Our Project to Create a Cost Effective Solid State HD Video Recording Device Begins

[Frank Brodkorb]

Ever thought of something like that:

Quad-CF PCI Controller from Addonics

Frank

[Mark Job]

Originally Posted by Frank Brodkorb

Ever thought of something like that:

Quad-CF PCI Controller from Addonics

Frank

...Hi Frank. This is not really a practical way for us to go. For one thing, we do not want to base our SSDR on CF card media because it is more expensive than other types of solid state media. Secondly, CF card media availability is not so universally accessible in all markets in yet. The larger physical foot print of the CF card socket and media forces a larger system layout and over box size, which we want to avoid.

[Roberto Lion]

Hi!!


any update with your new recorder?? i'm still interested about it..when you'll be ready, please let me know: info@onboardcamera.it

bye!

[Mark Job]

Right now folks are waiting for me to inject a little more cash, which I will do shortly, but our project is *far* from the prototype stage at this point I'm sad to say. I am seriously considering making the recorder a full uncompressed RAW recorder only, with perhaps, the capability to record HDV via FW interface.

....The reason why I'm leaning more toward a full uncompressed only recorder, is to.....

a) Get around paying expensive royalties to the MPEG folks.

b) The new SDXC spec will definitely be fast enough and large enough in capacity (1 TB +) to make uncompressed capture practical.

[Daniel McPake]

Hi there,

I know I may be jumping the gun but I was just wondering if you have had any updates, I have been following this project in the shadows and it seems amazingly promising!

Have you decided to make it an uncompressed RAW recorder only?

I wish you luck!

[Mark Job]

Hi friends:
As of today we have made the decision to make our new SD Card SSDR record totally in the uncompressed 10 Bit 4:2:2 color space. We will also build in 4:4:4 video recording capability. At this time, we will not be dealing with any compressed recording codec. We may develop our own proprietary codec for compressed recording at a later date.

[Mark Job]

Hi Friends:
My apologies for my lack of updates in this thread. The project has slowed to a crawl, but has not stopped. I got a fully functioning prototype board which recorded HD-SDI Full Raster 1920 x 1080 in 4:2:2 uncompressed in 10 bit once, then promptly failed and never recorded (or played back) another signal since ! I think I blew the circuit because our power supply is flawed. I think our power supply is unstable, and you need really good stability or you blow all your (Expensive chips).
I have decided to offer 2 modes: Compressed & uncompressed. The compression is our own proprietary codec (Because we're too cheap to pay anyone else any freaking royalties). I should start a contest to see what we should call our new codec ;-) (How about the new Markvision codec ?) I will suggest whoever winds up purchasing one of our boxes to playout our compressed stuff into their FCP or Avid NLE via the HD-SDI, so they have the option to capture in whatever codec they want Which we don't have to pay royalties for !). Uncompressed recording is well.......uncompressed. Play it out in realtime.

[Max Petin]

What about Avids DNxHD? I always thought its a free and open source codec? BTW: Thanks for your approach towards an affordable recording device. Hope to hear from it soon... :)

[Tim Kolb]

Originally Posted by Mark Job

Hi Friends:
My apologies for my lack of updates in this thread. The project has slowed to a crawl, but has not stopped. I got a fully functioning prototype board which recorded HD-SDI Full Raster 1920 x 1080 in 4:2:2 uncompressed in 10 bit once,...
I have decided to offer 2 modes: Compressed & uncompressed. The compression is our own proprietary codec (Because we're too cheap to pay anyone else any freaking royalties). I should start a contest to see what we should call our new codec ;-) (How about the new Markvision codec ?) I will suggest whoever winds up purchasing one of our boxes to playout our compressed stuff into their FCP or Avid NLE via the HD-SDI, so they have the option to capture in whatever codec they want Which we don't have to pay royalties for !). Uncompressed recording is well.......uncompressed. Play it out in realtime.

OK, I've been reading this thread on and off over the last year+...and I've been uncharacteristically mute.

It's OK to envision making a data recorder that records 10 bit uncompressed HD 4:4:4 to a device the size of a matchbook...but eventually you have to take a few things into account.

Not paying royalties is a great goal, but the bottom line is the industry has several reasonably established codec families that have spawned equally reasonable workflows for acquistion-to-post.

Uncompressed HD video (or 2K or 4K for that matter) end-to-end workflow is a concept that always sounds great when discussing how we all want to be uncompromising image makers, but when the practicality is explored, compressed acquisition just keeps winning. Uncompressed footage simply makes no sense in the context of a location crew...HDcamSR is clumsy enough, and that's MPEG4. Uncompressed footage would take incredible amounts of data storage to store redundantly in the field, not to mention that the data dumps would be less than even real time without some sort of a custom card reader and a serious workstation in the field (you'll likely not keep up with a laptop and USB drives for transferring this sort of stuff)

Now...edit. Uncompressed doesn't take a lot in the CPU torque department...it's about throughput. However, to get multiple streams of uncompressed requires serious harddrive speed with extremely well developed infrastructure, not to mention cavernous capacity.

At some point, a discussion about what the real advantages are between uncompressed and say, CineForm 10 bit, 4:2:2 needs to happen. For a customer, there are precious few (if any). Yes, CineForm is a proprietary codec, but if a customer can buy CineForm and streamline the whole facility and make post production easier, it's a small price to pay.

Whether or not -you- want to pay royalties has little to do with what makes sense for your customer. I use a wide variety of established codecs. My post process is predictable, my post and acquisition infrastructure is manageable and requires an investment that makes sense.

You're bent on making a device that will affect workflow downstream in a way that will make editing the footage completely impractical for those who are shopping in the price range you propose...and the part of the market that would actually find the specifications you propose compelling, is likely not large enough for you to really see a return on your investment through any serious quantity of sales, not to mention that the pricepoint you propose will make it hard for that group to take the device seriously. Not to mention proposing a data recorder that has to be transferred into a post workstation REAL TIME? Are you serious? Even HDcamSR can go in at 2X for crying out loud. You are now rationalizing the need for your device waaaay past a point where practicality exists anymore.

There are any number of innovative ways to deliver a product with really usable features to a market large enough to make your efforts worthwhile...with a file size small enough to be practical.

...at this point, I'd consider what is possible to actually sell and support before you continue to simply explore what you can conceivably -develop-.

(...in the nicest way possible.)

[Mark Job]

Originally Posted by Tim Kolb

OK, I've been reading this thread on and off over the last year+...and I've been uncharacteristically mute.

It's OK to envision making a data recorder that records 10 bit uncompressed HD 4:4:4 to a device the size of a matchbook...but eventually you have to take a few things into account.

...OK. Stop there. I am envisioning a device which records 10 bit 4:2:2 not 10 bit 4:4:4. The Device will *Also do 4:4:4 color space recording but in 12 bit color precision code. The actual size of our recorder will be larger than a match box. I'm not sure where you got the idea it would be the size of a matchbox ?

Originally Posted by Tim Kolb

Not paying royalties is a great goal, but the bottom line is the industry has several reasonably established codec families that have spawned equally reasonable workflows for acquistion-to-post.

...I think you may not have realized the simplicity of our approach ? If you record uncompressed, then play that signal straight out in realtime, you then can *Compress* the data when transferring to your Avid or Vegas, or FCP application using the codecs **in those NLE's.* It is not necessary to perform the compression when capturing the data to SD cards.

Originally Posted by Tim Kolb

Uncompressed HD video (or 2K or 4K for that matter) end-to-end workflow is a concept that always sounds great when discussing how we all want to be uncompromising image makers, but when the practicality is explored, compressed acquisition just keeps winning.

...It really depends on your application. I am envisioning a box which is primarily intended for high end work (Digital Cinema Origination) with an optional *Compressed mode (I already mentioned it would have compression as an option-We're just not sure what yet)

Originally Posted by Tim Kolb

Uncompressed footage simply makes no sense in the context of a location crew...HDcamSR is clumsy enough, and that's MPEG4. Uncompressed footage would take incredible amounts of data storage to store redundantly in the field, not to mention that the data dumps would be less than even real time without some sort of a custom card reader and a serious workstation in the field (you'll likely not keep up with a laptop and USB drives for transferring this sort of stuff)

....With the new SDXC format already launched by Sandisk (See their new 64 GB SD card) uncompressed recording not only makes sense, but is practical with 4 x 64 GB raid 0 striping.

Originally Posted by Tim Kolb

Now...edit. Uncompressed doesn't take a lot in the CPU torque department...it's about throughput. However, to get multiple streams of uncompressed requires serious harddrive speed with extremely well developed infrastructure, not to mention cavernous capacity.

....So ? The Harris Altitude HD NLE offers two streams of full 10 bit uncompressed realtime HD (Although it has just been EOL'd), so does Avid Media Composer, so does FCS 7.x if you use specified raids. Yes, I agree these are expensive. But you don't have to capture Uncompressed as uncompressed do you ? Let your NLE compress it in realtime during your video/audio capture. To edit practically, you only require enough uncompressed storage space for one hour. Not so big a deal as it used to be. HDD costs have come way down actually.

Originally Posted by Tim Kolb

At some point, a discussion about what the real advantages are between uncompressed and say, CineForm 10 bit, 4:2:2 needs to happen. For a customer, there are precious few (if any). Yes, CineForm is a proprietary codec, but if a customer can buy CineForm and streamline the whole facility and make post production easier, it's a small price to pay.

...Agreed. Buy your Cineform app and capture your uncompressed stream to that in realtime.

Originally Posted by Tim Kolb

Whether or not -you- want to pay royalties has little to do with what makes sense for your customer. I use a wide variety of established codecs. My post process is predictable, my post and acquisition infrastructure is manageable and requires an investment that makes sense.

...Yes, and it will continue to be predictable as long as you can capture in realtime to whatever NLE & codec you need to use.

Originally Posted by Tim Kolb

You're bent on making a device that will affect workflow downstream in a way that will make editing the footage completely impractical for those who are shopping in the price range you propose...

...I don't think so.

Originally Posted by Tim Kolb

and the part of the market that would actually find the specifications you propose compelling, is likely not large enough for you to really see a return on your investment through any serious quantity of sales, not to mention that the pricepoint you propose will make it hard for that group to take the device seriously.

....Really ? A small, portable recorder capable of both uncompressed and compressed recording in both 4:2:2 10 bit & 4:4:4 12 bit via HD-SDI, Dual Link HD-SDI, and 3G HD-SDI with FW and HDMI thrown in. The cameras will do the compression (HDV) not the recorder ;-) (Think about it), HDMI cameras will do their compression (Not the recorder again) (Think about that) - What's the mystery ?

Originally Posted by Tim Kolb

Not to mention proposing a data recorder that has to be transferred into a post workstation REAL TIME? Are you serious? Even HDcamSR can go in at 2X for crying out loud. You are now rationalizing the need for your device waaaay past a point where practicality exists anymore.

...Perhaps ? We think we're developing a device by thinking out side of the box (If you'll pardon the pun), and we'll get it right sooner or later. Right now it's half baked and not ready for prime time, and I won't show up at NAB with one in my pocket until it does !! BTW, realtime is not fast enough a capture rate for you ? Clip based import is often slower, except for Avid's fast import of certain codecs like HDV (Finally !)

Originally Posted by Tim Kolb

There are any number of innovative ways to deliver a product with really usable features to a market large enough to make your efforts worthwhile...with a file size small enough to be practical.

...at this point, I'd consider what is possible to actually sell and support before you continue to simply explore what you can conceivably -develop-.

(...in the nicest way possible.)

.....We're not Convergent Technologies ! We're not really in a position to compete with major companies like that, and frankly, we don't care to. We are a small team in a garage and this is all we are at this point . I don't care if I ever sell one. I *DO care if I make a successful prototype for my own use . I think creating your own SSDR is the coolest thing ever ! I'm trying to please myself :-) If what we perfect pleases others, then, and only then may we proceed as you suggested, but for now this is a hobby project created by a film maker in Montreal, Canada who drinks way too much coffee ! ;-)

[Tim Kolb]

I assumed you were making something to sell...you are obviously investing a lot of time and effort in this...

You've mentioned a price point a couple times in the thread...

If this device is only for high-end origination, the higher you go in the market place, the fewer buyers there are to capture...I think the range you've mentioned is far too inexpensive for that market.

As far as real-time capture...you must be referring to FCP I assume. i don't use FCP...I'm RT native to anything including DSLR and RED...so yes, real-time capture on the last project I did which was 15 minutes in edited length and had over ten hours of interview source material would have been quite a handicap.

...and obviously an hour of storage would not have been enough. I've done some television commercials where there has been an hour or less of source, but i think when you mention one hour of storage, you must be thinking that editors only do commercials...and they never have two in the house at the same time.

I'm not sure what you mean when you say the cameras will do the compression...HDSDI and HDMI are not general purpose data-pipes like FW. They send 'digital baseband'...uncompressed down the line. You can't use an HDV camcorder to record and run a recording device downstream via HDMI and get HDV... The downstream recorder gets uncompressed. HDSDI as well.

Good luck with it...I guess I'm thinking in terms of market, not product. It seems like a market that S Two and Codex pretty much have already...it's small and exclusive and requires post production resources that, while less expensive than ever before, are still at the top of the scale.

[Mark Job]

Originally Posted by Tim Kolb

I assumed you were making something to sell...you are obviously investing a lot of time and effort in this...

....Yes, I have invested a great deal of my own private cash into this project. You see, I can't think about the market or the marketing right now. My vision is narrow down a tunnel which only leads to the prototype stage at this point. The way we're looking at this device is more from a personal artistic and technical sensibility. I have asked myself what I want in a dream SSDR, and that something is pretty close to what I've been writing about in this thread over the past year. It is true I have asked the opinions of end users in this thread, and many folks have been kind enough to give extensive feedback (Like you as well :-) ). I think folks like you are much better *marketers* than I.

Originally Posted by Tim Kolb

mentioned a price point a couple times in the thread...

..Yes. We think this technology should be available to a broader range of independent videographers. Not just high end shooters, but I respect both, and have hoped for a more democratic device which could be easily adopted by the proletariat at large. I want to make this device retail for just under $1,000.00 US Dollars IF we ever get to the retail stage in this project. For now, I must build *The First* device which I want to see and use. Only when we have one in hand to take to NAB will we be happy and see what we can see at that point.

Originally Posted by Tim Kolb

If this device is only for high-end origination, the higher you go in the market place, the fewer buyers there are to capture...I think the range you've mentioned is far too inexpensive for that market.

...You are probably correct. The box is too inexpensive for what it can do. However high and low and digi-underground gorilla shooters will be able to embrace this device (If I ever actually get it to work !)

Originally Posted by Tim Kolb

As far as real-time capture...you must be referring to FCP I assume. i don't use FCP...I'm RT native to anything including DSLR and RED...so yes, real-time capture on the last project I did which was 15 minutes in edited length and had over ten hours of interview source material would have been quite a handicap.

...I operated for nearly 11 years in post with an old DPS Perception RT system with about an hour of storage. I used to work long edits in two ways....

A) Batch capture in really low quality just good enough to see what I was doing, then batch recapture at whatever online quality was required. In this way you could edit long or short form projects with little storage space. It seems as though many editors either don't know, or have forgotten the good old batch capture and re-capture feature. This is why most NLE applications still have this feature. It works through both the FW and RS-422 interface. Avid Media Composer, Final Cut Pro, Sony Vegas, Adobe Premiere all possess this basic post capability.

B) Capture and work in halves for long form projects. Edit one hour at a time of a 3 hour or 2 hour program.

Originally Posted by Tim Kolb

and obviously an hour of storage would not have been enough. I've done some television commercials where there has been an hour or less of source, but i think when you mention one hour of storage, you must be thinking that editors only do commercials...and they never have two in the house at the same time.

...See above. I was thinking BC and Re-BC, or one hour at a time and dump, then the next hour, etc.

Originally Posted by Tim Kolb

I'm not sure what you mean when you say the cameras will do the compression...

....This is a pretty straightforward, uncomplicated procedure. The HDV camcorder has an internal MPEG -2 encoder which dumps data into a recorder that results in an m2t transport stream file. In this case the recorder does nothing more than store the data - it *does not encode it.*

Originally Posted by Tim Kolb

HDSDI and HDMI are not general purpose data-pipes like FW. They send 'digital baseband'...uncompressed down the line. You can't use an HDV camcorder to record and run a recording device downstream via HDMI and get HDV...

...I never intended to. This would be a distortion of how my recorder functions.

Originally Posted by Tim Kolb

Good luck with it...I guess I'm thinking in terms of market, not product. It seems like a market that S Two and Codex pretty much have already...it's small and exclusive and requires post production resources that, while less expensive than ever before, are still at the top of the scale.

....I can't think about that right now. My goal now is to make it work. If I can't make it work, then it's all academic anyway.