KOMODO....

[Robert Hofmeyr]

IIRC Jarred mentioned something about a new RGB recording mode in the discussion during the first Red Solitary series livestream. Was busy with kids so only half watching. Anybody else catch it?

Found it: "Komodo you can record a RGB compressed format.. Just not allowed to talk about that yet"

 

[Phil Holland]

500MB/s would seem feasible

I'm done testing for now, but I've done a fairly elaborate round of media tests of compact commercially available media cards. CFast 2.0, SD, and CFexpress. I dedicated two of my systems to run long term tests for a week and I have developed some data as well as criticisms regarding how this style of media is even marketed/labeled.

Sustained writes even at some of the published tested speeds seems to be not accurate in most cases. CFexpress suffers from this as well, and it's just getting a bit weird.

The problems come from a few different directions, but when you fully test the read/write of all compact commercial grade media, a lot needs to be looked at. It's been rather sobering actually testing the newer cards and standards fully and it has me moderately in "sad face", but still the newer stuff is faster and reputable brands show more consistency than others for sure.

There's an incredible lack of full blown data rate tests out there, which is why I'm testing for myself, but quickly you'll see mostly benchmarks that don't take into account:
- performance at various capacity landmarks (even effects reads!)
- degradation over life of media
- temperature performance variance
- consistency/variance across multiple copies

All of those things aren't discerned without much deeper and unfortunately longer form testing and that will be what truly reveals what potential the media holds for any camera system's data rate needs.

I'll be penning a strong, very strong worded letter to various media manufacturers regarding writing larger files, testing fully, and publishing accurate data rates for motion picture professionals. Commercial grade media is used in many applications, but we need a much clearer and more accurate "answer" when it comes to burst, sustained, and performance across capacity levels. Stickers are cool to attract kids to media. Quick data rate checks are nice too, but it's not how we're using it. It's not how anybody uses it. Often we might not be filling up cards, but really we need to know what happens at 1/4, 1/2, and full capacity and what to expect.

Manufacturers test and will work within a limited data rate range (as well as whatever the hardware potential is) to have enough over provisioning to ensure no dropped frame, recording errors, etc. Generally you work in some sort of math to predict the performance across the lifespan of the media as well.

I'm certain RED is testing media as well for Komodo, we've been told as much, and I'm looking forward to what makes the recommended media list as well as guidelines for general buying.

 

[Robert Hofmeyr]

I'm done testing for now, but I've done a fairly elaborate round of media tests of compact commercially available media cards. CFast 2.0, SD, and CFexpress. I dedicated two of my systems to run long term tests for a week and I have developed some data as well as criticisms regarding how this style of media is even marketed/labeled.

Sustained writes even at some of the published tested speeds seems to be not accurate in most cases. CFexpress suffers from this as well, and it's just getting a bit weird.

The problems come from a few different directions, but when you fully test the read/write of all compact commercial grade media, a lot needs to be looked at. It's been rather sobering actually testing the newer cards and standards fully and it has me moderately in "sad face", but still the newer stuff is faster and reputable brands show more consistency than others for sure.

There's an incredible lack of full blown data rate tests out there, which is why I'm testing for myself, but quickly you'll see mostly benchmarks that don't take into account:
- performance at various capacity landmarks (even effects reads!)
- degradation over life of media
- temperature performance variance
- consistency/variance across multiple copies

All of those things aren't discerned without much deeper and unfortunately longer form testing and that will be what truly reveals what potential the media holds for any camera system's data rate needs.

I'll be penning a strong, very strong worded letter to various media manufacturers regarding writing larger files, testing fully, and publishing accurate data rates for motion picture professionals. Commercial grade media is used in many applications, but we need a much clearer and more accurate "answer" when it comes to burst, sustained, and performance across capacity levels. Stickers are cool to attract kids to media. Quick data rate checks are nice too, but it's not how we're using it. It's not how anybody uses it. Often we might not be filling up cards, but really we need to know what happens at 1/4, 1/2, and full capacity and what to expect.

Manufacturers test and will work within a limited data rate range (as well as whatever the hardware potential is) to have enough over provisioning to ensure no dropped frame, recording errors, etc. Generally you work in some sort of math to predict the performance across the lifespan of the media as well.

I'm certain RED is testing media as well for Komodo, we've been told as much, and I'm looking forward to what makes the recommended media list as well as guidelines for general buying.

Are you going to publish your tests Phil?

 

[Gregory Karydis]

Thank you for all the incredibly thorough testing you are performing Phil!

 

[Phil Holland]

Are you going to publish your tests Phil?

Probably. Still waiting on a couple cards to drop. But I suspect this would be a job for Linus Tech Tips to make a video on to sort of expose some of the issues here. Like even the print burst speeds have some variance once cards are over half full. Might be too hot of a topic though and could piss off sponsors.

Thank you for all the incredibly thorough testing you are performing Phil!

Anytime Gregory. Thus far I think I've done the most "work" surrounding the lens mount and adapters related stuff, which is stickied already in the Komodo section. I think creatively and technically that's likely the most useful information because we know we can power and record already

Once things really get rolling I'm hoping to show off some rigging and use case scenario stuff. As well as footage and all that.

 

[Gregory Karydis]

...
Once things really get rolling I'm hoping to show off some rigging and use case scenario stuff. As well as footage and all that.

That's awesome!

I'm preparing some rigging options as well

 

[Nick Morrison]

Don't know if this was already answered, but will the Komodo have simultaneous Red Raw and ProRes at any resolution(2k, 4k)?

Not sure, but given that Komodo has 4K SDI I think recording R3D's on board, and 4K Prores on a separate recorder should be totally doable. This dual recording also adds a nice safety net we haven't had before on a RED platform. Having a 4K safety net is especially welcome.

 

[Blair S. Paulsen]

Not sure, but given that Komodo has 4K SDI I think recording R3D's on board, and 4K Prores on a separate recorder should be totally doable. This dual recording also adds a nice safety net we haven't had before on a RED platform. Having a 4K safety net is especially welcome.

I alluded to this in my previous post. IMO, best practice would be to have two independent recordings created on separate media. One would be R3Ds in whatever format/compression ratio you desire and an RGB version at up to 4K recorded from the 12g SDI tap into an outboard recorder. Since Komodo's built in LCD screen is not what most people would want to use as an operator's monitor, one of the monitor/recorder combos in the 5-9" range would seem like the obvious play.

I would expect that you could choose whatever standard resolution output over SDI you wanted even if you were writing 6K R3Ds to CFast2 internally.

My other note was about using the SDI pipeline to allow off camera recording of R3Ds. The base SDI spec is just a data signaling protocol - Serial Digital Interface. In our industry, the chipsets/firmware at both ends are designed to honor a more restrictive set of specs to facilitate compatibility. (Details can be found on the SMPTE site).

I guess I'm that "what if" guy. What if that 12g SDI spigot could be used for an R3D bitstream instead of an RGB video signal? Since 12g is theoretically capable of 1,500MB/s it could potentially support uncompressed recording to external storage. Yes, Virginia, that might be serious overkill and would eat drive space like a teenager at a free pizza buffet. That said, what if you chose a light compression ratio like 3:1 or 4:1. You would still need a lot of fast storage, but for some use cases it really opens up possibilities. Perhaps the EXT port will support 10gE for tethered recording and make this moot. If not, and if it's not a technical Manhattan Project to implement, it could make the Komodo very attractive for a lot of advanced applications beyond the entertainment space.

Cheers - #19

 

[Blair S. Paulsen]

I'm done testing for now, but I've done a fairly elaborate round of media tests of compact commercially available media cards. CFast 2.0, SD, and CFexpress. (snip)
Manufacturers test and will work within a limited data rate range (as well as whatever the hardware potential is) to have enough over provisioning to ensure no dropped frame, recording errors, etc. Generally you work in some sort of math to predict the performance across the lifespan of the media as well. (snip)

Don't get me started on the phrase "up to ..." in media specs. :emote_head_explode:

Buying actual media and testing it in the kind of conditions we need it to perform in is critical. I can't thank Phil enough for his work on this issue.

Cheers - #19

 

[rand thompson]

Not sure, but given that Komodo has 4K SDI I think recording R3D's on board, and 4K Prores on a separate recorder should be totally doable. This dual recording also adds a nice safety net we haven't had before on a RED platform. Having a 4K safety net is especially welcome.

Thanks Nick!

I was also thinking about the recording of R3d files to the Cfast card internally and Prores to an external device. Would this come at the price of limiting the compression ratios of the R3d files so that output through the 4K SDI out can be enabled. To put it simply, would a 10:1 or 12:1 compression ratio be required in order to output through the 4K SDI out? Anything lower like an 8:1 compression ratio would disable the 4K SDI output in the menu.

 

[Nick Morrison]

Thanks Nick!

I was also thinking about the recording of R3d files to the Cfast card internally and Prores to an external device. Would this come at the price of limiting the compression ratios of the R3d files so that output through the 4K SDI out can be enabled. To put it simply, would a 10:1 or 12:1 compression ratio be required in order to output through the 4K SDI out? Anything lower like an 8:1 compression ratio would disable the 4K SDI output in the menu.

I don't think the R3D compression internally has anything to do with the signal being sent out SDI. They are separate. Even now for example on DSCM1 or DSMC2 your compression settings have no affect whatsoever on the SDI signal. That never changes. Should be same here. The SDI signal on Komodo should always pump out 4K and you can always record that (whether 4:4:4 or 4:2:2 I haven't checked or kept up with that).

If the 4K SDI signal can always capture 4K 4:4:4 that would be amazing indeed for external recording. But also the 4K SDI out is going to make Komodo incredibly useful for 4K streaming etc, esp as live streaming has become so relevant now and probably will remain going forward.

 

[rand thompson]

Nick,

I never owned a red camera so I didn't know how that would work, Thanks for laying that out for me! Hopefully we'll get as low as a 5:1 compression ratio but more than likely it will probably be 8:1.

 

[Nick Vera]

I am not sure if this has been asked. I know you can power a Red Monitor through the top of the camera like a traditional Red DSMC body. Are there any plans to allow third party monitors like SmallHD or Atomos monitors to be powered through the camera, to avoid using batteries for the monitors? Thank you.

 

[Bob Gundu]

I am not sure if this has been asked. I know you can power a Red Monitor through the top of the camera like a traditional Red DSMC body. Are there any plans to allow third party monitors like SmallHD or Atomos monitors to be powered through the camera, to avoid using batteries for the monitors? Thank you.

Actually existing Red monitors are not functional on the Komodo. You’ll have to use a 3rd party monitor. It will
Most likely have to be powered off its own power unless you use bigger voltage batteries like Vmount batteries that can handle external accessories.

 

[Nick Vera]

Thank you. Yeah having a gold mount battery pack would be great! But I thought the camera used Canon batteries. Hopefully Red or Tilta make a Gold Mount back!

 

[mikeburton]

Jesus, this thread is getting out of hand long. Is it possible to get a locked thread with just up to date information / specs? Not sure why that is always so hard here?

 

[Gregory Karydis]

Jesus, this thread is getting out of hand long. Is it possible to get a locked thread with just up to date information / specs? Not sure why that is always so hard here?

 

[Andrew Reese]

Actually existing Red monitors are not functional on the Komodo. You’ll have to use a 3rd party monitor. It will
Most likely have to be powered off its own power unless you use bigger voltage batteries like Vmount batteries that can handle external accessories.

What is the attachment / connectors shown on the top of Komodo for then? Looks very similar to DSMC2 connector for monitors but altered slightly. Maybe this has been covered but I was not aware if so....

 

[yann LHENORET]

What is the attachment / connectors shown on the top of Komodo for then? Looks very similar to DSMC2 connector for monitors but altered slightly. Maybe this has been covered but I was not aware if so....

Every answer you’re looking for can be found in Jarred’s posts:
#2036
#2046
#2444
#2463

Enjoy!

 

[Andrew Reese]

Well thats a shame... I know one guys opinion doesn't matter that much but I was really hoping for some DSMC2 integration for the reasons noted below and to be able to use accessories back and forth between my Gemini and if I get Komodo. I used to have an FS5 + shogun inferno and the need for external monitoring solutions, cables, and extra batteries is something I was really happy to get away from. Small HD / Atomos monitors are nice but I am still sold on the simplicity and cleanness of DSMC2 monitors despite the lack of brightness. I was really hoping Red would make a new version of the ultra bright in a 4.7" or 5" which in my opinion would be the perfect monitor for Komodo. Similarly considering batter life and reliability of mobiles I would prefer to use that as a secondary control device rather than for primary monitoring. I shoot in a lot of low temperatures in winter where phone reliability just won't cut it.

My wishlist for komodo from what we know now.


- Make adapters for Red EVFs and touchscreens, Think I´m not alone when I say I possibly buy a Komodo, but I really don't feel like kitting out another camera with a new set of accessories, I got plenty laying on the shelfs allready. I rather take the EVF from the main cam and put it on Komodo and run the main cam with only monitor. And I hate to have stuff that is not interchangable. I preferably use red cameras for that reason, I dont want to rent a lite sony cam for the behind the scenes as it comes with its own set of batteries, media, card readers, etc. Rather use a scarlet then. So even if it´s a different camera make it compatible with what we got all ready.

- Make good adapters for IOS phones. Skip the whole Hydrogen/android thing. I Love the DJIGO iphone app. Something like that is all thats needed. Make a plastic, low cost loupe to use with iphones. Then I can keep the screen and EVF on the main cam...

- Make an adapter that sits in-between the two battery mounts off Komodo and a V-lock battery. Sure, small batteries is nice at times. But frankly 90% of the time battery size is no issue especially not compared to buying and carrying multiple battery types and chargers around. Being able to power everything with one sort of power is gold. Whatever the small betteries are Im sure I for one would rate have a big brick on the back of the back of the camera pretty much always.

- Make a komodo Minimag module. Same there, having to buy new media types is not really on anyones wishlist. Instead of just going for a new standard, make a CFast to minimag module and just lower the cost for mini mags and sell a shitload of them.

- Finally and this is key. Make sure Komodo has a high frequency gyro. Preferably one that writes a few times per frame. Without saying to much I can confirm that we have been hard at coding the last couple of weeks and without saying to much. I can only confirm we are onto something. Something in camera gyro related and yes, Komodo with it´s global shutter would be sort of the king of cameras if it also has a good highfreqency interal gyro that writes multiple values per frame. That would be bigger than RF mount, small formfactor and low price and high framerates combind. Im not shitting.