anyone would want to invest now in an old design 720P front end. It cannot ever be magically turned into a 1080P front end. Yes there are differences between CCD and CMOS, but again these are largely over exaggerated differences that in the real world are not visible. The exception would be filming lots of strobe lighting etc where CCD would be an advantage, but even then I would want 1080 CCD not 720. You don't have to shoot 1080 with a 1080 camera, but at least you have the choice.

CMOS is being used day in day out to produce very high quality images including high end drama. The Arri D21 is CMOS as is RED. Would I prefer CCD on the 350? No. The power, weight and cost advantages far outweigh the different way that flash and strobe lights are handled. Remember, tube cameras used to respond to strobe lights and skew in a similar fashion, yet when CCD cameras first came along it was the CCD's that were seen by some to be "wrong".

At the end of the day you can take a native 1080 camera and shoot 1080 or 720. You can add an external recorder to up the bit rate, or a DA to add additional outputs. But you can never make a 720 camera into a 1080 one.
In the UK one of the most profitable SD rental camera lines is the DSR570/DSR450 DVCAM range, these are often more profitable than Digibeta or HDCAM because they are used day-in day-out in vast numbers as work horse cameras. They may not be the "best" cameras in the world or use the "best" codec but they are reliable workhorse cameras that if used well produce picture just about indistinguishable from the more expensive Digibeta cameras. This is the slot the PMW-350 is designed to fill and I believe it will turn heads as the image quality is stunning. It might not have the "best" codec, but what it will do is deliver 1080 HD out of the box that rivals or even surpasses many of the cameras currently on the market and offer a comprehensive range of modes and functions that will make it suitable for a huge range of productions. I believe it will make an excellent workhorse camera. It is expandable via external recorders should you choose. There is also an internal expansion slot for future upgrades, so who knows what will come in the future.

Jeff, you can put a mono VF on the 350. It has 2 viewfinder connectors, one for the colour VF and one for a CRT viewfinder. Using a NanoFlash is fit and forget. It powers on with the camera, it goes into record with the camera. It's solid state, it is not a firestore.

Alister, didn't I read somewhere that the BW viewfinder you can put on the 350 is an SD one? Anyhow, presumably the colour one can be switched to bw so probably not a big deal.
Funnily enough been playing a bit with the EX3 over the Christmas period and I do have a small preference for having the VF in colour now! It's not the greatest VF though, compared to all the other small cams on the market it's amazing, but not great compared to the full size cams.
It'll be interesting to see just what impact the 350 makes in high end work, but for wildlife I really don't expect to need to do a BBC training course on it any day soon, but that's just my guess.
Steve

You can put any DXF viewfinder on the 350. The DXF-20W is designed for HD. it is the viewfinder used on the PDW-F355 etc and is specified as "over" 480 lines resolution which is actually very close to the 500 lines offered by the HDVF-20A Sony's high end HD CRT finder.

Having used cameras with the high end C35W colour finder and the PMW-350 I would say there is little to choose between them. The C35W has a better range of options and controls but there is only a small difference in image quality.

Yes it will be interesting to see whether the 350 can make inroads into natural history.

Alister,

The specs for AVC-Intra 100 speak for themselves.

In my world 4:2:0 vs. 4:2:2 is important, not as big a deal as 4:1:1 vs. 4:2:2, but still a difference. I own a DSR-450WS, had two at one point, had a DSR-500. Digi Beta is clearly superior due to its color space and bit depth--it's not a subtle difference.

If Sony thought XDCAM HD and XDCAM EX were so great, why the need to introduce XDCAM 422? Obviously the 4:2:2 color space was seen to be worthwhile, and required raising the data rate to 50Mbps.

What we know about high end studio quality mastering formats is that they are 10-bit, high bit rate, 4:2:2 or 4:4:4, full sample, I-Frame. AVC-Intra 100 is the acquisition format that most closely matches a studio format specs wise.

AVC-Intra 100 is the only codec found inside a 1/3" and 2/3" one-piece camera that is 4:2:2, full sample, 10-bit other than an SRW-9000 for $100K. It is a Native, progressive codec that doesn't record 24P and 30P over 720/60P or 1080/60i to be compatible with legacy tape formats. MPEG4 is much newer and more efficient than MPEG2.

My editor did grading and color correction tests with Intra 100 vs. DVCPRO HD and found the former capable of being pushed much further. Many editors prefer I-Frame structure.

When editing for deliverable formats like HDCAM SR or D-5, both of which are 10-bit, it makes sense to shoot in a 10-bit format to begin with. If shooting 8-bit, the information is gone at the front end.

Sony was the first camera manufacturer to deliver an AVC HD camcorder, so they have embraced AVC codecs due to their superior efficiency vs. MPEG2. This is the same reason AVC h.264 is used with Blu Ray so often.

Final Cut Pro 7, Adobe CS4, Edius 4.5 and Avid 4.5 make editing in AVC-Intra much easier, some natively.

Regarding CMOS artifacts, I have clients who won't rent my EX1 because of the image skew(jello cam) issues. This is a bigger concern to them than the flash band artifacts.

I agree that the lens is the first quality limiter, and that is why I would be very reticent about asking a DP client to accept a $1600 HD lens on a 2/3" camera.

Again, Varicam is a proven name that has been improved in the P2 versions. If a full raster camera is paramount, there is the 3700, albeit not as flexible in frame rates and no 720P, but I have no doubt a 2700 and 3700 are taken more seriously for network episodic work than an XDCAM EX camera would be, primarily due to the master quality AVC-Intra 100 codec. A Fox Network prime time show called "Dollhouse" was shot in 35mm the first season and switched to 3700's and 2700's for the second. The show looks very good.

Jeff Regan
Shooting Star Video (http://www.ssv.com)

The specs for AVC-Intra 100 speak for themselves.

If Sony thought XDCAM HD and XDCAM EX were so great, why the need to introduce XDCAM 422? Obviously the 4:2:2 color space was seen to be worthwhile, and required raising the data rate to 50Mbps.

The specs for AVC-Intra speak for themselves but they don't yet speak to me. You asked a great question about why Sony felt the need to introduce XDCAM 422, but left without comment about why if AVC-Intra is the best codec, people would be editing with it and apple would not have had to introduce ProRes and Avid wouldn't recommend transcoding it to DNxHD etc. What it says to me is that the question was thrown up to mute a point that was made. The reasons in support of XDCAM-EX have been stated, speed in native editing and smart render to Blu-ray target.

Neither is the argument persuasive that AVC-Intra should be compelling because you would have to start with $100,000 to find similar codec features in the Sony line in the SRW-9000. This goes more to brand loyalty than logic.

Perhaps I would find those points compelling, if it could be justified why those extra 2 bits of grayscale are worth throwing away half of your luminance samples for 720p versus 1080p. People right here DO see that difference. I can't see the justification for 720p, nor can I find the answer by turning to the past and pointing to high end studio mastering formats, full sample 4:4:4 like HDCam SR when it appears we are trending away from that. There must be a reason. My guess would be that in the post tape world order, there is no longer room for inefficiency that wastes space on expensive storage media, for extremely minor benefits.


Sony was the first camera manufacturer to deliver an AVC HD camcorder, so they have embraced AVC codecs due to their superior efficiency vs. MPEG2. This is the same reason AVC h.264 is used with Blu Ray so often.

Although your two points are correct, it is not the reason h.264 was adopted by Blu-ray. In fact, Blu-ray was conceived from the beginning to accommodate a two hour recording in mpeg-2. The h.264 was only adopted later by the BR association in response to the competition from rival HD DVD and Microsoft, that a new optical format was not needed because of the recent advances in codec compression, to witness VC-1. If not for that competitive push, h.264 would likely not have been adopted by Blu-ray at all.


Regarding CMOS artifacts, I have clients who won't rent my EX1 because of the image skew(jello cam) issues. This is a bigger concern to them than the flash band artifacts.

You know your customers best. I would only counter that I own and use the EX1 now, am aware of the potential skew issues, they would not stand in the way of repurchasing into the CMOS technology. The benefits of light weight, low power, low noise and high sensitivity (and low cost) justify the use of CMOS for me.

I agree that the lens is the first quality limiter, and that is why I would be very reticent about asking a DP client to accept a $1600 HD lens on a 2/3" camera.

Your point is understandable, but no one is forced to buy this lens.


Again, Varicam is a proven name that has been improved in the P2 versions. If a full raster camera is paramount, there is the 3700, albeit not as flexible in frame rates and no 720P, but I have no doubt a 2700 and 3700 are taken more seriously for network episodic work than an XDCAM EX camera would be, primarily due to the master quality AVC-Intra 100 codec. A Fox Network prime time show called "Dollhouse" was shot in 35mm the first season and switched to 3700's and 2700's for the second. The show looks very good.

I'm looking for more than just reputation or brand name recognition. There are tangible benefits with CMOS. The PMW350 raises the bar in meaningful ways, reduced power consumption, lighter weight, lower cost 2/3 inch 1080p acquisition, lower cost storage per GB on a variety of solid state media types. It includes its own extremely fast and friendly editing codec that ports straight to Blu-ray with smart rendering, yet remains compatible with high bitrate 4:2:2 solid state capture solutions, i.e. Nanoflash.

AVC-Intra 100 is the only codec found inside a 1/3" and 2/3" one-piece camera that is 4:2:2, full sample, 10-bit other than an SRW-9000 for $100K. It is a Native, progressive codec that doesn't record 24P and 30P over 720/60P or 1080/60i to be compatible with legacy tape formats. MPEG4 is much newer and more efficient than MPEG2.

And that 1/3" camera is not approved for use by Discovery or Nat Geo while the EX's are. XDCAM, HDV, MPEG2, H264 and many other codecs can be native progressive recording just the actual used frames, there is nothing special about the way AVC-I encodes progressive. PsF is the way the signal is output over HDSDi as originally there was no provision in the the HDSDi specs for 1080P. Panasonic do exactly the same using PsF for 1080P over HDSDi, the recording itself is true progressive. PsF is part of the ITU part 709 specifications for HD broadcast systems.

My editor did grading and color correction tests with Intra 100 vs. DVCPRO HD and found the former capable of being pushed much further. Many editors prefer I-Frame structure.

Well there are no surprises there. DVCPRO HD is no where near full frame and that causes all kinds of concatenation issues. This just proves my point that just because a codec uses "X" bit rate or 4:2:2 over 4:2:0 it doesn't automatically make it a better codec.

When editing for deliverable formats like HDCAM SR or D-5, both of which are 10-bit, it makes sense to shoot in a 10-bit format to begin with. If shooting 8-bit, the information is gone at the front end. But ONLY if the data isn't compromised by not using a high enough bit rate to allow for the extra bits to be recorded without raising the compression ratio. As I keep saying this was a point raised by the EBU in their tests with AVC-I and XDCAM HD422.

Sony was the first camera manufacturer to deliver an AVC HD camcorder, so they have embraced AVC codecs due to their superior efficiency vs. MPEG2. This is the same reason AVC h.264 is used with Blu Ray so often. Yes in consumer cameras and it's a swine to edit which normally means you end up having to transcode it. It is now creeping in to pro level cameras but even at it's maximum data rate of 24Mb/s the jury is still out as to whether it's any better than HDV. AVC or Mpeg4 was designed to excel at very low bit rates and high compression ratios, which is where it is at it's best. It's amazing how good a 4Mb/s 1080P clip can look. That's why it ended up in consumer cameras where people want to be able to record hours and hours on a single SD card or memory stick. At low bit rates/high compression ratios it is very efficient. But scale that up to 24Mb/s, AVC's maximum, where you would think it's going to kick HDV's butt and to be honest it's not really any better. Some scenes can look terrible where HDV would still look good. I have yet to see an AVC camera that produces superior images to a comparable HDV camera. At high bit rates/lower compression ratios AVC is no more efficient than MPEG2. If AVC really was so good then I'm sure that Canon and JVC would have adopted it in their pro level cameras, but they have not. Then you have to remember that AVC-I has almost non of the efficiency adding attributes of AVC. AVC (mepg4) was designed as a long GoP codec. It's the long GoP that make it so efficient, but AVC-I has no long GoP and thus the bulk of the efficiency has gone.

Final Cut Pro 7, Adobe CS4, Edius 4.5 and Avid 4.5 make editing in AVC-Intra much easier, some natively. They all edit XDCAM natively.

I agree that the lens is the first quality limiter, and that is why I would be very reticent about asking a DP client to accept a $1600 HD lens on a 2/3" camera.Then don't use it. The 350 has a standard B4 mount so just use whatever lens you feel is appropriate. You can by the camera without the lens if you wish.

Well my friend and officemate Don Lenzer got over to see the camera at Abel today and followed up with a report. It certainly puts on hold his plans for the 2700 although to be fair he was leaning against the trade in before this. He isn't rushing to buy the 350 but is holding off on the 2700.

Tom, Alister,

I think a big reason Sony tried to stay with MPEG2(and needed 50Gb dual layer discs because of that) is due to not wanting to pay licensing fees to Microsoft for VC-1 or fees to use MPEG4 h.264, ultimately they did so because of pressure from the studios who didn't want to have multiple encoding schemes to deal with when HD DVD was alive and because of the extra content needed for special features--even 50Gb wasn't adequate with MPEG2. They had to do this with consumer cameras due to that market wanting long record times.

It wouldn't surprise me if Canon and JVC don't want to pay for licensing new codecs either.

I agree that if 10-bit is of huge importance, it is conceivable that full raster 1080 is also important and that is why Panasonic offers the HPX300, 3000 and 3700. 2 extra bits translates to 4X more shades of gray, not inconsequential.

The 300 had terrible CMOS skewing in 1080/24P, not sure if this has been improved with new firmware. EX1 and EX3 global shutter issues are well known, now we just need to hear how the 350 performs. I Sony to keep improving CMOS technology and the 350 may already show benefits.

I think we are dealing with two different strata of production requirements, the clients that seek the Varicam look and functionality would not consider 4:2:0 low bit rate XDCAM EX. It's really that simple. Nor would they consider a $1600 lens. Nor would they want to take a chance with rolling shutter artifacts.

So while you may not care about proven CCD and Varicam history, many producers do.

Could this change? Of course. We've already seen the RED One make a big impact on film style production and video DSLR's are being used in places we would have never expected.

This doesn't change the fact that high end productions seek 10-bit, 4:4:4 or at least 4:2:2 at high bit rates, just as they are more likely to seek full raster sensors going forward. Whether tape or solid state doesn't make any difference--even if SR and D-5 went away, similar or better specs would be expected with whatever media.

This is why Panasonic is developing AVC Ultra at 200Mbps with 4:4:4, 1080/60P, 12-bit recording with 3D capability as well. This would use the same P2 card media currently in use.

I would be happy to see the 350 get some traction in my market and I would offer it in my rental inventory if I could get at least twice as much day rate than for my EX1, using the stock lens. I have no doubt that many people will be happy with the 350 and it should do exactly what they want, but to dismiss the HPX2700 Varicam due to native 720P CCD's or AVC-Intra 100 due to it requiring transcoding for some NLE's or is "inefficient"(even though a much newer codec than the long in the tooth MPEG2 variants) is not valid.

I think we've all made a good case for either camera, and they both deserve consideration.
Ultimately, the types of projects being shot daily with the cameras and formats is what is most telling. So far, Varicam has seen continuous usage for many years on all types of high end projects. EX1 and EX3 are doing projects that one wouldn't expect to see them on, such as a B-Camera for "Trauma" or "Public Enemy". I would be concerned right now if I owned an F900R or 700/800 because of the 350.

Of course, I'm concerned by all palmcorders, RED One and video DSLR's, not to mention Flip HD at this point! It seems 2/3" cameras just aren't sexy anymore.

Jeff Regan
Shooting Star Video (http://www.ssv.com)