XLH2 features you want.

[Yi Fong Yu]

1. cheap=$10/2 hour 1080p media. that's pretty cheap (like miniDV) prices.

2. i said abandon existing designs (like HDV) and come up with a tape transport that is able to record uncompressed 1080p for 1-2 hours for $10/tape. i'd be happy by then. but right now, we don't have that option. i'm no engineer but i'm damned sure it can be done. tape is old, but it is much cheaper than new media and can hold more capacity for less $. it's a quick&effective way.

3. if they can continue to produce cheaper, higher capacity 2.5" drives, that would be the one cam operators would love. lugging around 4 3.5 drives ain't my idea of fun while shooting. looks innocent enough until you pack it onto the cam and the cam gets damn heavy.

4. but how much uncompressed footage do we need for consumer-prosumer-level needs? i say record&encode to a mpeg4 codec live. save space&easy editing. all i need to know is when i project 1080p to a 250' screen, it'll look good 'nuff. 25mbps is good 'nuff, 100 is dimishing returns.

Originally Posted by Alexander Ibrahim

That depends entirely on your definition of cheap enough. Given how much I spend on tape I am looking for options.

In case you missed my sick sarchasm (sic) HDV not only meets but exceeds your specifications.

Also, tape transports are incredibly complex systems. I suggest getting an VHS deck and a service manual for it to see what I mean. Break the deck a few times, then fix it to learn. Remember- that's a consumer technology that's over a quarter of a century old.

Apple has also been doing a rather spry business with mobile hard drives on this iPod thing of theirs, you may have heard. It uses 1.5" hard drives in capacities up to 60GB... and they are fast enough for video recording.

Uh... because greater bandwith recordings have higher fidelity.

[Jacques Mersereau]

>The thing is this: I believe that people from the manufacturers
>R&D and engineering departments lurk in pretty much every
>public forum. It is in their interests. They pay attention to
>off the wall threads like this because they need feature ideas.


Canon does indeed listen and respond.
When Mr. Hurd asked what features we wanted to give to Canon
for the next XL series, you'll find in the archives that I said they should
take two paths. One for an 'evolutionary' model of XL2 and another
for a revolutionary model of HD. It was I who kept asking for SDI and
HD-SDI outputs. They did listen, just as several years before that
when complaining about the 16X lens I suggested that since the lens
was servo controlled, Canon should put a lens setting 'remember button'
to allow super accurate rack focus between two subjects,
and _also add a timing factor_ for the duration.
Six months later this feature appeared on on 'pro' Canon
lens. Controlled rack focus will now also be on the HD1.

It was also I who several years ago, being a early flag carrier for
720P, was the first to suggest on the DV-L that bandwidth could be saved by
repeating frames for 24P or any other frame rate by rewriting the codec.
Six months later Panasonic came out with a box that would do just that.
This feature is now being incorperated on the HVX200 and will allow
a far longer record time than straight 720P@60fps.

>I believe (well its more like blind faith) that if we post
>reasonable and well thought out ideas we may see them
>in future products.

True enough, so where are my royalties?
(I'll be a beta tester on a HD1 Canon :)
In fact, I am also highly aware that at any time, any great new invention
is being worked on by several other people, but I think I can take
credit for at least one of the above features.

Onion skinning is and will be another.

[Jacques Mersereau]

>2. i said abandon existing designs (like HDV) and come up with a
>tape transport that is able to record uncompressed 1080p for
>1-2 hours for $10/tape.

I have to say that this is as much a pipe dream as my 1 TB memory
pack. Jan has reported that a DVCPro tape transport costs over
$7,000.00.

The main problems with storage are size, cost, speed and reliability.
Hard disks are close, but tape is still the storage king. Removable
disk is the only real contender for now imo. When 30g P2 cards are
$40 each, they may challenge blue laser. That may never happen me
thinks.

[Yi Fong Yu]

i also agree that HD is a contender but methinks tape is physically lighter than HD, hence my concern for cam ops who are mobile.

i wouldn't mind 2 of the 500GB SATA2 (NCQ) hard drives in raid0. then we can plug it into editing machine live (hotswappable is another SATA benefit). though the array-moving would be an issue.

[Alexander Ibrahim]

Originally Posted by Yi Fong Yu

1. cheap=$10/2 hour 1080p media. that's pretty cheap (like miniDV) prices.

2. i said abandon existing designs (like HDV) and come up with a tape transport that is able to record uncompressed 1080p for 1-2 hours for $10/tape. i'd be happy by then. but right now, we don't have that option. i'm no engineer but i'm damned sure it can be done. tape is old, but it is much cheaper than new media and can hold more capacity for less $. it's a quick&effective way.

This is not going to happen unless there is a breakthrough in how we store data on tape.

HDV is a VERY good tape format, with a good transport design that stores 25Mbps. (You argue for 25Mbps later which makes no sense, but I'll come to that.)

Nobody is going to design a new transport for the same task. Such design is extraordinarily expensive.

Further more, you have not considered the costs of creating a manufacturing infrastructure for your new tape design. You have to manufacture the tape, the shelss, the transports and all the parts. You have to make machines to handle those functions, and tools for the decks, tapes and machines you use in manufacturing.

There will never be a huge market for tape transport manufacturing machines, so these machines will likely be hand made from high end materials: In other words ungodly expensive.

Why do you think VHS still exists ? JVC retooled it to make SVHS, DVHS and D-9.

Why is Sony still making NEW Betacam transport based products ?

Panasonic makes a D-5HD system that can record full HD resolution 4:2:2 video. It's only $99,000 USD. Blank tapes cost $325.95 for 124 minutes.

Sony makes HDCAM SR (which is a Betacam transport based product), which can record full HD resolution at 4:4:4. The SRW5000 deck is $88,000 USD. The Dual Link SDI I/O board is an additional $10,000. A format conversion board is $10,000. The ability to play back Digital Betacam adds on another $4,700. Realistically you need the SDI and format conversion board, so you are looking at $108,000 for a basic HDCAM SR deck.

HDCAM SR 40 minute tapes are $71.67 each. For two hours you need 3.

That sir is the state of the art. The best we can do for now. It costs so much because it is VERY HARD TO DO.

3. if they can continue to produce cheaper, higher capacity 2.5" drives, that would be the one cam operators would love. lugging around 4 3.5 drives ain't my idea of fun while shooting. looks innocent enough until you pack it onto the cam and the cam gets damn heavy.

They already make high capacity 2.5" drives, as I already wrote. They also make high capacity 1.5" drives. See the iPod, also as I already wrote. (I am not entirely sure if the 1.5" drives can handle 100Mbps, but they can handle 50Mbps readily.)

Why don't you try projecting HD first. Then I think you'll find that 25Mbps is not enough at all for a 250' screen. You will see big bad artifacts.

How much uncompressed HD do you need ? That depends entirely on your application.

In a two hour "effect's driven" feature you only need uncompressed quality for about 5 minutes of on screen footage. A film like Lord of the Rings needs this for almost every frame.

For a home video you need absolutely none.

Most of us need something in between.

Personally I can shoot a format like HDV 75% of the time, because that footage will be edited cuts only with no effects. 20-24% of the time I need higher quality footage, but it can be compressed. DVCPRO HD is absolutely fine. (I shoot a lot more on sequences that need high quality footage, often because I need the coverage, and also because I typically need more takes.) The last little bit requires the highest quality I can budget for, so that's how often I need Uncompressed HD or better if available.

HDV is pretty good, but it artifacts more than DV. You can see it on even modest screens.

Heck, while I am at it HDCAM, at 135 or so Mbps isn't good enough for projection. Star Wars Episode 2 was shot using what were essentially HDCAM SR cameras, but was recorded almost entirely as uncompressed dual link SDI.

DV was widely disregarded as a medium for theatrical release. There are quite a few films that dispute that notion. Still, while the world stares agog at a feature film shot with an XL-1 and is awed, they fail to understand WHERE in the production flow the truly awesome technology lies.

In post ladies and gentlemen. They upconvert the DV footage to 2K or 4K resolution and 4:4:4 color then they can do some amazing things.

Lord of the Rings went through a similiar process using Discreet's lustre.

http://www.uemedia.net/CPC/digitalci...le_13360.shtml

Take a LONG time and study image quality. Shoot some DV, HDV, medium format and 35mm of the same subjects and compare them carefully yourself. Look not for what is the same, but rather what is different.

Whenever a trade show for video/film comes to your area go look carefully at the cameras vendors are selling. You can see HDCAM across from DVCPRO HD and next to HDV and DV. Compare them critically.

Create in your mind specific images and go create them with a couple of different cameras. I'd pick a specific scene from a movie you like. Once you get your results, post process them to try and match your vision exactly. Take all the time in the world if you like.

The point is to train your eyes.

[Alexander Ibrahim]

Originally Posted by Yi Fong Yu

i also agree that HD is a contender but methinks tape is physically lighter than HD, hence my concern for cam ops who are mobile.

i wouldn't mind 2 of the 500GB SATA2 (NCQ) hard drives in raid0. then we can plug it into editing machine live (hotswappable is another SATA benefit). though the array-moving would be an issue.

???

I think you are saying you want the best of both worlds. That is cool I suppose.

Yes tape is lighter than a hard disk, even a 1.5" drive. But the tape is not the whole thing.

You have to account for the weight of the tape transport. I am suggesting that we completely replace the tape transport. That leaves the camera with only solid state components and some I/O interfaces and controls.

That will make your camera much more reliable.

All the expensive moving parts will be in the hard disk, which as I mentioned before has all sorts of economies of scale plus demand for ruggedized versions already going. It only makes sense for manufacturers to start thinking right now about taking advantage of that industry.

Instead of tape manufacturers can build hot swappable SATA enclosures for hard drives. If they get computer manufacturers on board you could pull a drive out of your camera, plug it into your computer and start editing.

So, its a weight neutral solution compared to tape, and it increases the camera's lifetime and decreases camera maintenance requirements.

All the mechanical complexity of the system will be completely contained in your media and the lens. Two user replaceable parts. Anything that brings service issues under my control as a small studio owner is a GOOD thing.

[Nick Hiltgen]

Alexander you make a few good (albeit long) points. One thing that stuck out specifically in my mind was the ipod mention. With the new Nano and the shuffle we've seen them move from hard drives to flash, Do you believe it's possible that someone (some brand or other) could create a flash based raid array (distributing the information to a series of NVRAM flash chips) I mean two nano's are about the same size as a P2card (though that's really more of an estimate but if one were to remove the screen I'm sure you'd be able to fit all the ram in there) and that would be 8gb vs 4 on the P2. I just can't help but wonder if it's possible to just use that technology to create a capturing device with no moving parts? not to mention 500 bucks vs what is it now 4k?

[Alexander Ibrahim]

Originally Posted by Nick Hiltgen

Alexander you make a few good (albeit long) points. One thing that stuck out specifically in my mind was the ipod mention. With the new Nano and the shuffle we've seen them move from hard drives to flash, Do you believe it's possible that someone (some brand or other) could create a flash based raid array (distributing the information to a series of NVRAM flash chips) I mean two nano's are about the same size as a P2card (though that's really more of an estimate but if one were to remove the screen I'm sure you'd be able to fit all the ram in there) and that would be 8gb vs 4 on the P2. I just can't help but wonder if it's possible to just use that technology to create a capturing device with no moving parts? not to mention 500 bucks vs what is it now 4k?

Sorry about being long winded, I have no time to be succinct!

That is exactly what a P2 card is. A striped array of flash drives.

I expect we will see 16GB P2 cards pretty soon now. We should see 32GB ones in the near future, 64 and 128 in late 2006, and 256GB in late 2007.

I also expect we'll see companies come out with competitive systems. That is a good thing because competition always brings price down.

Still, in the near term price of flash RAM is so high I expect we'll really see more hard drive solutions. Apple is consuming about 40% of the world's flash production just for iPods! (Shuffle and Nano) (Actually they are taking 50% of Samsung's production, and I think Samsung produces about 80% of the world's flash.)

[Terry Johnson]

3. any of the mpeg4 variant codecs.
.

somn like this is ~2009?[/QUOTE]

With any of the mpeg4 variant codec's it should be possible to output 4.2.2 HD video to the dv tape deck at 25 Mbs. This would solve the long term video storage problems present with the Panasonic P2 cameras and be in the format which is required for the forthcoming Blue Laser disc's.

The difficulty is real time encoding to mpeg4. I'm sure this will take multi thread microprocessors ( which will soon be available) and a good deal of cooling and battery power but it would solve a large number of problems. The battery and processor could be located in a back pack if need be.

With the 1.5 M pixel ccds and Canon's 20X lens and image stabilization this would be a killer product.

[Alexander Ibrahim]

Originally Posted by Terry Johnson

With any of the mpeg4 variant codec's it should be possible to output 4.2.2 HD video to the dv tape deck at 25 Mbs.

MPEG-4 is a container format, like .avi Quicktime and Windows Media.

I could very easily write a codec that encodes full resolution HD video at 4:4:4 to a 1kbps stream. That would mean nearly complete image loss however.

In addition you have to consider the requirements for MPEG 4 Studio profile recording. That's a fancy standard for specifying what a codec must do in order for it to be suitable for studio use- i.e. for editing.

For example, h.264. That's one heck of a codec, but it isn't studio profile by a long shot.

You can't cut it at a particular frame with re-encoding both sides of the cut. Heck, have you tried scrubbing h.264 backwards ? That alone can bring many fairly new machines to their knees.

Also, from a quality standpoint it is highly lossy. It uses all sorts of motion estimation and other smart techniques to hide exactly how much information it tosses out the window, and so it looks decent.

As soon as you start manipulating it though you create a huge mess.

I usually convert any h.264 video I get into DVCPRO or uncompressed HD for editing. That doesn't improve the image of course, but it does allow for a smoother edit workflow and a better end product.

Codecs are a whole lot more than a color sample and data rates.

[Yi Fong Yu]

this thread is all about features we WANT, doesn't mean it's reality. we live in an age where consumer miniDV debuted at $1,000+ a few years ago and can be had now for around $300. i don't see why those full-res HD systems couldn't be had for less. economies of scale. it can be done. only engineering, time and market needs. it's 2simple, but like i said, i'm no engineer. i do have an idea what i'm asking for. $10/uncompressed 4:4:4 HD tape&$300 4:4:4 1080pHD consumer cams. how many years? i dunno 10 years? 20 years? why? because it's the future and i think that 1920x1080p will be the last great consumer format. after that, it'll be be niche markets. i'm willing to settle at that resolution.

smaller HDs aren't practical yet. 3.5 has the volume to justify the cost. the cost/MB is much lower on 3.5 HDs than 2.5.

i have a front screen projector and HDTV cable. @100" most of the time HDTV streams run under 25 and it looks fine by me. no big bad artifacts
PS do you re-read some of the stuff you write?

[Boyd Ostroff]

I've had to edit this thread due to violations of DVinfo policy. Everyone: please keep the discussion civil and avoid making personal references to your opinion of other forum members' abilities. If you aren't already familiar with DVinfo policy then please read the following. By registering at DVinfo you agree to abide by these principles. Thanks for your cooperation!

http://www.dvinfo.net/network/policy.php

[Terry Johnson]

[QUOTE=Alexander Ibrahim]MPEG-4 is a container format, like .avi Quicktime and Windows Media.


"I usually convert any h.264 video I get into DVCPRO or uncompressed HD for editing. That doesn't improve the image of course, but it does allow for a smoother edit workflow and a better end product."

This is a process that I would anticipate that would be used to edit MPEG 4 encoded video. It is my understanding that HDV will be edited in a similar fashion. And it is my assumption that with a 25 Mb per second recording bandwidth the recording quality could be very good. I would also hazard to guess that the motion artifacts associated with MPEG-4 (with a 25 Mb per second bandwidth) would be significantly less than those associated with HDV. I'm not an expert in this area so this is just speculation.

[Alexander Ibrahim]

Originally Posted by Yi Fong Yu

this thread is all about features we WANT, doesn't mean it's reality. we live in an age where consumer miniDV debuted at $1,000+ a few years ago and can be had now for around $300. i don't see why those full-res HD systems couldn't be had for less. economies of scale. it can be done. only engineering, time and market needs. it's 2simple, but like i said, i'm no engineer. i do have an idea what i'm asking for. $10/uncompressed 4:4:4 HD tape&$300 4:4:4 1080pHD consumer cams. how many years? i dunno 10 years? 20 years? why? because it's the future and i think that 1920x1080p will be the last great consumer format. after that, it'll be be niche markets. i'm willing to settle at that resolution.

smaller HDs aren't practical yet. 3.5 has the volume to justify the cost. the cost/MB is much lower on 3.5 HDs than 2.5.

i have a front screen projector and HDTV cable. @100" most of the time HDTV streams run under 25 and it looks fine by me. no big bad artifacts
PS do you re-read some of the stuff you write?

First off, though it has been removed by the "wrangler", apparently you thought I offended you. I apologize.

Second, I can appreciate a typo and it looks like you made one. You wrote 250', not 250". 250' is a big IMAX screen, and you are talking instead about a huge home theater or a very small theater. It was so off the wall (pun sadly intended) I should have asked rather than going insane.

Still even at 100" I suggest that you may not be seeing artifacts present in the image. There may be a technical reason, or you may not be used to picking them out. With a home projection set up I'd ask if the projector is in an HD resolution. I'd ask if the projector was calibrated, what screen you were projecting onto, and if light levels are appropiate in your theater.

As for seeing issues with images, I can very clearly see faults with HDV and DVCPRO HD images on both my Sony CRT's and my Apple Cinema HD 23" (Over DVI most of the time, but on a few occasions via SDI-DVI converter.)

If you have software like Lightwave or Maya I suggest creating an animation in 1920x1080 32 or 48 bit TGA or TIFF uncompressed still sequence. Then bring it into your NLE. Then output it as DV, HDV, h.264, DVCPRO 50 and DVCPRO HD. (HDCAM and HDCAM SR if you have it.) Don't forget uncomressed SD and HD. The difference is clear.

I have the results of such an experiment. This is using 960x540 resolution. Using various codecs. The uncompressed image is at 4:2:2. H.264 is included, but not HDV or any SD codecs. Even this VERY simple image shows some flaws. Look at the lens flare. Also you can see the colors are different in all three versions.

http://www.zenera.com/images/HD_codec_comparison.tiff

Remember those are CG images! That's as clean as the codecs will get! Keying a strawberry blonde's hair while she is running out at the beach gets ugly fast. Can you say rotoscope?

Now, as far as the future of HD is concerned. HDV is one of the best possible 25Mbps codecs for HD video. I think we might be able to get 1920x1080 60i or 30p into 25Mbps, but the quality will be worse than what we see now.

I think the future will be higher bitrate studio profile codecs. I doubt very much we'll see another consumer 25Mbps HD editing format. HDV is barely an editing profile- actually it may not qualify by MPEG consortium standards.

As far as 3.5" vs 2.5" drives are concerned, a 100GB drive is excellent for 25, 50 amd 100Mbps recording.

Laptop drives have all sorts of features that are not available in 3.5 drives. For one they support fast drive parking. If you ever drop or bump your camera you'll be thankful for this. It is essential for all those laptop features (Apple and IBM/Lenovo laptops) that detect that the laptop is moving too fast and park the drive heads, preventing data damage.

They are just generally more rugged, since it is assumed they will be moving, as opposed to 3.5" drives which are expected to be installed in stationary enclosures.

Also, in 2004 laptop sales exceeded desktop sales. Economies of scale have tipped now in favor of laptop hardware.

Finally, and to reenforce a point I thought I had made: You talk about DV dropping in price from $1000+ to $300. Well I think that those numbers are wrong but my memory is fuzzy, so we'll use them.

That means that DV dropped in price by ~66% in ten years.

Today a full resolution 4:4:4 HD camera (HDCAM SR) is $117,000 USD. That does NOT include a lens, batteries etc. Tapes run ~$210.00 USD for two hours.

If we see the same rate of decrease in price in ten years a camera with the same specs will run about $39,000 USD. Tapes will run about $70 for two hours.

We won't see a full resolution 4:4:4 HD camera at a price of $300 for at least 15 years, and my guess would be 30+ years, if ever! As another hint, we don't have 4:4:4 SD cameras for anywhere NEAR $300 today.

By then, we'll all want to shoot "Digital IMAX" or some such. IF videography is even a viable business for most of us involved in it today. It may well go the way of Typists and Word Processing professionals.

[Alexander Ibrahim]

[QUOTE=Terry Johnson]

Originally Posted by Alexander Ibrahim

"I usually convert any h.264 video I get into DVCPRO or uncompressed HD for editing. That doesn't improve the image of course, but it does allow for a smoother edit workflow and a better end product."

This is a process that I would anticipate that would be used to edit MPEG 4 encoded video. It is my understanding that HDV will be edited in a similar fashion. And it is my assumption that with a 25 Mb per second recording bandwidth the recording quality could be very good. I would also hazard to guess that the motion artifacts associated with MPEG-4 (with a 25 Mb per second bandwidth) would be significantly less than those associated with HDV. I'm not an expert in this area so this is just speculation.

Well your anticipations accuracy depends entirely on the MPEG-4 codec you choose. Pretty much all the MPEG-4 we have seen to date is for consumer use. So called Simple Profile.

If you have a codec meant for editing, Studio Profile (There are provisions for Simple Studio Profile and Core Studio Profile.) then you don't have to.

MPEG-4, is very strongly based on Quicktime. So for a preview of the future of MPEG 4 editing compare editing HDV (Which is MPEG-2 in a simple profile) to editing DVCPRO HD (which is I think supports most features of Simple Studio Profile MPEG-4 ) in Final Cut Pro or another Quicktime based editor.

Did that clear up or confuse the matter ? I can never tell- it makes sense to me. <shrug&grin>