Two words... Atomos Ninja
Two words... Atomos Ninja
Funny enough, I actually checked that out yesterday, but problem is, it seems it cant do 4K capturing and the recordings are still lossy.
Well...I think it's got a big brother coming out sometime after called the Atmos Samuri which will do all of that for several hundred dollars more.
No one can read the Gospels without feeling the actual presence of Jesus. His personality pulsates in every word. No myth is filled with such life...Albert Einstein
You know we could in theory create the very same thing for a fraction of the price.
Also, I personally would want to keep all information i.e. RAW, thus I would want to have one record and capture footage into a CineDNG file format, it will be incredibly useful when filming at 4KIf this was possible, then I would be very happy until I actually get to use 35mm film
Oooo...neato! I wonder if the Ninja's Atom-OS is the same thing as Atomic-OS.Originally Posted by HD-tography
Good to see people have had some success with this. I'm still not in the market yet - I'm not sure I will be soon, the storage requirements (both for capture and long-term) are just killing me, and I think a GH2 (my soon-to-be) with AVCHD gets me close enough for most outdoor shots.
Anyway, check my post in the DVXuser 3D thread for cheap and lightweight 12V lithium-ion batteries. They even include their own charger & power switch, all in a neat self-contained case. A few of those may be enough for a low-power system.
Ex-Firmware HackDictator (check out my own firmware <cough> 'hack' - scamtastic).
Cam: HV20 PAL | DIY: glLight(LED light) The Ski-balizer (steadicam) glKey (reflective chromakey)
Hey guys, I'm revisiting this thread and posting this here since I'm loyal to this forum.
I'm seriously considering making a portable prototype to capture the HDMI output from our cameras. In order to keep cost low (no SSD's) and lightweight, it would use normal hard drive(s) and capture in DNxHD, Cineform, Matrox or RAW. Main problem I'm running into is cross platform compatibility. In order to keep cost low, the OS would have to be Linux or Android based to avoid licensing fees. Same goes for the codec itself, as well as being able to work on Windows, Mac and Linux. I'm not going to go into specifics as to equipment, but this would be in a package roughly 10x10x3 (hopefully smaller). A small screen in the neighborhood of 4 inches would be used for controlling the system and possibly to monitor footage.
Now for the million dollar question...if I can do this and sell them at around the $600-$700 (USD) price point (maybe $500 if I can get parts cheap enough), would there be enough interest to justify mass marketing? In other words, would they sell?
I think that you have some serious competition with the BlackMagic Hyperdeck Shuttle 2 at only $330 USD, as this records both HDMI and HD-SDI, in both DNxHD and ProRes.
OK, it isn't RAW, but for almost half the price and the fact that nearly everyone shoots ProRes these days (even Hugo), you might find yourself unable to sell this.
Of course, there might well be a market for the RAW element of your design. But you will have to make the powering options bullet proof as more often than not, this is where the issues lay!
Cineform is a solid codec by the way. I also love Matrox's VFW codecs which they thankfully made open for free in recent years. They used to cost a fortune!!
In the UK? Then post here. You might need a hand in the future from a real person... HV20 (PAL)
I gave up on the idea of compressing the video. Strictly going for uncompressed now as the selling point. I'd load it up with Cineform and DNxHD as well so the end user could use that if they wanted. Not sure about cross platform compatibility with Matrox or PreRES.
I agree that the Shuttle is only $330, but there's a catch. SSD's are far from "inexpensive" as the BM site claims. You're talking $400-$600 just for the drive. We're now in the $1k ballpark to use a Shuttle for 30 minutes of uncompressed footage. Tack on at least another $500 for another 30 minutes. SSD's are notoriously unreliable still too. I've read too many horror stories how "Joe's" new expensive SSD died after just two or three months.
Mine would allow for approx 1.5hrs of uncompressed 1080p30 and 24-bit, 192k uncompressed 2 channel audio. Not sure how my battery life would be at this point, as I haven't gotten that far yet. I need to make sure some of the newer, smaller, processors are up to the task of capturing. I don't see why not though. It's just handing off the information to the drives and not doing any compression. Something in the sub-$1k range that can do this has been missing from the market for a long time now and not everyone has the time, or knowledge, to hobble up something on their own.
I have been interested in this type of capture method and ended up getting a laptop (lenovo w520) and using the BM shuttle usb3.0 device. Battery life is good and capturing in cineform can be done to the internal drive for a few hours. Hard to justify any improvement by going uncompressed. Cineform capture is only beneficial when the codec is stressed due to high speed action or color correction.
BM also does these...
http://www.blackmagicdesign.com/products/ultrastudio
...which give you both compressed and uncompressed SDI/HDMI full HD recording starting at $145, if you have a computer with thunderbolt.
*Balanced audio hack* *Variable ND filters* *HV20 vs. Film* "All animals are equal, but some animals are more equal than others." (George Orwell: Animal Farm)
There are still people that want uncompressed though. I wouldn't even attempt at competing with codec based capture devices. Also, look at reviews on Amazon for the shuttle. The shuttle is very picky about the computer it's attached to.
http://www.amazon.com/Blackmagic-Des...owViewpoints=1.
Even though it's marketed as being a "low budget" capture solution, the required computer equipment is generally (not always) not low-budget.
I saw those the other day Halsu. They do look pretty nifty. But unless the end user is capturing to a SSD or RAID-0 array, they aren't capturing uncompressed 1080p to a single drive like that picture would lead one to believe. A single, standard, rotating drive isn't capable of handling the bandwidth required for such a task. An uncompressed 1080p30 stream requires in the neighborhood of 200MB/s of sustained throughput (give or take 20-30MB/s). The best drives out there barely get up to half of that and even 2 WD Velociraptor 10k drives in RAID-0 can struggle to hit 150MB/s. We just aren't there technologically yet to use a single rotating drive to capture 1080p in an uncompressed state. So again, not aimed at the low budget film maker as an uncompressed footage capture solution. And then there's the whole Thunderbolt interface thingy. Thanks for yet another interface Apple.
I did some rough number crunching. This "thing" I'm thinking of designing would get roughly 1-1.5 hours of record time from a 7 pound battery. Now mind you, this is a sealed lead-acid type battery. I wanted to see how much the cheapest power solution would weigh. Haven't looked at lighter alternatives yet. And so far, my production cost is still on track at being able to provide a product in the $600-$700 range. If this would cost more than $700 to sell, I will throw in the towel. I'd offer support and warranty of 1 year as well and no skimping on cheap, no-name parts.
**EDIT*
I don't want anyone thinking I'm bashing the products being offered by BM or Atomos. They are indeed fantastic products. Just not really aimed at the really low budget videographer.
Last edited by zagnut; 2012 December 15th at 00:10.
That thing was developed by Intel. Apple only had ius prima notcae.Thanks for yet another interface Apple.
...and this other (which it isn't) interface allows you to exactly get the speed you want from spinning drives.
"It is dark the other side. Very dark!" - "Oh, shut up and eat your toast!"
Ah ok. I was under the impression that Apple designed it.
On the topic of spinning drives, it is impossible to get 200MB sustained throughput on a single spinning drive. If it were a magical interface, then SATA II already would have worked with it's 300MB/s bandwidth. SATA III exceeds this even more with 600MB/s. Thunderbolt ups this to double again, at 1280MB/s. It's not the interface that can't do it. Heck, even the 1394 interface can capture an uncompressed 1080 stream. The bandwidth and throughput has been available on our equipment buses for a long time now. But technology still hasn't caught up to allowing sustainable transfer speeds, in a spinning drive, required by an uncompressed 1080p stream. We're talking about Megabytes, not Megabits.
If you don't believe me, look it up. Here's something to start with, updated in Oct this year. Scroll down to "3.5" Desktop Hard Drives" and look under performance -
http://www.storagereview.com/best_drives
Look at whats required to reach just over 200MB/s, two enterprise edition VelociRaptors. Even a single one of these awesome drives can't reach the kind of throughput that BM is leading us to believe -
http://www.storagereview.com/western...sd_alternative
More real world figures (these on the last page are benchmarks using SSD's. Stats from a year ago shouldn't really be any different than what we have today. So feel free to go back a page or two) -
http://hardforum.com/showthread.php?t=1284410&page=21
Some quick math -
1920 * 1080 = 2073600 number of pixels
2073600 * 24 = 49766400 24-bit color (16.7M colors)
49766400 * 30 = 1492992000 30 frames per second
1492992000 / 8 = 186624000 convert to bytes/sec
186624000 / 1024 = 182250 convert to KB/sec
182250 / 1024 = 178 convert to MB/sec
1080p30, 24bits, uncompressed requires 177.978515625 megabytes per second of bandwidth
1080p24 would require 142.3828125 megabytes per second.
1080p60...well, it would require in the neighborhood of 350MB/s (355.957). Looking at a RAID-0 array with a minimum of 3 velociraptor's. Even with 3, I wouldn't trust it to not lose the occasional odd frames.
Even a single enterprise velociraptor would probably struggle with 1080p24, leading to the occasional dropped frame(s). Don't move the mouse! Oops...gotta re-shoot.
I hate to be the one to burst everyone's bubble, but those devices and the Thunderbolt interface are incapable of making a laptop, or desktop, with a single spinning drive into some magical device capable of capturing uncompressed 1080p. Are they a fantastic device for capturing to a near lossless codec? Absolutely! Can they capture 720p30 uncompressed on "Joe Blows" laptop? Yup. But that's not the market I'm looking at targeting.
As for further development on my mind, I may be throwing in the towel. I'm looking at a cost of around $550 to make my solution. And that's just for components and not my time putting it together and configuring it. This doesn't leave much room profit. I really don't want to be selling in the $800 range. I need to look at a business license and see what cost I'm looking at as far as wholesale goes.
Last edited by zagnut; 2012 December 16th at 21:35.
Thunderbolt (as storage connector) was not meant for single drives, but for RAID arrays. The Pegasus and G RAID stuff is pretty impressive and fast enough - despite your numbers.
But I think you are a bit on an off track with your obsession for uncompressed. For field capturing you only need a laptop with either a card port or Thunderbolt, a RAID array (LaCie has some very compact stuff there), and a I/O card that compresses you SDI or HDMI output to ProRes or DNxHD. Those codecs are near oosless and good enough even for broadcast, so why would an editor want to fuck around with uncompressed?
I doubt your device would be a success.
"It is dark the other side. Very dark!" - "Oh, shut up and eat your toast!"
I do think you may be right in me being late to the game. I realize there isn't much call for uncompressed capture. But nonetheless, there is a market for it, albeit a small one. And all of this stuff between you and I about interfaces and their respectable bandwidth as being suitable for capturing uncompressed is really nonsense. There's no need for a new interface, as a eSATA would suffice just fine. Guess it just irks me how us, as consumers, are always being told that "this is better because it's faster!" Meh.
I haven't looked at or priced any portable RAID arrays, so I think I really need to do that before I even think about looking at a business license in order to obtain components at wholesale prices. Same goes for reasonably priced SDI cards, if a such thing exists. Thank you for opening my eyes to this.
Let me ask you and the others here...IF you wanted to capture uncompressed, what would you look for, aside from portability and price?
**EDIT**
Holy crap! Just priced arrays with transfer speeds spec'd high enough for capturing uncompressed, looking at about $400 to start and the limit is the sky. Ok, I can win on this aspect, time for SDI pricing.
Last edited by zagnut; 2012 December 16th at 23:25.
You would have to beat those guys:
http://www.matrox.com/video/en/products/mxo2/specs/
http://www.aja.com/pdf/Ki_Pro_Family_Brochure_WEB.pdf
I know you are looking into this already for years, but, to be honest, to build this project wouldn't be worth it for me.
What is your target group, btw?
"It is dark the other side. Very dark!" - "Oh, shut up and eat your toast!"
On a more practical note: What OS and capture software would you use? Did you think of plugins for common NLE?
"It is dark the other side. Very dark!" - "Oh, shut up and eat your toast!"
Well, before I continue more, I need to look into a business license so I can get wholesale pricing on components. Otherwise, it wouldn't be in anyone's interest. I want to keep cost to the end user at no more than $700. OS would have to be Linux based in order to avoid additional costs of licensing. As for capturing software, I haven't gotten to that point yet. But again, something in the open source domain. I haven't considered plug-ins for any NLE's, as it wouldn't be intended to be used for editing the footage itself. I would intend on making this connect to an editing station / computer in a similar fashion as an ordinary external drive would connect. Or possibly as an NAS.
Again, I've halted any further research into making something like this due to where the final price stands as of now. And unless I could secure better component pricing, this just sits as yet another idea that didn't go anywhere.
Linux? Then you'll have to work with ogg/Theora - still barely supported in the closed source world. As you'll sell your machine commercially, you won't be able to get around licenses for anything else.
"It is dark the other side. Very dark!" - "Oh, shut up and eat your toast!"
Why ogg? Linux supports WAV. As for the capturing software itself, VirtualDub is a viable option and is open source. The license states that it may be redistributed freely. However, it would have to be run in WINE. Not very user friendly and would have to make some type of executable script to start both with one click or button press. Avidemux is another option that wouldn't require WINE, but I'm not sure how compatible it is with uncompressed.
However, since AVI and WAV were both developed by Microsoft, there may be some type of odd, underlying licensing issue there. Will have to dig deeply into this subject to find out. Again, you made me think of another possible issue that I didn't think of. Thank you. Should give you a consultation fee if I continue to move forward.
Posting Permissions
Reply With Quote