Joseph Stunzi
Veteran
Wasn't this thread originally meant to discuss the HPG20. If you want to battle out P2 vs CF and other things, just start a new thread. I'm curious about new news on this device personally.
Panasonic AG-HPG20: an AVC-Intra P2 Gear
- Thread starter Barry_Green
- Start date
Jan_Crittenden
Veteran
Guys,
You misunderstood what I said, and that is that there is more going on in the P2 card than just recording data. I am not trying to insult anyones intellegence, but there really is more going on there than what even the current CF cards do.
The CF card is fine for a certain level of application and then the expection of CYA goes up.
And no, we are not making boatloads of money on the P2 card.
So as I said, unless you have been on the engineering team, you do not know all that is in the P2 card, nor do you know the quality of the memory that is in there, which btw you cannot buy commercially. You do not know about the LSI and all that it does. You can say that you do but unless you work for Panasonic, you don't. So saying that the CF card can do what the P2 card does is incorrect, it cannot.
Sorry,
Jan
You misunderstood what I said, and that is that there is more going on in the P2 card than just recording data. I am not trying to insult anyones intellegence, but there really is more going on there than what even the current CF cards do.
The CF card is fine for a certain level of application and then the expection of CYA goes up.
And no, we are not making boatloads of money on the P2 card.
So as I said, unless you have been on the engineering team, you do not know all that is in the P2 card, nor do you know the quality of the memory that is in there, which btw you cannot buy commercially. You do not know about the LSI and all that it does. You can say that you do but unless you work for Panasonic, you don't. So saying that the CF card can do what the P2 card does is incorrect, it cannot.
Sorry,
Jan
mikkowilson
Steadi-Guru
Jan,
You mentioned that the HPG20 can play back from a USB HDD in "confidence mode". What exactly is that mode?
I understand that the unit can record from the inputs to P2 cards. And can move files between P2 cards and USB HDDs. Can it also record directly to USB HDDs?
Also, is the Up/down/cross conversion available live? Or only when recording and/or playing back?
I assume that the unit does not have a Composite input (even for just DVCPRO-25/50 recording)?
- Mikko
You mentioned that the HPG20 can play back from a USB HDD in "confidence mode". What exactly is that mode?
I understand that the unit can record from the inputs to P2 cards. And can move files between P2 cards and USB HDDs. Can it also record directly to USB HDDs?
Also, is the Up/down/cross conversion available live? Or only when recording and/or playing back?
I assume that the unit does not have a Composite input (even for just DVCPRO-25/50 recording)?
- Mikko
Shipsides
DVXUser Sponsor
The ability to play P2 content directly off a hard drive is a long time coming. If confidence mode allows for playback of backed up content on a USB HDD then this device is even more impressive.
Jan,
My question is will the HPG20 backup data like it has in the past? Making a drive letter for each card copied? As you know there is a limit to the number of drive letters so this backup process has not been preferred. But with confidence mode I would recommended using this option if you wanted to make tape dubs or down converted output of P2 material. Overcoming the drive letter limitation would make this device even better. Also any idea if the P2 Mobile ever get a "confidence mode"? Post houses would love the idea of using the Mobile to play p2 content directly off a hard drive and into their decks.
Jan,
My question is will the HPG20 backup data like it has in the past? Making a drive letter for each card copied? As you know there is a limit to the number of drive letters so this backup process has not been preferred. But with confidence mode I would recommended using this option if you wanted to make tape dubs or down converted output of P2 material. Overcoming the drive letter limitation would make this device even better. Also any idea if the P2 Mobile ever get a "confidence mode"? Post houses would love the idea of using the Mobile to play p2 content directly off a hard drive and into their decks.
Barry_Green
Moderator
It is my understanding that yes, that's what it means. You can play clips straight off a hard drive. I think they call it "confidence" because they can't guarantee that it'll play back flawlessly (as that's up to the hard disk, of course). So you're spot-checking the footage for confidence that it got there.
I expect the same type of drive-letter process, but remember on the HPG10 they added USB host mode which boosted the number of potential partitions from 15 up to 23. USB mode is just better in all ways; it provides for more partitions and it also allows you to assign names to the volumes. And it has bus power. Panasonic has a downloadable program that lets you map an offloaded drive into a folder, rather than having it take up drive letters (like how the P2 Store Manager lets you map all the contents of the P2 Store into an NTFS folder, instead of each partition needing a drive letter).
I expect the same type of drive-letter process, but remember on the HPG10 they added USB host mode which boosted the number of potential partitions from 15 up to 23. USB mode is just better in all ways; it provides for more partitions and it also allows you to assign names to the volumes. And it has bus power. Panasonic has a downloadable program that lets you map an offloaded drive into a folder, rather than having it take up drive letters (like how the P2 Store Manager lets you map all the contents of the P2 Store into an NTFS folder, instead of each partition needing a drive letter).
Jan_Crittenden
Veteran
Confidence PB is that you can review the PB but isn't full 60oi/60P mode, plays some of the frames just not all. So the PB is not fully smooth but you can check the content. 24P looks more like full playback.
No, the video must record to the P2 Cards, first.
All the up/down/cross, must be from PB.
That is true.
Thanks,
Jan
mikkowilson
Steadi-Guru
As a follow-up to this...
Can the HPG20 copy from one card to a USB HDD while recording to another card?
This in the hopes that you could use the cards as a "buffer" and then dump on the fly. - The same of course could be achieved using two units (one as a "recorder") and the other as a "dumper" and swapping the cards between them.
Of course the ability to keep up would depend on the speed of the hard drive and the bitrate over time of the video being recorded. Not sure if this could be guaranteed/possible.
- Mikko
Jan,
do you have anything that compares the AG-HPG20 to the AG-HPG10 features and price?
Terry
do you have anything that compares the AG-HPG20 to the AG-HPG10 features and price?
Terry
Mike Schell
Active member
Hi Jan-
This is Mike Schell, from Convergent Design. I have a few comments on your recent post.
First-off, I am not an employee of Panasonic, but I am an Electrical Engineer with over 25 years of digital design experience, so I do know a little about Flash memory, especially since I was formally a EEPROM and Flash memory chip designer. I can tell you we're not talking rocket science here, as P2, SDHC and CF cards are all based on multiple NAND FLASH chips and a (relatively) simple controller chip.
Regarding the use of CF cards vs P2, I seriously doubt there is a significant difference in reliability and performance. CF cards do have built-in CRC error checking, so every write and read transaction is checked. I don't doubt that the SLC (single level cell) NAND Flash chips are slightly more reliable than the MLC (Muli-level cell), but either of these technologies are available in CF cards today. Our own internal testing has shown that both MLC and SLC CF cards are extremely reliable. To date, I am unaware of any recording problems whatsoever related to CF cards. I bet the Red camera folks, as well as 1000's of professional DSLR users will also second this observation.
Regarding performance, today's CF cards are screamers. The lowly 133X 32GB MLC Transcend card (US $82) has a write speed of over 160 Mbps and a read speed of over 320 Mbps. The 300X 16GB SLC Transcend card (US $150) is rated at 360 Mbps write and read speeds. These speeds are more than sufficient for compressed HD video from either of our CODECs.
Yes, 5 or 6 years ago, this was an entirely different story, as CF cards were expensive and slow. But the technology has advanced by leaps and bounds. Today's CF cards are very fast, very affordable and very reliable.
Regarding the Long-GOP MPEG2 CODEC used in the Flash XDR / nanoFlash. Yes, the MPEG2 algorithm is over 15 years old. But the silicon implementation has also advanced considerable in the last 15 years! We're now using a 7th generation CODEC, which produces amazing results at 100 Mbps (check out some of the image comparisons in the EX1 area of this forum).
If you try to run Long-GOP at HDV rates, such as 25 Mbps, then yes the video quality does suffer. But we're taking about 100 Mbps 4:2:2 Full-raster (1920x1080) quality, that plays back in Final Cut Pro (today) without transcode or re-wrap...straight off the CF card!
Yes, AVC-I does offer 10-bit, while we are limited to 8-bit, but I bet the differences are very minimal once the video has been compressed. Any compression will tend to reduce the overall resolution of the data. But, MPEG2 Long-GOP does have the distinct advantage of both temporal (I-Frame) and spatial (P,B Frames) compression, which improves the overall efficiency some 2X to 3X over an I-Frame only CODEC.
The AVC-I CODEC should provide superior results over MPEG2 CODEC running in I-Frame only mode. Many people have touted a 2X improvement, but this depends on the bit-rate. All the studies I have read (I also have an extensive background in CODEC technology) has shown that this 2X advantage disappears as the bit rate increases. At 20Mbps, the 2X improvement certainly applies, at 100Mbps, the improvement drops to 30% or less.
Naturally we are comparing AVC-I to MPEG-I, not MPEG2 Long-LOP, which enjoys at least a 2X advantage in compression efficiency over an I-Frame only CODEC.
My point here is that technology moves at a very rapid rate. Five or six years is a lifetime in the electronics industry. So, let's not assume that products or technologies initially developed years ago have not made significant advances in speed, reliability and cost.
Mike Schell
Convergent Design
Jan_Crittenden
Veteran
Hi Mike, Thanks for posting, it gives me an opportunity to give a little more analysis to the situation.
It isn’t just the memory Mike, it is what else that is in there that matters. Flash memory by itself may have ECC, and yes I have a number of DSLRs that have worked with CF and I have no complaints. What the CF doesn’t have is the parity check and the ability to shift its writing speed to whatever we may choose down the pike and will be able to work with anything that we have produced in the past. So 4 years ago that 4 GB card could work with any thing we may come up with even if it were a 4:4:4 AVC-Intra codec that needed 400 mbs to write it. Of course there wouldn’t be a lot of time but there would be compatibility. Additionally the memory is held within a diecast chassis which of course also gives it the ruggedness factor that it can handle just about anything that gets thrown its way. (For those that have seen my demo, you can smile)
Well I have yet to see a CF card built into a diecast nor have I seen a CF card that has the ability to write at 800mbs like our newest 64GB cards do. People on the P2 bandwagon are paying for reliability, compatibility and a system that does not have a dead end.
But Mike it doesn’t do the rest of the secret sauce. How is it parsing out the video? What MXF structure is it written in? Is the audio separate for the video? How rich is the metadata, this is even more important in a data domain? Can I edit directly on the card? Write back to the card from the timeline in a universally coded file structure that can be read by any other P2 device or NLE? There are a number of housekeeping issues that are handled by the card with the LSI that is built in the card. It is more than memory.
I couldn’t agree more but as I said in the beginning of this, the CF card does not do everything that the P2 card does. And because of that Panasonic would not use a CF card. Doesn't have the right stuff.
I did see those and it is an improvement over the EX1 codec, but frankly you are still betting the bank on MPEG2 and here is where you and I will just have to agree to disagree. The future is not wrapped around MPEG2, it is moving toward H.264.
I would take that bet actually. And I am not a betting person unless I know the answer. Keep in mind that the efficiency of the codec is not the priority of the Production person that has to get his images through the rest of the production stream. If you have not compared 10 bit AVC-Intra to an 8 bit MPEG2 codec at similar bit rates then you are just making an assumption that is easily proven wrong. The difference between AVC-Intra and any 8 bit codec is frankly night and day when it comes to render tonality, noise and speed of handling each frame on the Desktop.
Please site your source and not that AVC-Intra (as implemented by Panasonic) is specifically optimized for high data rates. Tests done with reference codecs will not give an accurate assessment. Here are a couple of sources that discuss the beaty of the codec, one is the HPA Award for AVC-Intra and the other is on Compression Technologies.
http://broadcastengineering.com/pro...st-alliance-engineering-award-1116/index.html
http://broadcastengineering.com/infrastructure/compression-technology-0601/index1.html
And my point remains that the CF card does not do what the P2 card does. It does some of it just not all of it. Speed isn’t the only argument here. Our customers are looking for reassurances, reliability, future-based decision-making in product design, and believes in a true file-based workflow. Keeping data on cheap cards sort of feels like a tape based workflow, the data is not as accessible and it means that you have to keep a device that will read those cards in perpetuity. If that is what you are looking for; inexpensive media, long GOP recording, efficient codecs, I suggest you look at AVCHD.
All the best,
Jan
Jan_Crittenden
Veteran
Hi Terry,
Pricing has not yet been released but it will be over the $5000 price point. The strongest features in contrast to the HPG10 is that the HPG20 has an HDSDI input, AVC-Intra capability, and the ability to upload Metadata, just like you would on a camcorder with data that you have composed in P2-Log or P2CMS.
It will only stream the 1394 connection, no hosting to 1394 drives, only hosting to USB, 23 partitions, and like the HPG10 in USB mode, it can verifiy data transfer.
You will see a marketing bulltin in late December on this little guy. It is indeed a pretty nice product.
Best,
jan
Willis Chung
DVXuser Sponsor
Is there any way for P2Gear or P2Mobile to record the HPX3700's 4:4:4 color spacing via their avc-intra 100's codec?
Jan_Crittenden
Veteran
Hi,
No, you have to go to a 4:4:4 recorder.
Best,
Jan
No, you have to go to a 4:4:4 recorder.
Best,
Jan
Ooh, I love a good fight.
Actually it would be great if you guys can give each other one of your recorders, then do a comparison test. We'll probably get different results, but I think we'll also be able to see the best of each recorder. Any takers?
Mike Schell
Active member
Hi Jan-
It's true that CF cards do not have parity checking, but neither does the incoming HD-SDI video, which is even more error prone. HD-SDI relies on the same CRC check employed in both CF and P2 cards. Also, while CF cards do not have a die-cast chassis, neither does the tape cartridges we have depended on for 30+ years. I am sure this helps, but CF cards have a very reliable track record across an installed base which is certainly over 1000X larger than P2 cards.
CF cards now have plenty of read/write performance for professional HD video. You can now readily purchase CF cards with 360Mbps read/write performance. We're no where close to this performance requirement with either of our CODECs.
Compact Flash is no where near a dead end as we have seen new 64 and 100GB cards introduced from several manufacturers this year. The current CF spec supports a data-rate up to 1000Mbps, so we have plenty of growth potential. CF has a very large following with DSLR cameras and of course there are 1000's of Red users that rely on CF media every day.
Also, we implemented the "secret sauce" capabilities inside our box. We have a powerful FPGA to manage the CF media and to keep track of the video, audio, metadata and file structure data. Yes, you do need a hardware based controller to manage these functions, but why not implement these functions (once) in the recorder (or camera) instead of adding cost to the media?
Yes we can play/edit directly off the CF media without transcode or re-wrap, using a $60 Firewire-800 CF reader. Last time I checked you could edit AVC-I directly off the P2 card in Edius, but you are forced to trancode to ProRes when using FCP. That's because the processor load for AVC-I decode is substantially greater that MPEG2, which plays back fine on 3-4 year old MACs.
I know you can play / edit DVCProHD files directly off the P2 card, but this CODEC is not competitive (in quality) to our 100Mbps MPEG2 4:2:2 full-raster CODEC.
Let's just do some side by side comparisons. We can test AVC-I against our Long-GOP and I-Frame only CODEC and overlay the images in Photoshop. I am very curious to compare 10-bit and 8-bit compressed images, particulaly in shallow gradient scenes. I would like to see the differences in an I-Frame only CODEC and a Long-GOP CODEC running at the same bit-rate.
And we can also compare and contrast the workflows.
Mike Schell
Convergent Design
Jan_Crittenden
Veteran
Hi Mike,
But the HDSDI is not where the parity checking needs to happen it needs to happen on the P2 card.
But we have very large corporations/broadcasters that make decisions based on the product, the product line-up, its future support and reliabilty form factor. It isn't just about the memory writing speed factor, it is virtually everything else. I will say that when we started this back in 2004 we did have to work to get the writing speed to where we needed it, and then added overhead for future based concepts. So we built a system that includes cameras, workflow tools and have worked with a group of companies that have made P2 one of the best supported in the industry.
Because the card can do it where it counts. That way no one is sitting there with a card that doesn't work because it is in the card. Sort of like my experience this last weekend with my new D700. Wasn't paying close attention to which Flash Cards worked and which didn't. Used an older one and it didn't work. If it is P2 it works.
Having something work on 3-4 year old computers should not be the defining line of leading technology and codecs. The AVC-Intra codec and its full implementation is just beginning. An MPEG-2 codec should be easy as that is an older codec. Keep in mind I have watched a good number of codecs implemented and deployed in my years at Panasonic, most of them by the competition, it is a matter of time and popularity to enjoy full support. Frankly I see D5HD quality out of AVC-Intra. I see a 10 bit format and its subtleties, frankly, there is more stuff there than what an 8 bit codec can do.
And here I would have to look at this, as it is not always about just resolution but many other factors as well.
Well the way we would need to do this is to feed both devices a signal from the HPX3000, which frankly is the most gorgeous picture I have seen from a camera in years and then butterfly the images, an overlay in photoshop does one frame at a time and frankly we are evaluating video. I can say that when you look at gradients in a singular color, the results are amazing and we should take it multigenerational on the time line working through color correction and some transitions. Then let’s look at the skin tones. And while we are at it we should make sure that we have some water falls and maybe a 10 bit gradient chart for starters.
And yes the total workflow, on all platforms. The tough thing about platforms is that they change all the time so what is true on that day will be different three weeks later. Since there is only one HPG20 at this point and it is an Engineering sample I don't see that a test is really appropriate at this time but late February, maybe or after NAB for sure. In NJ.
Best regards,
Jan
Mike Schell
Active member
Hi Jan-
OK, let's just agree to disagree on the issue of P2 vs CF. Obviously there are some advantages that P2 enjoys over CF, such as parity and a die-cast chassis. Beyond that, I can't identify a clear advantage, especially given our hardware based CF memory manager in our Flash XDR /nanoFlash.
I find any discussion about reliability differences to be very minor. The Panasonic AVCCAM web page touts "Ultra Reliability" for a camera that uses SDHC memory. SDHC and CF are basically the same memory card (they use the same NAND Flash chips, only a different controller and form factor). So, I would have to believe that this "Ultra Reliability" descriptor would also apply to Compact Flash?
So, I would agree that P2 has a slight advantage over CF, but at 20X the price, customers should expect some additional benefits.
I'm ready to test the AVC-I va our MPEG-2 Long-GOP. Could we also test with some moderately priced cameras (like the Sony EX3)? I don't think the majority of people on this forum are using a $48K camera for their work.
Let's compare still images, moving video and of course workflow. To be complete, let's also include equipment costs and ease of use considerations.
Mike Schell
Convergent Design
OK, let's just agree to disagree on the issue of P2 vs CF. Obviously there are some advantages that P2 enjoys over CF, such as parity and a die-cast chassis. Beyond that, I can't identify a clear advantage, especially given our hardware based CF memory manager in our Flash XDR /nanoFlash.
I find any discussion about reliability differences to be very minor. The Panasonic AVCCAM web page touts "Ultra Reliability" for a camera that uses SDHC memory. SDHC and CF are basically the same memory card (they use the same NAND Flash chips, only a different controller and form factor). So, I would have to believe that this "Ultra Reliability" descriptor would also apply to Compact Flash?
So, I would agree that P2 has a slight advantage over CF, but at 20X the price, customers should expect some additional benefits.
I'm ready to test the AVC-I va our MPEG-2 Long-GOP. Could we also test with some moderately priced cameras (like the Sony EX3)? I don't think the majority of people on this forum are using a $48K camera for their work.
Let's compare still images, moving video and of course workflow. To be complete, let's also include equipment costs and ease of use considerations.
Mike Schell
Convergent Design
Jan_Crittenden
Veteran
I think that is a safe bet, as we will not agree. As I see it there is a very clear advantage to the P2 card.
Actually I don't know about that statement being universally applicable, I have seen the tests that they perform on the Panasonic cards, with twisting and flexing. And I think the statement is about the Panasonic cards. As an example I read a post earlier today here where another brand's card went bad. Card memory is not infallible.
We are going to be testing the codecs. I don't really care about whether Sony's camera works well or not. I do know that the 3000 puts out a picture that will show the differences between the two codecs and the noise, and imperfections of the camera's output won't get in the way of the codecs' performance.
I think the ease of use needs to be a sliding scale as it will all change at NAB as it usually does. So perhaps we should wait until after NAB. Most video is not a still image, there is generally something moving. But I will conceed to a couple of charts including the standard resolution charts and the 10 Bit gradient charts.
As far as costs are concerned, hey I think that seems pretty obvious, your device is cheaper and has CF cards. But it would really be hard to make a final tabulation until we expended the life of the cards, and on the P2 it will take 100,000 rewrites. And since the P2 Portable actually does more than the Convergent design product, like play video to its own monitor, upload metadata, add metadata, allow the user to judge video and audio quality, add markers and host its own transfer with verification to a Host Drive, is this really really possible? I mean really, you want to compare the product pricing on these items with the starting assumption that they are equal in performance and capabilities? Please let's not get silly here.
Best,
Jan