SanDisk intros 8GB, 16GB ExpressCards for camcorders (xdcam EX)

[Nik Manning]

Somehow I thought these cards where going to be cheaper than P2 but I guess not. The SanDisk version is $500 for a 8gb and $900 for 16gb. What! Come on now you got to be kidding me. They took an open format and mucked it up in that special Sony way.
Link

Now look at the Lexar express card pricing. Link
Sony stop this foolish game. Every format cannot be your format. Lets try to limit things to 1 proprietary format every 5 years. Can we do that please!:violin:

 

[Jason Ramsey]

Yeah. Right now the cost of SxS cards are comparable to P2. 16gig p2 retails for the same price as the 16gig in that link you provided.

So, the cost of the workflow is the same basically. It comes down to the cost difference in the cameras and whether the features of one are worth the price over the other. P2 support is pretty widespread now and improving and proven and well established in the broadcast industry as well. SxS will be brand new and it will take time for NLE's to support it and for the workflow to be put through its paces. I'm not saying that's a 'negative' towards the camera. Just the nature of the beast. It's new.

Late,r
Jason

 

[Nik Manning]

I was really thinking that the Sony would be the camera to buy very cheap tapeless workflow with great low light and better dof. This crap that they pulled with these cards however is just heart breaking. On of the benefits of mini dv was that you could just go to any store and buy some more tapes. Can't just buy an express card for these ex cams cause they won't work. Obviously it isn't a huge deal but man it sucks. Plus I just saw this. LINK LEXAR 8gb for $99

 

[Barry_Green]

I was really thinking that the Sony would be the camera to buy very cheap tapeless workflow with great low light and better dof.

Why?

Okay, let's put it this way: when has Sony ever, ever, ever, ever, ever been the low-cost option? Or the open standard? It's simply not something that company does, or will ever do. They are proprietary by nature -- their slogans mean what they say. "Sony: Like No Other." and "Sony: The One And Only."

I am not surprised in the slightest. It's exactly what I expected -- if they want to make a system that has the performance of a P2 card, they're going to have to use similar manufacturing techniques, and the result is that it's going to cost as much as a P2 card does. And now we have confirmation from both Sony and Sandisk that yes, their cards cost just as much. There's nothing new here, this is par for the course.

 

[baquajim]

Yeah. Right now the cost of SxS cards are comparable to P2. 16gig p2 retails for the same price as the 16gig in that link you provided.

So, the cost of the workflow is the same basically.

I think what is not being considered is that with a Sony 16GB card, you can store 50 minutes of 1080P but with a 16GB P2 card, you can store 16 minutes.

To get the equivalent recording time in P2, you would have to spend $2700 (3-16gb P2) versus $900 (1-16gb SxS)

 

[marketmd]

Somehow I thought these cards where going to be cheaper than P2 but I guess not. The SanDisk version is $500 for a 8gb and $900 for 16gb. What!

Why should SanDisk make less of a profit than Panasonic? The whole area is absurdly overpriced, like 35mm adapter market. Wait a year and both will be brought down to earth (but who can wait?)

 

[marketmd]

I am hearing from my DIY buddy that these new cards will be easier to hack/create your own than the P2 which was impossible. In theory that is. When he gets his hands on them I will let you all know more...

 

[MiniMan13]

Well at NAB they did say that there card would be affordable...

 

[harddrive]

Yeah. Right now the cost of SxS cards are comparable to P2. 16gig p2 retails for the same price as the 16gig in that link you provided.

So, the cost of the workflow is the same basically.

No - because even in the worst possible case (the SxS cards do sell for the price quoted in the link, which is the recommended price) the data rate is only about 1/3 that of DVCProHD. Hence the *cost per minute* of SxS easily trumps that of P2, even if the costs per GB were similar. Likewise, the run time of a camera with 2 16GB cards is about 3 x as long for SxS as for P2.

Anybody seen any ACTUAL dealer prices for SxS cards yet, as opposed to recommended prices?

 

[Nik Manning]

I am comparing cost versus other Express Cards. If the Lexar 8gb is $100 I think overpriced might be if you charge $400 more than competing products. That is not affordable! They should remove that from any and all of there marketing material. Please somebody release a open Express card camera that records to the 50mb hdv spec and crush Sony and there stuck up ways.

 

[Jason Ramsey]

Run times will be longer per gig with the SxS card, but I would stil say that the prices of the cards will be about the same. p2 doesn't sell for its recommended price either. They've been had for as cheap as 850 from authorized dealers the day they were realeased.

The price per gig will be cheaper, but you still have to buy the whole card, so I don't see *much* difference in the cost of the workflow. Less offloads maybe, which is good, but all of that will go away for both of these cams relatively soon.

Jason

 

[Barry_Green]

Hence the *cost per minute* of SxS easily trumps that of P2

Yes, and the cost of an Aiptek HD camcorder ($169) easily trumps that of the Sony EX1. They're both solid-state-recording camcorders, right?

Let's at least try to maintain a semblance of apples-to-apples here. One is 35 megabit long-GOP 4:2:0, the other is 100 megabit intraframe 4:2:2.

I mean, if you are willing to discard format differences and just compare minutes of recording time, then I dare say you could put a P2 camcorder in DV mode and get longer recording times on P2 than on SxS, right? Fair's fair, yes?

 

[harddrive]

I am comparing cost versus other Express Cards. If the Lexar 8gb is $100 I think overpriced might be if you charge $400 more than competing products.

But what speed is the Lexar? To compete with the SxS card it would have to be 800Mbs (100MBs) and I strongly doubt it is that.

Frankly, it's very early days and we haven't even got a street price for SxS yet - only manufacturers recommended prices. There are also conflicting stories about whether or not it will work with non-SxS ExpressCards - one report saying "no", the other that it WILL, albeit with limitations. Maybe those two tests reached such different conclusions because the people concerned used different spec cards? (Albeit both being sub SxS?)

For most people I suspect the 800Mbs speed is unnecessarily high. But maybe the *current* range of non-SxS ExpressCards just isn't high enough. That's not to say we won't see middle of the road performance cards - equivalent to the Extreme III range, say - at around 20MBs (160Mbs). That would allow 5-6x downloading of XDCAM-HD, yet shouldn't be any more expensive than current CF cards.

Far too early to draw hard conclusions yet.

 

[harddrive]

Let's at least try to maintain a semblance of apples-to-apples here. One is 35 megabit long-GOP 4:2:0, the other is 100 megabit intraframe 4:2:2.

I mean, if you are willing to discard format differences and just compare minutes of recording time, ........... Fair's fair, yes?

If fair really is fair, then maybe it's worth mentioning that one also has a horizontal luminance recording resolution (in 1080 mode) of 1920 versus the other of 1280, chrominance figures of 960 v 640 as well?

Resolutions aside, whilst DV v HD is obviously apples and oranges, I would maintain DVCProHD v XDCAM HD IS apples v apples. Albeit different varieties.

The interframe/intraframe debate is a bit of a boring one, but I thought it had been long settled that interframe can give comparable quality to intraframe at lower bitrates - the penalty is added complexity, rather than (necessarily) quality.

 

[Barry_Green]

but I thought it had been long settled that interframe can give comparable quality to intraframe at lower bitrates

Long settled? By who?

Interframe works fine up until the point that it doesn't work fine, and can in fact be ruinous to your shots. Intraframe always works fine.

There's a reason that every professional format ever devised has used intraframe, even when using a codec that could be used for interframe. MPEG-IMX, for example, a professional implementation of MPEG-2 using intraframe-only. HDCAM-SR, for another example; it's MPEG-4 but intraframe-only. Even the new Convergent Designs flash recorder box uses intraframe-only MPEG-2, at 160 mbps.

Inter-frame was developed specifically for low-bandwidth requirements, such as transmission over limited bandwidth airwaves, or implementation in tiny-bandwidth DVDs. It most certainly was not developed for quality purposes! The only reason interframe codecs came about in camcorders is because the recording media has such limited bandwidth. The DV tape was maxxed out at 25 megabits, so HDV was limited to 25 megabits. The XDCAM disc had a maximum write speed of 35 megabits, so XDCAM-HD is 35 megabits. The data rate certainly wasn't selected because it's high-quality! Sony keeps pushing up their write speed, and correspondingly they keep increasing the bitrate, trying to get it to the point where it *is* competitive with intraframe. I think at 50mbps they might actually get there, but I'd have to test it to be sure. I would still rather have the 160mbps Convergent Designs implementation of intraframe 4:2:2 than a long-GOP 50-megabit version though.

 

[harddrive]

Interframe works fine up until the point that it doesn't work fine, and can in fact be ruinous to your shots. Intraframe always works fine.

Not necessarily, and it's hard to generalise, but intraframe can be challenged by excessive fine detail, interframe is more likely to have problems with excessive motion. Normally both manage fine, and I think far too much is made of the quality debate. Within reason, the downside to interframe is complexity (more power etc needed to edit) - the upside is greater efficiency per bit.

There's a reason that every professional format ever devised has used intraframe, even when using a codec that could be used for interframe. MPEG-IMX, for example, ..................

Not true - Beta SX? I'd also say that in the past the technology of the day just couldn't viably cope with the processing demands that inter-frame would make on it. It was simply cheaper to use an intra-frame system, and by and large storage capacity wasn't an issue with tape. Tapeless media and far higher, cheaper processing capability have turned the economics upside down.

Inter-frame was developed specifically for low-bandwidth requirements, such as transmission over limited bandwidth airwaves, or implementation in tiny-bandwidth DVDs. It most certainly was not developed for quality purposes!

I don't think in this case the limited bandwidth/quality arguments are exclusive. By and large no one complains about DVD quality, so here are we not looking at maintaining quality, whilst reducing bandwidth? Can it not also be said (with reference to DVDs) that it enabled far, far better than VHS quality at a sensible cost? So it WAS developed for quality?

The only reason interframe codecs came about in camcorders is because the recording media has such limited bandwidth. The DV tape was maxxed out at 25 megabits, so HDV was limited to 25 megabits.

Is storage capacity not more of an issue than bandwidth? (And in the case of DV/HDV the cost desirability of using off the shelf, cheap mechanisms?)

I would still rather have the 160mbps Convergent Designs implementation of intraframe 4:2:2 than a long-GOP 50-megabit version though.

I've just looked at their website again ( http://www.convergent-design.com/downloads/Flash XDR.pdf ) and note (last page) they now make reference to a 100Mbs mode - 422P@HL, long GOP, 4:2:2 1920x1080 4:2:2 mode. That may even beat the 160Mbs mode for quality, whilst giving longer recording times. At the expense of extra processing, of course.

But all this aside, I don't think the Sony system is perfect by any means. As far as media goes, I'd be the first person to agree that the Convergent Designs approach of using Compact Flash is probably far, far more sensible at the price range of such as the HVX200 and the EX, and Convergent Design have PROVED that the performance of Compact Flash is more than adequate. 160Mbs recording - 60% greater than DVCProHD!?! Capable of being downloaded at 2x real speed (other hardware permitting)!?!

I suspect that SxS has really been developed less for the EX than for higher end, yet to be announced, product, and it's inclusion in the EX is more in line with all their products using the same media.

What will be really interesting IMO is what Canon and JVC do next. A JVC "300 series" camera using Compact Flash could really raise the bar far further still at this price range. It's not long since I heard it said that Compact Flash would never be able to handle high grade pro video - the device you refer to shows that it obviously can, and at 160Mbs, even!

 

[themagickite]

about sony never ever using a standard format, that's not really true
they made cassette recorders, and heaps of miniDV cameras and such
hopefully we can expect to see a standardised solid state replacement to miniDV in the next couple of years. proprietary formats just slow down development, when competing companies get together and work something out then we get major advances (like MIDI) that benefit users and make a lot more money for the companies in question.

 

[baquajim]

The pictures always prove themselves, this number measuring and comparing is meaningless.

If the HVX-200 could record 800Mb/s but the lens and sensor aren't capable of generating more than 540 lines resolution, then it is irrelevant.

Besides, Red claims they are getting 4k into a visually lossless codec at 27Mb/s. Why can't Sony provide more with less?

 

[Barry_Green]

The pictures always prove themselves, this number measuring and comparing is meaningless.

I agree with you that the pictures are far more important than the numbers. Far, far more important.

To say it's meaningless, however, only applies to those who don't know what the numbers mean. Comparing MPEG-2 1920x1080 @ 35 megabits, vs. MPEG-2 1440x1080 @ 35 megabits, is not meaningless to those who know how to compare them, and know what the implications are.

There are those who say that HDV has no "motion artifacts" or they say "I've never seen 'em". Then again, there are those who have played Russian Roulette and have not been shot. Does that mean there's not a bullet in the gun?

I prefer to know what the systems are capable of, and what they are not capable of. If a system works fine 90% of the time, but there are situations where it fails, I want to know about it. To me that's not "meaningless", it's prudence.

Why can't Sony provide more with less?

They could, if they've drastically increased the efficiency of their encoder. That's something we'd all be very interested in knowing about, if it's happened.

 

[Barry_Green]

Not necessarily, and it's hard to generalise, but intraframe can be challenged by excessive fine detail

No more so than interframe would be.

interframe is more likely to have problems with excessive motion.

Motion is just one of the many things that can cause issues. Smoke, fire, fog, lightning, flashbulbs, strobelights, confetti, rippling water, splashing water, all sorts of things that cause changes in the frame beyond the codec's ability to cope.

Normally both manage fine, and I think far too much is made of the quality debate.

Both certainly do manage fine most of the time. I'm only talking about the times when they don't manage fine -- that's where the worry is.

In business, what's important is not how quickly you get the good news, it's how quickly you get the bad news. Knowing what the downside is, and knowing how best to mitigate it, is far more valuable.

Not true - Beta SX?

Beta SX was a dismal failure, barely a blip on the radar. Certainly nowhere near the impact, within its respective price bracket, of HDCAM or HDCAM-SR or Digital Betacam or even MPEG-IMX or DVCAM.

I'd also say that in the past the technology of the day just couldn't viably cope with the processing demands that inter-frame would make on it.

Undoubtedly some truth to that -- but that doesn't change that even when one could cope with it, intraframe is still a superior reliable way to encode.

I'll tell you what -- I'll concede interframe has a role where it is appropriate, and that's when it's compared to intraframe at equivalent bandwidth. If I had a choice of 160-megabit MPEG-2 Interframe, or 160-megabit MPEG-2 Intraframe, I'd probably choose interframe. But when the choice is 25 or 35 megabits, vs. 100 to 440 megabits, intraframe is a much more reliable system.

And really, what's with these tiny bandwidths on the cards anyway? Why do they feel they have to squeeze HD video into just 35 megabits? They've upped the bar repeatedly, acknowledging that 25 megabits wasn't enough so they made 35, and now acknowledging that 35 isn't really enough so they're going to 50 megabits for the 2/3" cameras. But these bandwidth restrictions are due solely to the tape or optical disc they're using. That card is capable of 800 megabits! Why not just open up the throttle and let us record 100 megabits?

By and large no one complains about DVD quality, so here are we not looking at maintaining quality, whilst reducing bandwidth? Can it not also be said (with reference to DVDs) that it enabled far, far better than VHS quality at a sensible cost? So it WAS developed for quality?

Yes, but now you've stumbled exactly into the quagmire that is MPEG-2 interframe recording. MPEG-2 works quite well as a DELIVERY medium. MPEG-2 is not a synchronous codec with equivalent encode/decode times. DV and its variants are. It takes equal amount of time to encode as it does to decode. So realtime encoding at best quality is easily achievable.

But MPEG-2 is not that way. MPEG-2 was designed as a DELIVERY codec, not for acquisition! It can be very extensively processed, running multiple passes to optimize it to extraordinarily high quality within a given bitrate. But that means the encoding time might be 2x, or 10x, or 50x, or 100x as long as the decode time. Optimizing and improving and making the very most of the available bandwidth.

Can't happen in a camcorder. In a camcorder you're getting realtime encoding at nowhere near the efficiency of what's possible in a standalone encoding station. By very design you're getting the least-efficient MPEG encoding possible, being done not on a full program stream but on each individual group, and the differences from group to group can be extreme.

It's a great idea for delivery. But it's a lousy idea for source footage.

Is storage capacity not more of an issue than bandwidth? (And in the case of DV/HDV the cost desirability of using off the shelf, cheap mechanisms?)

Why would it be? Storage is cheap. Quality is the determining factor. With adequate bandwidth quality can be assured. But in the case of HDV, as you say, the determining factor was re-using the DV tape drive, which dictated the maximum bandwidth available to them.

They most certainly didn't design it from a quality standpoint and say "folks, what's the minimum acceptable bandwidth we need to reliably encode a great HD signal?" Instead, they designed it from the opposite spectrum: "Engineers, you've got 25 megabits, and that's all you get. Stuff some HD into that bandwidth." The compromises made were not made in the quest of better quality, they were made around the idea that they wanted to use 25 megabit storage (or, in the case of XDCAM discs, 35 megabit).

I've just looked at their website again ( http://www.convergent-design.com/downloads/Flash%20XDR.pdf ) and note (last page) they now make reference to a 100Mbs mode - 422P@HL, long GOP, 4:2:2 1920x1080 4:2:2 mode. That may even beat the 160Mbs mode for quality, whilst giving longer recording times. At the expense of extra processing, of course.

100 megabits of interframe is about the best I've seen. I'd probably go so far as to say that's good enough, assuming the encoder is appropriately efficient.

I suspect that SxS has really been developed less for the EX than for higher end, yet to be announced, product, and it's inclusion in the EX is more in line with all their products using the same media.

But what though? See, that's what's so very curious about this whole SxS thing -- Sony insists on touting how great it is that they use long-GOP compression... so what do they need an 800-megabit storage card for? Are they going to abandon long-GOP and start encoding HDCAM-SR on those cards (which would be fine with me)? If so, how do they reconcile their current mantra of "long-GOP is better than intraframe"? I dunno, it seems to me that they could easily have stored the data stream on a cheap nonproprietary media... Red One is storing 27 megaBytes on an off-the-shelf flash card, so why did Sony need a 100-megabyte/sec storage medium to store their 6-megabyte/sec footage?

What will be really interesting IMO is what Canon and JVC do next.

Canon's doing AVC-HD. Since they haven't gone further than HDV before, I don't really expect them to go beyond AVC-HD. Canon doesn't innovate formats, they use existing formats. I look to them to develop 24-megabit AVC-HD (which should be toe-to-toe competitive with 35-megabit XDCAM-EX).

JVC, though -- that's an odd one. JVC is rabidly innovative,(or has historically been so; it remains to be seen if the sale to Kenwood will have any effect on them). They've already introduced their own 1920x1080 MPEG-2 recording format, but I think it's to hard disk only.

It's not long since I heard it said that Compact Flash would never be able to handle high grade pro video - the device you refer to shows that it obviously can, and at 160Mbs, even!

And the Red One is using it today, too.