URSA 4.6k VS RED Raven

i have been enjoying the debate between other cameras and what not on here... but i would like to ask you guys what you think of the RED Raven? is it the Dragon but way cheaper ?

i ask this cause i have the 4.6k on pre order and my line of freelance clients lose there crap over "well known" cameras even though the image is not much different. Is the Raven something BMD should be worried about ? RED is saying to get started you need around 10k pretty much the same as the 4.6.

All other differences aside, the costs aren't even close. Consumables aside, Raven 4K Brain + Red/DSMC2 1920x1136 Touch Monitor + DSMC2 EVF W/Mount Kit + RedStation Mini Mag + Redvolt XL module + DSMC Base production pack (front/top handles, shoulder mount and rails) + cables for everything = almost $20,000. The equivalent setup for an URSA Mini 4.6K would be camera + shoulder mount kit (with top, front/side handle w/extension, rods, and shoulder mount) + CFast card reader + URSA viewfinder + AB/Gold mount plate = ~$7000.

Batteries are about equal, with each 89Wh Redvolt XL at $350 and higher end 98 Wh gold mount battery being about the same (though you can get name brands, even entry level AB, for around $200 per 90 Wh batt). Not sure about chargers, though many gold mount batts can have cable chargers built in while the cheapest BM travel charger can run around $200 individually. However, a 120GB Red Mini Mag 240GB costs $1400, while an approved 256GB CFast2.0 costs ~$550 with higher speed to boot, and if you buy the C-box for $500, you can use even faster and larger SSDs, with the approved and highly regarded Samsung 850 Pros costing only about $200 for 512GB. So with 2 bats + charger + 1TB of storage, this runs the total up to ~$26,000 for the Raven and $8300 for the Ursa Mini 4.6K with the less expensive batts and the C-box + SSDs, or around $9,900 if you go all Cfast (and prices should continue to drop for both SSDs and Cfast, not so much for Redmag).

It would cost you around 2x as much to get a fully functional Raven as it would an Ursa Mini 4.6K.

Beyond that, since neither is available, the rest is just specs.

It's not just a cheap Dragon though; it's quite limited, probably so that it doesn't cannibalize the Scarlet line.

And to debayer Red footage with their own gear (Read: staying in their own ecosystem), you'd be looking at the Red Rocket which is around $5,000 as well..

Here is my view on it.

It will cost you more to get the Raven, but not 2x as much (probably 3-5k more for a basic kit). And keep in mind that redcode is really flexible, so in the media side you get more space for your money (even with cfast being more affordable than red mags). It will go down something like 9k for Mini 4.6k and 13k for Raven (for similar kit).

Now for 4/4.5k R3D you don't need a rocket to edit it. Computers came a long long way since the early days of r3d, and with a pretty OK one ($2K) you can edit things pretty fast (even with imacs - to some extend - you can edit 6k r3d). Now the Raven also support simultaneous ProRes recording, so you can do offline/online without penalizing the workflow, or just shoot prores (thought its only 2k, which is a big minus here).

I had my eyes on the Mini 4.6k since it's presentation (and with the footage now getting out I'm really impressed). With that said, my main reason of getting a camera next year is to rent out, and man people here don't know BlackMagic that much but they really know Red (producers and directors like it). So instead of starting some fist fights with my clients saying why they should go with the mini, for me it makes a business sense getting the Raven, paying the premium, but being able to charge a little extra and renting more without going "against the stream". And at the end of the day the Raven images will look great, just like the Mini 4.6k will look great.

Now something that might make it a little better than the current dragon is the fact that the Raven uses some of the new designs (internally and externally) found on the Weapon. Which gave the dragon sensor a little bit more of DR and a more clean images overall. I know its a though call, and I can't make the decision for you, but for me and my market here it makes more sense getting the Raven.

Just an addendum here, if I were going to use a camera only for personal projects I would save the extra money and be getting the mini.

It's all specs right now....granted we have some graded samples on the 4.6K, it's still all hearsay.

If you don't know what you want, wait it out and then research further once both are out in the wild. Then, way your business needs.

I will say that BM has come along a little bit in the professional industry and I think once the new sensor is out, it will make some extra headway.

At this point, there is no better. Specs are so close, it's personal preference. Since there's no footage to compare, I'd say look at the specs that are important to you, including price, and see what makes sense. It's so close that it basically doesn't matter. A talented person behind either is going to yield amazing results.

Decide what you personally like better and get that.

At this point, there is no better. Specs are so close, it's personal preference. Since there's no footage to compare, I'd say look at the specs that are important to you, including price, and see what makes sense. It's so close that it basically doesn't matter. A talented person behind either is going to yield amazing results.

Decide what you personally like better and get that.

Yes. Don't forget that one of the big draws to the mini in the first place was its more "ENG-ish" ergonomics.

It's funny that a lot of people were clambering for this, but then when RED announced the Raven, all the sudden people were willing to go back to that "box that needs rigged."

Not that there's anything wrong with that but you really need to decide what YOUR needs are.

It's a shame that specs and brands get in the way of good indie filmmaking these days.

I agree Justin. Find out what your needs are and the choice should be easy. I want to use my Ronin M and drone shots, so Raven is what I need. If I were just using shoulderrig, tripod etc I would get the Mini

I agree Justin. Find out what your needs are and the choice should be easy. I want to use my Ronin M and drone shots, so Raven is what I need. If I were just using shoulderrig, tripod etc I would get the Mini

Dustin, but yes.

Here is my view on it.

It will cost you more to get the Raven, but not 2x as much (probably 3-5k more for a basic kit). And keep in mind that redcode is really flexible, so in the media side you get more space for your money (even with cfast being more affordable than red mags). It will go down something like 9k for Mini 4.6k and 13k for Raven (for similar kit).

Now for 4/4.5k R3D you don't need a rocket to edit it. Computers came a long long way since the early days of r3d, and with a pretty OK one ($2K) you can edit things pretty fast (even with imacs - to some extend - you can edit 6k r3d). Now the Raven also support simultaneous ProRes recording, so you can do offline/online without penalizing the workflow, or just shoot prores (thought its only 2k, which is a big minus here).

I had my eyes on the Mini 4.6k since it's presentation (and with the footage now getting out I'm really impressed). With that said, my main reason of getting a camera next year is to rent out, and man people here don't know BlackMagic that much but they really know Red (producers and directors like it). So instead of starting some fist fights with my clients saying why they should go with the mini, for me it makes a business sense getting the Raven, paying the premium, but being able to charge a little extra and renting more without going "against the stream". And at the end of the day the Raven images will look great, just like the Mini 4.6k will look great.

Now something that might make it a little better than the current dragon is the fact that the Raven uses some of the new designs (internally and externally) found on the Weapon. Which gave the dragon sensor a little bit more of DR and a more clean images overall. I know its a though call, and I can't make the decision for you, but for me and my market here it makes more sense getting the Raven.

Just an addendum here, if I were going to use a camera only for personal projects I would save the extra money and be getting the mini.

Thanks for the response man it really add's to the decision

I see a huge huge huge red flag here....


...I would not buy either a RED Raven or URSA Mini for the purpose of renting out!!

Alexas and Epics rent out. Minis and Ravens are for owning and being part of a commercial day rate package...and maybe wet-renting out.

All other differences aside, the costs aren't even close. Consumables aside, Raven 4K Brain + Red/DSMC2 1920x1136 Touch Monitor + DSMC2 EVF W/Mount Kit + RedStation Mini Mag + Redvolt XL module + DSMC Base production pack (front/top handles, shoulder mount and rails) + cables for everything = almost $20,000. The equivalent setup for an URSA Mini 4.6K would be camera + shoulder mount kit (with top, front/side handle w/extension, rods, and shoulder mount) + CFast card reader + URSA viewfinder + AB/Gold mount plate = ~$7000.

This seems rather blown out of proportion. First, the $9800 Raven I/O package includes a 120GB SSD (which retails for $850 and not $1400), and includes a v-mount battery plate, so no need for the XL module. It also includes the 5" LCD (same size as the Mini, so I'm not sure why you're referencing the larger 7" as a requirement). The system is completely cable-less, so I'm also not sure why you're referring to having to buy cables. Plus the kit also includes a top handle and such.

Point being, you're looking at $10k Raven vs $6k for the Mini.

That being said, you're correct that kitting out the Raven for handheld is much pricier, but for many people (myself included), that's not something I need. The Mini is a much better value proposition either way, however for those in need of high quality high frame rates, light weight, modular, or for gimbal use, the Raven is worth the additional investment.

Thanks for the all the information and corrections!

This seems rather blown out of proportion.

Perhaps a bit, to be fair. I guess I was looking too closely at matching capabilities for both locked down/remote and handheld/shoulder mount when that's clearly not always to case for many folks.

First, the $9800 Raven I/O package includes a 120GB SSD (which retails for $850 and not $1400), and includes a v-mount battery plate, so no need for the XL module.

My mistake, I didn't take the basic I/O package into account as opposed to a la carte. Also I did indeed compare the $1400 240GB mini mag with the $550 256GB CFast, the 120GB before that was a typo.

It also includes the 5" LCD (same size as the Mini, so I'm not sure why you're referencing the larger 7" as a requirement).

I matched them based on resolution, not size, since the 5" RED is only 720p, not 1080p like the Mini or the 7" (technically the latter is a tad larger). So if you didn't need full 1080p or had another proper monitor anyway for that, then that drops $1500 from the RED's price.

The system is completely cable-less, so I'm also not sure why you're referring to having to buy cables. Plus the kit also includes a top handle and such.

My mistake, thanks for the correction.

Point being, you're looking at $10k Raven vs $6k for the Mini.

Point taken, given a non handheld/shoulder shooting configuration and a 720p screen—at least until you include media. At this point, RED Mini Mags are around 2.5x the cost per GB of CFast, and around 10-15x that of a good SSD (850 Pro). But again, as you mention, depending on what you are shooting (if you have time and/or a DIT, which seems more likely for non-handheld/shoulder productions) then I'd imagine this isn't so much of a concern.

Thanks for the all the information and corrections!

Given that I criticized your earlier post, I'm rather impressed by your incredibly humble and respectful response! Cheers!

As for the LCD, I hear you there. The added resolution of the Mini's LCD is a nice touch. Speaking of touch, is the Mini's a touch screen? Red's touch based menu system really appeals to me, especially after my recent usage of the FS7 (what a nightmare!).

As for the 120GB SSD, it may seem like it's a rip off compared to the cost of a 240GB Cfast, however when you take into account the usability and efficiency of Redcode, the capacities end up being about the same. That being said, Red's media prices still seem like absolute robbery, but the performance and reliability is hard to argue with, especially given the countless accounts of Cfast disasters on the big Ursa.

Anyway, to reiterate, both are going to be incredibly capable cameras. Sure, the price tags are different, but so are the usage scenarios.

Well hello again!

Given that I criticized your earlier post, I'm rather impressed by your incredibly humble and respectful response! Cheers!

Sometimes I surprise myself

Seriously though, I'm quite tickled pink that you'd say that; certainly has made my day. Though I can't say it had anything to do with the similarly gracious and reasoned tone and points made in your original reply--although you'll sometimes hear me trumpeting how I don't mind taking criticism, I certainly don't always live up to that.

Speaking of touch, is the Mini's a touch screen? Red's touch based menu system really appeals to me, especially after my recent usage of the FS7 (what a nightmare!).

Yes indeed, it most certainly is touch based, and quite a bit of praise has been foisted on its menu system's design in general (particularly relative to the Sony--then again, a rusty old tin can would look classy next to that.) And I can certainly say the same about my experiance with Sony cams like that, and if anything even worse, about my own Panny AC130A I just sold to help raise funds for my 4.6K.

As for the 120GB SSD, it may seem like it's a rip off compared to the cost of a 240GB Cfast, however when you take into account the usability and efficiency of Redcode, the capacities end up being about the same. That being said, Red's media prices still seem like absolute robbery, but the performance and reliability is hard to argue with, especially given the countless accounts of Cfast disasters on the big Ursa.

Redcode certainly has its perks, and it is indeed a very efficient compression medium, all things considering--more so than CDNG, certainly, and does help at least partially equalize the playing field in that respect. Though, even with the most efficient AVC, there is only so much even the best intraframe compression can do to preserve all the data present in the original--you can certainly debate whether R3D is at least 2.5x as efficient as CDNG, but not even HEVC can come close to the 10-15x efficiency ratio that would be required to match 850 Pro SSDs with a C-box. They may not be bulletproof, but they have a pretty sterling record, faster speeds than the Red, I can dock them with my hotswap SATA6 bay, and I can record onto two at once at double the frame rate so, if worst comes to worst, I have a frame by frame backup (more or less).

Anyway, to reiterate, both are going to be incredibly capable cameras. Sure, the price tags are different, but so are the usage scenarios.

For sure! That's something I think we sometimes can loose track of, that not only must one use the right tool for the job, but that we often do very different jobs.

In any case, not sure if you celebrate it, but happy thanksgiving!

What comparison, the BMC going to be cheaper, but apart from other factors mentioned here, consider the form factor difference, and the size and handability fully kitted out for different jobs. They went really wrong, in my opinion, after they dumped the Redone inclusive form factor for the modular system. If only we had a RedOne like camera with dragon like image today, hold it wait, we do, the Ursa mini.

As for the red rocket, it was suggested to top management as a stop gap measure for handling the original red footage. But when it was pointed out that GPU's were becoming capable of assisting processing instead, they appeared less lucid, denying it was viable. I know, that person that suggested these things to them was me. Since the cancelation of.the fixed, I am still p o. And p o about how long the aspertus has taken and they still about less than $10k when they need to talk 8kk res, or less than $3k for marketing. BM wins both battles, even though the mini is not setup competitively for eng or prosuner camcorder use. Now the question is what will we get in 4k under $2k next year, and what will BM bring. The "race to the bottom" might lead some high and dry.

countless accounts of Cfast disasters on the big Ursa.

Countless disasters? Now who's blowing things out of proportion?

countless accounts of Cfast disasters on the big Ursa.

Countless disasters? Now who's blowing things out of proportion?

Shots fired

While there seemed to be quite a few reports of this problem (though far from countless), they all seemed to occur with a specific version of the firmware (2.0), on a specific card make (SanDisk), on a specific camera (Ursa Major 4K 1.0) and if my understanding is correct, only on a specific model of that card (the original 128GB version). Besides, the Lexar once it came out is both less expensive and has has much higher read and write speeds, so there was little reason to use the SanDisk anyway. So its not really an issue anymore, to the extent of my knowledge. Which, as we all know, does not extend as far as I think it does haha

Countless disasters? Now who's blowing things out of proportion?

lol nice!

But to be fair, what one considers "countless" is subjective. I easily lost count after only two or three. :p

While there seemed to be quite a few reports of this problem (though far from countless), they all seemed to occur with a specific version of the firmware (2.0), on a specific card make (SanDisk), on a specific camera (Ursa Major 4K 1.0) and if my understanding is correct, only on a specific model of that card (the original 128GB version).

All jokes aside, over at the BMCU forum there was a thread just this week where the Ursa killed all 4 of a guys cards. Granted I haven't read past the first few posts in the thread, and hopefully it's indeed limited to exactly the scenario that you're describing. Either way, it's yet another case that I've noticed, and one that would concern me as a BM shooter on a major production.

Here's the thread in reference:

http://www.bmcuser.com/showthread.php?1 ... fast-Cards

That being said, I should probably clarify that although I have a Raven preorder, I'm also a proponent of BM cameras and equipment. My current production uses an ATEM 4k ME1, three Micro Studio Cameras, 6 hyperdecks, numerous BM converters, etc. I highly value the advances that Blackmagic has made, support their cameras, and do indeed think that the Mini 4.6K is a killer camera that will be perfect for many shooters. It's just not the camera for me.

Further, if I had to list the pros and cons of the Ursa Mini 4.6K, the list would probably be 95% pros, and the few cons would be very minor. For most shooters, thats one hell of a good ratio for a camera that costs $5k.

RED is more robust and industry known. I'm sure the image will be fantastic as well. And let's not forget the fantastic REDCode. That said, the Mini 4.6K will be a very big competitor, image-wise.

One thing that I found funny is that people that had pre-orders for the mini and were pumped for the ergonomics, all the sudden found it not for them when red announced the raven with similar box ergonomics.

I'm not talking about anyone in particular, I have just noticed this whole BM vs RED discussion silly.

Here's the thread in reference:
http://www.bmcuser.com/showthread.php?1 ... fast-Cards

Thanks! Hmm, well looking at that thread, we see my earlier conjecture was likely a bit hasty, as these involved new Lexar 256GB cards. However, the cards didn't actually fail; the problems were due to formatting in camera rather than a full format on a computer. Apparently the former is known to be unreliable, which really shouldn't be the case--after all, back in the day, I was told to *always* format in camera and not on a computer or the same thing could happen. However, the problem is avoided either with a firmware update, formatting on a computer, or (presumably) using good SSDs with the C-box and capturing to 2 at once.

Further, if I had to list the pros and cons of the Ursa Mini 4.6K, the list would probably be 95% pros, and the few cons would be very minor. For most shooters, thats one hell of a good ratio for a camera that costs $5k.

That's good to hear--we all have to make the right choices for our very different productions, of course. If you'd care to look over my post on the Ursa Mini feature request thread, if there are any cons you see that aren't included there, I'd be curious to know them for myself as well as for adding there. Thanks!

If you'd care to look over my post on the Ursa Mini feature request thread, if there are any cons you see that aren't included there, I'd be curious to know them for myself as well as for adding there. Thanks!

It's not so much that BM did anything wrong. Quite the contrary, as the Mini 4.6K looks to be a spectacular achievement!

For my needs however I like the Raven's 4K at 120fps, built in wifi control (remotely controlling the electronic focus on an ipad without focus gears/motors is amazing), Redcode efficiency (6:1 to 8:1 compression ratio is ideal for 99% of work), as well as the modularity (I'll be getting the 3rd party OMOD modules for wireless video and dual XLR with pro quality preamps).

Not exactly simple things for BM to add to the Mini in a quick firmware update. :p

But again, I may be a Raven supporter, but I'm very certain that I'm a minority, and with good reason. The Mini does 90% of what a Raven can do at nearly half the cost.

Ah, thanks for the insight. I think sometimes, as much as I say it myself, I loose track of the fact that no camera is perfect for every production and use case (at least at a reasonable price, that is). Nevertheless, I actually do have a section in that post where I mention some longer term features that would take one or more revisions of new hardware, and in fact I happened to include some of those exact things on the list: 120 fps, more efficient codecs for certain use cases, additional modularity, and believe it or not,

Some sort of wireless remote control/streaming addon? Could be useful to allow remote viewing/control via an iPad

So maybe someday in the future, though like you say most of that would take new hardware for sure. But the Ursa 4.6K, all things considered, is pretty darn close to a perfect product for a heck of a lot of needs, especially considering the price. Who knows what BM might come up with in a couple of years?

Nevertheless, I actually do have a section in that post where I mention some longer term features that would take one or more revisions of new hardware, and in fact I happened to include some of those exact things on the list: 120 fps, more efficient codecs for certain use cases, additional modularity, ...

Over at the RU forums, there's a lot of hostility towards the Mini, while over here on BM, there's equal hostility towards the Raven. When people invest a lot of money into something, they often become protective of their investment. Unfortunately they occasionally become irrationally defensive as well. :p

I like to think that I see both sides from a relatively objective standpoint, as I'm a BM advocate that just happens to feel that the Raven fits my needs a bit better. As I said earlier, I still own numerous BM products though that I employ for different use cases.

I think over here at the BM forum, one of the main criticisms over the Raven is the higher compression ratios of Redcode. I think much of this criticism is misplaced.

I suspect that it stems back to a few years ago when many of us were super pleased when the BMPC and BMCC first launched, and for the first time ever we had the opportunity to purchase a product that shot Raw internally, while maintaining an entry level price point. It was so innovative at the time, that a large user base quickly became addicted to shooting Raw, which was a massive improvement over the older AVCHD and such that we were previously accustomed. More recently we've seen BM update the firmware to more efficient 2:1, 3:1, and even 4:1 Raw on the Ursa line, which has been met with confusingly mixed feelings. Some people seem to feel they're losing quantifiable quality, which although mathematically is true, it's pretty much imperceptible to the human eye.

Which brings me back to Redcode for a second, where the perceived "sweet spot" is typically 6:1 to 10:1. That may seem rather surprising coming from BM cameras, but the reality is that most Hollywood films are filmed around 7.5:1. We're talking $150,000,000 movies, again, using 7.5:1 as the standard! The benefits are clear, as this ratio is less data than Pro Res, while maintaining the flexibility of Raw.

When put into perspective of the Raven, although the ratios can go as low as 3:1 for 24fps, it can actually do 4K 16:9 at 100fps in 10:1. Again, 10:1 is actually very good when you consider the Hollywood standard has been 7.5:1. Sure, the ratios get a little questionable at 120fps, but even then the new Dragon samples on the DSMC2 form factor are showing 12:1 to be quite clean.

But the Ursa 4.6K, all things considered, is pretty darn close to a perfect product for a heck of a lot of needs, especially considering the price. Who knows what BM might come up with in a couple of years?

Exactly. And when considering how far BM has come in such a short period of time, that's truly impressive.

The reality is that both the Raven and the Mini will be very capable cameras. People can criticize one or the other, but in the end, what one may perceive as a negative, another person may deem a positive. It's not a question of which camera is better anymore. It's which one best fits your needs.

It's not a question of which camera is better anymore. It's which one best fits your needs.

Excellent points as always, and the above does a pretty bang up job of summing up what I think both of us were trying to articulate. Perhaps one can say that was always the case, though now more than ever with how dramatically cameras' technical capabilities have improved while costs have come down such that its less of a matter of picking a camera with the fewest necessary compromises, but rather the one that does all of what you need it to for the forseeable future, without paying for what you don't.

Honestly, no matter what brand we've bought into (and I haven't actually plunked down the dough for my 4.6K yet), strong competition and filling other niches cannot fail to be a good thing for us. Blackmagic has surely helped contribute toward the industry offering more accessible options while raising data rates and expectations, and conversely as higher-end brands like RED come out with competitive features that the Ursas lack, hopefully it will help spur BM and others toward raising the bar further. It's a good amplifying feedback loop, for sure, and no matter what brand we use, we can be thankful for it. Defensiveness, as you say, can actually just be a sign of perhaps a bit of spec insecurity about our own camera system (and I've shown some of it in the past), but no matter the exact specs, we're probably better off working with what we have as long as we can create great images with it, such that we can better learn to maximize our own capabilities. Plus, we might be able to earn a little something to allow us to upgrade should the tech specs of our gear ever become the limiting factor that, assuming we have anything any of these companies have produced recently, it so rarely is.

It's fair to say you are paying a bit extra for the RED brand for what you get on paper, but that's true with anything, I think, and it undeniably does offer a few key things the mini doesn't, particularly in the 4K120 and modularity/expandability departments. REDCODE's advantages only partially nullify the much greater cost of on-camera storage, but for a large production where you have a few REDmags and a DIT to transfer everything to main storage, with a complex post workflow that would get unwieldy with less compressed raw, I can certainly see the advantages there.

In the end, it isn't really a question of the 4.6K "vs." the Raven, in the sense of picking one over the other. It really boils down to whether you need what the Raven offers, and which form factor/ergonomics works better for what you need and you probably already know which is true for you (as you certainly do). For the other folks who don't have these specific needs like you do then the 4.6K will almost certainly do the job, unless you just like RED products for whatever reason (and can afford them).

No question the Raven has some pluses in the comparison and I am not familiar with any direct comparisons in video quality of deliverables that were produced using various REDCode compression ratios. I'd like to see some complex scenes, not sky and sea, that illustrate 3:1, 6:1, 10:1, 13:1 for example shooting the same scene. If have seen the comparisons done on the BMPC4K shooting ProRes 422 HQ, 422, LT, and Proxy which were very instructional. The various ProRes codecs were actually very decent portraying colour but they did show degradation of detail. Of course it was easy to see the loss of details pixel peeping a still, but wouldn't be as evident in the moving picture.

Has any fair comparison been done without a lot of post processing to try and compensate for the detail loss? 3:1 and 4:1 CinemaDNG are considered visually lossless but there is some loss of information. I would think REDCode 3:1 would be able to make the same claim, but at what point does a complex scene show a degradation in detail? And by the time you consider using 13:1, it would be good to know what that is likely going to do to a complex scene before you commit to purchase the Raven. Maybe you already know, but might be good to share with us here.

Until I see otherwise, I would think that amount of compression is something I'd want to avoid by staying with BMD's approach.

The greatest benefit of the high compression ratios is recording times on media and storage costs. But these make the process of filming easier, the means of getting to the end. They don't make the product better, the end result. Maybe BMD is a bit cruel in wanting your raw to be an evident improvement over ProRes. Raw 3:1 and ProRes 444 XQ nay be close to equivalent even. RED seems to be taking the more practical tact showing mercy to decrease the storage requirements and increase recording durations, but it might also impact the usefulness of the data (for compositing, CGI, and VFX perhaps), to produce the best image possible. Nathan, would you agree?


Sent from my iPad using Tapatalk

RED Raven using the total image sensor at 120fps is 15:1 rather than 13:1

I'd like to see some complex scenes, not sky and sea, that illustrate 3:1, 6:1, 10:1, 13:1 for example shooting the same scene.

Not sure if it is exactly what you are looking for, but I found one in particular with a pretty complex scene (tree leaves waving in the wind, stopped down for maximum sharpness) on Reduser shot with the Epic in 4K; I can't link to the forum post directly but here is the link to (of all things) imgur. He claimed he checked the minimally compressed JPEGs and there was no difference to the original TIFFs produced right off the Redcode, ungraded and with no sharpening/detail applied. But something more formal would be interesting.

Has any fair comparison been done without a lot of post processing to try and compensate for the detail loss? 3:1 and 4:1 CinemaDNG are considered visually lossless but there is some loss of information.

I'd actually be interested to see the tests of these two vs. lossless, to see what visible differences, if any, arise in a complex scene after moderate to heavy grading. While my minimal budget and situation (episodic doc-style series with local cable and web distribution, maybe longer-form doc and other venues down the line) doesn't really call for lossless anyway, I do need to choose between 3:1 and 4:1 (leaning toward the latter) and would like to know what I'm missing, if anything.

I would think REDCode 3:1 would be able to make the same claim

In at least some places Red has claimed that 3:1 is in fact mathematically lossless, but I was under the impression that it was virtually impossible to get much above 2:1 losslessly no matter the compression technique, and I've seen spotty evidence for that claim. However, I'd certainly put stock in the fact that it is very likely visually lossless under a wide variety of conditions.

Raw 3:1 and ProRes 444 XQ nay be close to equivalent even.

At equivalent frame sizes, Lossless raw and XQ have nigh identical bitrates (judging from the 4K). Therefore, my question on the other thread as to why one would use the latter, given anyone working on a produciton needing and able to handle such bitrates would in all likelihood be prepared and able to take advantage of the benefits of raw.[/quote]

RED seems to be taking the more practical tact showing mercy to decrease the storage requirements and increase recording durations

Funny enough, Red's got nothing on Sony in that department. The F55 on down still offers 50 Mbps 8-bit MPEG-2 (XDCAM), which isn't far from AVCHD when it comes to quality given a similar compression ratio as XDCAM EX (which AVCHD at 24 Mbps equals or beats in all the comparisons I've seen). While I would never argue that the Ursa Mini should offer something like that, there was a time not long ago where that was actually considered "good" quality next to EX, AVCHD and piss-poor HDV.

You shoot the higher quality ProRes codecs rather than a higher quality raw when your workflow dictates which to use. For example, you are handing off your captured video to someone else for editing and they request ProRes or more likely refuse to accept raw. If you are in charge of the workflow, raw normally requires more work to handle with the raw debayer step and setting the parameters involved compared to native editing of ProRes.

Thanks for the tiny shots of the tree with featureless white sky, but a better subject might be a static but complex roof or wall of bricks or other detailed material so the image is more repeatable.

It's frankly absurd to claim complex images at 18:1 were indistinguishable from 3:1 if your test is intended to show differences. Obviously this individual has determined there is no discernible difference and then made statements to support that. JPEG images, at least the traditional version of them, are inherently horrid at the edges as colour and luminosity change. Not good to use that to judge anything at the level of the pixel peeking we are talking about. Yes, at a distance a JPEG can look good enough. TIFF can be as good as your source since each pixel is fully defined colour as long as the source is fully defined colour. 3:1 source isn't fully defined in a complex scene, but it certainly is hard to identify any loss of anything that matters. 18:1 a different story unless you are shooting the sky.

Mathematically lossless may get a 2:1 compression when we are talking data, perhaps even 3:1 for normal computer data, but you don't get anywhere near that with a complex scene's video data. BMD liked to say you could use a 1.5x compression value for estimates and do doubt that can be true, but a much safer estimate is 1.3x or 1.25x for very complex scenes. Somewhere CaptainHook did comment on this and at that time his sample video was a little less than 1.3x mathematically compressed.

When Raven is out in the wild, someone will do the tests from 3:1 to 15:1 in a fair and accurate manner where the scene is complex. I hope it's not their last post on RED user (joking), but people will be curious. I'm confident 15:1 will be usable in most shooting and viewing situations so I don't want to come across as saying 15:1 is absurd marketing hype and should never be used. The saving grace of motion pictures is that they are moving and you don't notice what you can't see usually except when doing these kind of static comparison tests.

But some people, including Grant Petty, don't want to use 15:1 raw because they don't want you losing detail in that way, you can be confident the detail is in the BMD raw data. If your situation requires significant video compression you have the complete palette of ProRes codecs at your disposal, but please, shoot CinemaDNG raw or raw 3:1 or raw 4:1 when you require the most detail to be captured and the greatest flexibility in post. REDCode is different and I'll leave it to someone shooting with it to determine its optimum range, but I think I'd heard even 7:1 or 8:1 could be very good.

Sent from my iPad using Tapatalk

Thanks for the tiny shots of the tree with featureless white sky, but a better subject might be a static but complex roof or wall of bricks or other detailed material so the image is more repeatable.

Yeah, it's not so useful (to be fair, it is a crop of the original image; you could compress something with DV and it would look no different if that was the full original side) but it was what I happened to find after a bit of searching so I figured I'd leave it there. As the poster mentioned, a non overcast day with more contrast and wider DR would likely reveal more pronounced differences, as would heavier (but consistent) grading, even more detail in the shot, etc. That's the reason I mentioned something more formal (and thus more valid) would be interesting, and I didn't intend to claim the above was any evidence of Redcode holding up at high ratios.

It's frankly absurd to claim complex images at 18:1 were indistinguishable from 3:1 if your test is intended to show differences.

I didn't mean to say the dude claimed there were no differences between 3:1 and 18:1, though would indeed be very much absurd and indicative of a very poor testing procedure as you mention. What he was claiming is that there was minimal difference between the rendered TIFFs and the rendered JPEGs (when not modified obviously) which can, in fact be true for most scenes according to a number of tests I've seen, if the JPEG is lightly enough compressed and is first-generation. However, he might have been referring to other JPEGs uploaded to the forum that weren't visible to me, rather than the JPEGs I linked which were likely recompressed on "imgur." He had links to the actual TIFFs and R3Ds, but they were old and dead. He himself claimed there was a significant visible difference between the various R3D flavors, and the majority of claims I've seen on both RU and BMCuser (esp the latter) maintain that 3:1 is the only level of R3D truly visually indistinguishable from lossless, below 5:1 or below was were pretty meaningful degredation became apparent, though by no means the end of the world.

Mathematically lossless may get a 2:1 compression when we are talking data, perhaps even 3:1 for normal computer data, but you don't get anywhere near that with a complex scene's video data.

Right, which is why I found RED's claim of (mathematical) losslessness for their "3:1" compression quite unbelievable, given even the most "typical" data can be compressed is around 2:1, and I've never seen any lossless video codec do better than that on a real scene. So the claim, if it really was made, likely referred to visual losslessness, which, at least as most people seem to define it, is about met as you say.

When Raven is out in the wild, someone will do the tests from 3:1 to 15:1 in a fair and accurate manner where the scene is complex.

You and I both are looking forward to that, and hopefully they can compare BM CDNG at lossless, 3:1 and 4:1 alongside that. Maybe Mr. DuMoulin with his forthcoming Raven and one of his BM cameras? One can only hope.

RED seems to be taking the more practical tact showing mercy to decrease the storage requirements and increase recording durations, but it might also impact the usefulness of the data (for compositing, CGI, and VFX perhaps), to produce the best image possible. Nathan, would you agree?

In terms of Redcode, 6:1 to 8:1 is the "sweet spot", although I've heard 10:1 is great and many people shoot at 5:1 for additional confidence. The beauty of it is the ability to change the ratio depending on the complexity of the scene. Water or trees? Sure, 5:1. Static talking head? 10:1 sounds good.

Speaking of which, I have yet to hear of anyone (outside of VFX) who shoots at 3:1 Redcode. It just doesn't happen, as I think it's mostly a question of diminishing returns.

Take 3:1 and 6:1 for example. The data rate is 100% more for 3:1, yet the perceivable difference is maybe 5% (random number for the sake of argument). Now downscale that for your 1080p and you have yourself a 1.25% perceivable difference. Then transcode the finished product to ProRes HQ for your final deliverable, and your 1.25% improvement is now 1% at the absolute most.

So yes, you're right when you say that you've maximized the quality, and produced the best image possible. The difference is quantifiable, albeit small.

However consider the post process, while working on a project that has a typical turn around time (anything with a limited number of hours). The 3:1 files that are twice as big are going to slow down your editing, effects, and grading process significantly. Be it from having to generate proxies, longer transfer speeds, stuttering playback, difficulties in sending shots to VFX, etc, you will waste a significant amount of your time frame using 3:1 compared to having simply used 6:1.

So in the end, sure, you got an image that was marginally nicer at 1080p, but how much inferior is your end product after you spent twice as much hassle dealing with the 3:1 files? That 1% increase in picture quality cost you in every other department of post production, rendering the actual product much inferior.

It's so easy to get caught up in the technical. Yes, you're absolutely right that lower compression yields a better image. But at what expense? What we do here is predominantly a creative process, often simply outlined by the technical. It's great to be as technically sound as possible, but not when it comes at the expense of the creative process. 6:1 or 8:1 ratios are good enough for $150,000,000 blockbusters with nearly unlimited resources. Sure they're good enough for me, especially knowing that I can put those smaller files sizes to use by maximizing my post workflow, and focusing on what matters most.

Let's talk about compression being lossy or lossless. This is the established terminology.

None of this visually lossless mathematically lossless mumbo jumbo marketing speak.

Redcode is lossy no matter the compression rate.

cDNG is lossy at 3:1

CDNG is otherwise lossless as either RAW or it's more mild iterations.

JB

Let's talk about compression being lossy or lossless. This is the established terminology.

Agreed, though I admit to at times being an offender in that department (though mostly to refer to said marketing speak). After all, confusion on this point is why Red seems to have duped at least some people into thinking 3:1 Redcode is somehow lossless, in the technical (accurate) sense of the word.

After all, confusion on this point is why Red seems to have duped at least some people into thinking 3:1 Redcode is somehow lossless, in the technical (accurate) sense of the word.

Red doesn't claim that 3:1 is lossless. It clearly states that on their Redcode page. They do however state that 2:1 is lossless, and 5:1 is visually lossless. Whether it's a false claim or not is irrelevant though. We're analyzing the value of the Mini vs the Raven, and not the business practices of either company (which would get out of hand from both sides). More specifically, I'm addressing the value of the ability to use varying lossy compression ratios, above what BM offers. To me, this is invaluable and a very appealing feature of the Raven.

We can talk all day about what Ratio is the best or most useful, but the fact that it's up for debate is exactly my point. With Raven, I have the option and flexibility to choose as I please. With BM, I don't.

The flexibility of Redcode is a major advantage in post, and something that draws me to Red. Not everyone will agree, but that's exactly the beauty of having multiple platforms to choose from.

PS. This continued obsession with the technical is exhausting. The notion that projects should utilize ratios superior to what Hollywood uses seems so ridiculous to me. People can continue to shoot with giant file sizes for an extremely marginal gain in picture quality and ultimately make their post workflow an absolute nightmare, while I go and get things done. :p

(I'm just teasing so please take that into consideration before flaming me lol)

Nathan, well put, and I will continue to use HD camera and ProRes, and even, once in a while AVCHD 2, may the gods forgive me!

(I'm just teasing so please take that into consideration before flaming me lol)

FLAME ON!
Just kidding haha. But I do have (actual) OCD which at least recently has been going full tilt into this sort of thing. Maybe I was born to be a DIT lol. I'm sorry that I keep obsessing about it; I'm trying to break the habit by not visiting the forum so often and replying at such length and worrying more about more pressing and important matters. Like getting sleep.

Nathan, well put, and I will continue to use HD camera and ProRes, and even, once in a while AVCHD 2, may the gods forgive me!

Hear, here--that was my bread and butter up until recently for me with 14-bit MLV raw, higher bit rate H.264 (~150 Mbit/s long-GOP for HD with log gamma using Magic Latern) and hopefully soon 4:1 12-bit log raw with the 4.6K, and I kept my AC130A (AVCCAM not even 2.0) around up until now. Its kinda funny though that on one hand I'm asking probing questions about pushing the limit with bit depth, compression and raw and yet also advancing the notion that it would be helpful for a large chunk of people, including myself in some cases, to include more efficient compressed options on the Ursa Mini given its at least nominal focus on capturing some of the doco/run and gun market.

But I guess it boils down to different strokes for different folks--why not offer those options for those who need them, higher quality for high end cinema (if it in fact helps), and more importantly better efficiency for docu, run and gun, and maybe even limited event and broadcast work (obviously you wouldn't buy the camera for that, but might need it sometime). Of course, there would be at least nominal liscnesing feeds and probably the need for a hardware encoder chip, but I'm more than willing to pay a little extra cash for a lot of extra flexibility, even though most of the time I'll end up in the middle shooting what it already has (and certainly no higher than 3:1 12-bit log myself or below 100 mbit/s, and even that low only with a high efficiency codec).

The flexibility of Redcode is a major advantage in post, and something that draws me to Red. Not everyone will agree, but that's exactly the beauty of having multiple platforms to choose from.

I actually think REDCODE's best feature isn't the compression, but the way they handle metadata though from camera to post.

REDCODE *was* great at first because uncompressed was simply unwieldy. Well, that was 5 years ago and now uncompressed isn't so unmanageable.

But what i like about REDCODE is the way that it manages it's metadata. That once you make a change in camera or in post, that metadata stays with it and I wish cDNG did the same thing.

JB

John we are patiently/impatiently waiting on your footage. When you post that, i think the red raven vs Ursa mini thread will go in favor of the Ursa mini. Even for those considering getting a raven based on specs.

John we are patiently/impatiently waiting on your footage. When you post that, i think the red raven vs Ursa mini thread will go in favor of the Ursa mini. Even for those considering getting a raven based on specs.

This is probably my fault, as half this thread is me arguing extensively regarding the merits of Raven, as that's what I ended up pre ordering. lol Conversely though, I could definitely ramble equally about the aspects of Mini that trump the Raven.

- Cost!
- Ergonomics for run & gun (although the Raven wins when it comes to gimbal and studio work)
- True 4.6K, whereas the Raven's 4.5K is still only 2160p (not useful when shooting 16:9)
- Larger sensor size (the biggest factor for me, as 16:9 on Raven is nearly a 2x crop)
- Cheaper media (although I trust MiniMags over Cfast any day)

I was about to add XLR inputs, but a 3rd party manufacturer is currently producing a very high quality I/O+audio module that replaces Red's I/O module.

I've said it before and I'll say it again. It's really not a question of which camera is better. It's just whichever best suits your shooting needs. They will both be incredible pieces of gear in the right hands.

I've said it before and I'll say it again. It's really not a question of which camera is better. It's just whichever best suits your shooting needs. They will both be incredible pieces of gear in the right hands.

I agree. We need to replace the "VS" with "OR" and focus on the job at hand.

This argument is kind of like saying, "boots vs shoes?" My reply would be "What are you about to go do?"

Good one Nathan, But here on this thread we want everyone to go for the Ursa mini. Kidding of course and you are absolutely right. Different cameras for different needs as per the user.

This is probably my fault, as half this thread is me arguing extensively regarding the merits of Raven, as that's what I ended up pre ordering. lol Conversely though, I could definitely ramble equally about the aspects of Mini that trump the Raven.

In your defense, pretty much in every post you stressed how those merits were highly dependent on one's shooting situation, and gave the Mini due credit for its advantages as well. If anything I derailed the thread more than anyone with my rambling about specs and whatnot.

By the way, are you the same Nathan DuMoulin with Runout Media who shot Venom Trickshots? I happened to be re-(re-)watching one of those videos and noticed some people in the comments complimenting the camera work, and upon seeing your name mentioned I got the chills, and wondered if it could be true that I was speaking with the man himself.

I agree. We need to replace the "VS" with "OR" and focus on the job at hand. This argument is kind of like saying, "boots vs shoes?" My reply would be "What are you about to go do?"

Absolutely. Like Mr. DuMoulin said, its not a matter of which camera is better, but rather which is the best fit for what you shoot (and your budget).

By the way, are you the same Nathan DuMoulin with Runout Media who shot Venom Trickshots? I

Yeah, that's me. Florian (Venom) is a friend of mine, and I've made a few low budget videos for him. The first was actually with an HDV tape based camera where the max res was like 1440i lol, and the sequel was with a Canon t3i. Super high end stuff. :p

(probably the most frustrating thing to film, as when shots take 20-50 tries, the framing, focusing, or jib movements are always sloppy by the time he finally makes it)

It's crazy to look back at how fast things have changed, and how drastically inferior the tech was just a few short years ago. It really puts this thread into perspective, as we're sitting here arguing over 3:1 vs 6:1 Raw and such, when 3-4 years ago some of us were still using 2/3" camcorders, HDV cams, or entry grade SLRs. Too funny.

If I only knew then what I know now, things would have been a lot different. Looking back, it's borderline embarrassing.

Do you think we shall be embarrassed five years from now about today's state of the art? Likely in ten years!


Sent from my iPad using Tapatalk

In one year we will be embarassed.

Guess it's time to make another demo reel...

I'm already embarrassed haha...

Wow, whaddaya know—I've met someone famous xD! Low budget or high, 1440i HDV or 4.6Kp RAW, those videos and especially the camera moves still look pretty darn slick, and I'm not the only one who thinks that if you look at the comments. Though I figured it took quite a few tries to get such impossible looking shots to connect, honestly 20-50 tries was on the low end of what I though it might take for some of them. Still, if only I could take shots like that, whether the video or billiards kind...one can dream, haha.

Hey, even I was up until very a few months ago I was still (infrequently) shooting on my 3x1/3" AC130A in glorious 8-bit 4:2:0 24 Mbit/s 1080p30 ACCAM (the bigger brother of the HMC150 you are posing with on the Runout Media website), and I still shoot similar (with the bitrate overclocked to 100 Mbit/s) on my Canon DSLRs when practicality is not in favor of MLV raw. Heck, it has been less than three years since the best we had in school was the DVX100B--loved that cam. Of course, the biggest difference besides what's in front of the camera for me and you/everyone is the skills and creativity (or lack thereof, in my case) of the person behind it. And it is mostly due to that, not the gear, that I'm rather embarrassed to release anything even recent that I've done for various purposes, given how much I myself have (hopefully) changed, gear notwithstanding.

The really mind blowing thing to me, reading some old video magazines online, is how far we've come since right before the introduction of DV in 1995. 20 years later, its incredible to think that, at least in terms of technical quality, Joe Schmo can shoot 1080p60 or "4K" video on his new iPhone's $20 camera module smaller than a fingernail that can surpass even what a massive, highest end $100,000+ Beta SP broadcast rig could do in almost every respect, and easily do fairly complex editing right on the device, not to mention on the computer, that would take a whole edit suite or top of the line custom computer to do back then. Does that mean Joe Schmo "will" produce better footage than a broadcast pro or trained DoP with 1995 technology? Heck no, but it does mean that, from the iPhone up to the Ursa mini, we are in an age where the gear we can afford matters less and less, while with an ever-rising skillcap given the greater abilities, operator ability and everything else that goes into a production matters more and more.

I recently read that an emoji character was the 'word' of the year. That's ridiculous. It's obvious to me that the word of the year should be 'viral' and I very rarely watch any clip that is described as viral, but when I do, the clip has always been of the poorest quality imaginable. People seem more comfortable with garbage. Sort of like Blair Witch but clips of thirty seconds or less. I did watch Blair Witch from beginning to end and hated every frame and con and concept and manipulation.


Sent from my iPad using Tapatalk

Wow, whaddaya know—I've met someone famous xD!

Hardly!

Coincidentally I actually used five Ac160s after I sold my two Hmc150s. Great cameras at the time. I actually miss those days a bit (even the newest content on the Runout site is nearly 5 years old now). I've since been working instead as a producer in live production gigs (ESPN3 and such), so the skill set and tools are quite different. I still film on occasion though, hence my Raven purchase.

What's amusing about this little tangent were on, is that back then anything sub $10K produced a really mediocre picture. The upgrades between each new camera generation were drastic, the improvements largely quantifiable, and each increment was well worth obsessing over. I think that finally, cameras like the Mini 4.6K and Raven have allowed us to reach a level where it no longer makes sense to over analyze the feature set. Each camera, although different, offers nearly unrestricted professional quality. We've hit a plateau where the consumer will no longer be able to tell the difference, and the quality of the end product will solely be based on operator skill, and the content which is being recorded.

Once the Mini and Raven ship, it's really just a matter of putting these new tools to good use, expanding our skillsets, and enabling opportunities for good content.