Recommended Media for BRAW - BMPCC4K

Hello,

Do we still need the CFast 2.0 memory card to be able to record BRAW 3:1, Q0 and Q5 at 30/60fps?

Or can we record BRAW on the lower cost memory cards?

Thank you!

I can confirm, though take this with a grain of salt, that my $12 80Mb/s READ 10MB/s WRITE SD card recorded 4K30 BRAW 12:1. I am utterly blown away that it was able to do so. Recorded 2 takes for 30 minutes each on a 64GB card. I dont know if there is any issue with it randomly working sometimes and not sometimes. I also thought the write speed needed for 12:1 would be greater than 10mb/s. Or maybe my SD card is writing much faster and I am killing it way sooner, but this is a 4 year old SD card that I have used many times for other purposes. Your mileage may vary.

I am hoping to see if the cheap $50 256GB 90Mb/s WRITE cards can handle 4K60 and hopefully at Q5, though I doubt it. Even 4K30 which is what I typically shoot for 1.5 hours at a time, at Q5 or 8:1 would be so freaking impressive!

Testing my T5 512GB via USBC right now, indicates at 4K60 DCI, Q0, I get between 30 and 60 minutes of record time. I am up to 5 minutes so far and no dropped frames.

Curious if this level of activity causes any heating issues with the processor in the camera. Hopefully not!

I was able to record 12:1 4K DCI raw in 24p continuously on my SanDisk Extreme Pro 95 MB/s 128GB card. When switching to 60p it dropped frames after about 5-10 seconds but that didn't surprise me. Still good enough for quick slow motion clips though! Also didn't notice any dropped frames in 120p 12:1 raw (HD obviously) which was a nice surprise.

I got 4K30DCI in 12:1 on a SanDisk Pixtor.. not sure what else it is called, but it is 80Mb/s read and 10mb/s write.. only 10.. not even the 70mb/s write that the SanDisk Extreme claims. I dont know how much difference 10mb/s to 70mb/s write speeds makes for 4K30 to 4K60.. I would guess that it is more than 2x faster at writing continuously, but I have seen a lot of disk speed tests where things start out slow, then get faster, then slow down again.. not sure if SD cards have varying speeds depending on duration, etc. That said, I would hope that the Extreme could handle 4K60 or maybe 4K30 at 8:1 or even Q5?

I was able to record 12:1 4K DCI raw in 24p continuously on my SanDisk Extreme Pro 95 MB/s 128GB card. When switching to 60p it dropped frames after about 5-10 seconds but that didn't surprise me. Still good enough for quick slow motion clips though! Also didn't notice any dropped frames in 120p 12:1 raw (HD obviously) which was a nice surprise.

I also have a SanDisk Extreme Pro 95 MB/s , but a 256GB version. I was able to record a 5 minute clip in 12:1 4K DCI raw in 60p without dropping frames! I don't know how reliable this is, but I was pleasantly surprised!

Thanks for the replies!

I got the camera last week and I bought 2 angelbird AV PRO CF 256GB cards to be able to shoot ProResHQ 60fps. I shoot stock footage and I don't record more than 30 seconds at the time. I'm trying to I'm trying to figure out if I should return the angelbird AV PRO CF 256GB cards and get few Lexar Professional 1000 cards.

I'm very happy with the look of PreRes HQ, and I usually don't upgrade the firmware of my equipment.

Please try UltraHD 60p Q0/ Q5 on the lower cost SD cards!

Thank you!

I tried Q5 with my SD Extreme Pro (128 GB) and could record 4K25p without problems, too (no drops after 1+ minutes).

OK.. so this is surprising to me.. I took my cheap $7 32GB microSD card with SD adapter. I am recording 4KDCI 60p on it right now, 8 minutes in, using Q5. What? That is nuts to me! Oddly though, the 12:1 barely goes 3 seconds before it stops. I thought 12:1 and Q5 were pretty similar?

Also, I am recording with a lens cap on most of it. I moved camera around fast/slow, zoom in/out, etc.. still recording.

So color me impressed and baffled in a good way.

I still would MUCH rather rely on more rugged SD media than cheap ones. I am assuming the likes of the ToughCard and others that cost like $200 for 128GB are built to handle a lot more rugged situations, like cold/hot weather, and continuous recording? I assume the memory used in those is much more durable and long lasting than the cheap cards.

But still.. $7.99 for a 32GB microSD that records 4K60 BRAW at Q5.. holy crap!

What are you recording? A steady image or something complex? Since the lowest bitrate in Q5 is under 12:1 it can have something to do with what you are filming. Try shaking the camera or moving it fast.

Right..static.. so I expected that.. but even so..that such a cheap SD card can record that level of quality.. is nuts. I did shake it around for about a minute. no loss frames. Will do more testing when the forking rain stops in our area of the world and I can get outside a bit with it.

You got that wrong!
Shaking or moving doesn't matter, Braw is I-frame only.
Motion blur might even make it easier because of the loss of detail.

High contrast detail matters, so try full-screen leaves against a clear sky or pebbles on a beach, all in focus.

So question.. all these SD card formats with varying degrees of read/write speeds. I believe SDXC is the card to get, though I did use SDHC card for my 4K30 12:1 test and that worked.

There are a LOT of options, from 633x speed, to UHSI or II, with some saying 170MB/s but that is read speed, to the newer V30/60/90 ratings for video.

Would a V30 rating, which supposedly means 300Mb/s sustained writes... is that fast enough for a 4KDCI 60 Q5? What about 12:1, 8:1, etc?
What about V60 and V90 variants? Basically you can find for around $75 a 256GB V30 SD card. I would love to be able to go that low in price/size, if possible. I have the T5 500GB SSD for those times when I need more space for higher quality stuff. But for day to day sports video that is mostly to turn in to a youtube video after editing, I want to maximize space with still good quality. I also dont want to have to replace the SD card after 5 recordings.. so hopefully these SD cards have a decent life span.. e.g. hopefully you can use it a few hundred times before you have to replace it.

I was able to record 12:1 4K DCI raw in 24p continuously on my SanDisk Extreme Pro 95 MB/s 128GB card. When switching to 60p it dropped frames after about 5-10 seconds but that didn't surprise me.

Do you format your cards in camera ? I always do.

If you format on the PC, it could assign a storage block size that isn't optimal for the camera, which in turn could reduce write performance, causing your card to drop frames at 60p?

Always format in camera. Always exFAT.

Justin V30 means a guaranteed sustained 30 MB/s write I believe.


Sent from my iPhone using Tapatalk

But which ones.. will a V30 work well enough to do 4K60 in 8:1? Like.. how do you know which one to get for which quality setting? The V30 ones can be had for about $75 for 256GB. But if you need V60 or V90, now we are talking a lot more money. Will UHS-I work, or you need UHS-II be needed?

For example, can this one (https://www.bhphotovideo.com/c/product/ ... A514&smp=Y) support 4K60 at 8:1 or 3:1? Or Q0? Or will that at best support 12:1 at 4K30?

More so, there is the Extreme Pro that is ALSO V30, that is about $20 more for 256GB. I mean if the $57 256GB V30 Extreme can support Q5 at 4K60 (which I have to assume it can since my much cheaper 10MB/s card apparently can), I would hope it could support 4K60 at 12:1 or even 8:1?

Also, 12:1 is better quality than Q5? Because Q5 seems to record at 4KDCI60 on my cheap 10MB/s card continuously, but 12:1 stopped after 3 seconds. It did record 4K30 though continuously.

You got that wrong!
Shaking or moving doesn't matter, Braw is I-frame only.
Motion blur might even make it easier because of the loss of detail.

High contrast detail matters, so try full-screen leaves against a clear sky or pebbles on a beach, all in focus.

Oh, ok. Didn't know that, makes sense. Thanks.

Good catch, Uli.


Sent from my iPhone using Tapatalk

Hey Justin,

With my Extreme Pro 95 V30 256GB, recording 4K DCI 60fps, is a no go with Q5. During my test, it dropped frames after 3 mins, 38 secs.. This was shooting indoors and occasionally pointing out a sunny window.

Also 8:1 fails as well after 10 seconds.

12:1 seems OK, and I recorded over 6 mins before I stopped it.

Is there any reason for microSD card with adapter to be slower than normal size SD card? They have exactely same parameters and name, both are R/W 170/90MB/s. In adapter is no electronic, it only connect contacts. But maybe smaller is on different technology and can be slower in reality. What do you think? I ask because microSD is 15% cheaper and I have reader for them.

JMY.. crazy... I am using a very cheap PNY 32GB Elite Performance. I set it to BRAW Q5, 60fps, 4K DCI. I dont know if I did something wrong but at first I set it to 60fps, and it was recording for 15 minutes.. when I Stopped it showed 23.94fps. I dont know why/how that happened. Anyway, I reformatted SD card in camera, tried again. First two times, it stopped after a couple seconds. Then, and I had to triple check, it recorded for over 7 minutes before I finally stopped it. I was walking around house, moving camera constantly.

So I am a little confused.. does the Q5 setting mean it can vary the amount of processing per frame based on complexity of image it is recording? Or does movement cause it to need more processing power?

I dont know the write speed of this SD card. It says 95MB/s read speed, but I cant find anything on write speed. It is a Class 10 UHS-I U1 card. Super cheap too.

Regardless.. it is crazy to think you can record this quality on this card. I record long durations usually but I assume when shooting short shots for a film/commercial/etc they arent typically very long most of the time. So that you could use Q5 or 12:1 and record a minute or two at a time at 4KDCI 60 is just absurd to me! Or rather, I am very very happy that this is now possible, and I suspect BM is the only manufacturer that is able to make this happen due to BRAW.

Now they need more cameras that use it!

Troska, often Devices with smaller capacity will not perform as well as larger devices. So a 32GB device will fail whereas a 128GB will pass.

The “up to 95MB/s” claims are completely useless in evaluating the write speed and even if you found a published write speed, it’s not dependable. If it has a badge like V60 or V90, that means it can sustain writing video up to 60MB/s or 90MB/s. V30 might work at 12:1 as it can be more capable than 30MB/s. It could be actually 50MB/s but I think there is no industry badge V50 so they print V30.

The BMPCC takes UHS-I cards. The BMPCX4K supports UHS-II cards which are much faster as they have two rows of contacts transferring data. You can use an UHS-I card in a UHS-II device, but you get a much slower data transfer.

However...
If you have some, can you try and let us know? If you don’t have a few already, I’m not sure if I’d be buying very fragile microSD cards instead of an already fragile SD card and risking a problem with precious footage.

Were not these cards originally designed to be inserted in devices like a stills camera once and forgotten about as you transfer photos to a computer via a cable? When used for video and you’re pulling them and inserting them perhaps hundreds of times in a year, I’d be worried they will crack or deform or break.

That durability is one advantage of the faster CFast 2 cards but the cost is significantly higher. When I’m shooting with the BMPCC4K, my CFast 2 cards will be my principal cards and my SDXC card will be there if I need the space for overflow as long as my codec doesn’t drop frames. I usually shoot 30 fps and I use a fast Wise 128GB SDXC so it will be good.

It’s possible I’ll only push and pull on the regular SDXC card a few times a year.

Justin, the variable bitrate of Q5 does vary with the complexity of the frame. As ULI pointed out, if you have a lot of edges in frame as you might find waking around the home and picking up details like upholstery patterns or plants or dishes and so on.

It memory serves me, Q5 can record as much as 74MB/s (7:1) shooting a leafy tree at midday or as little as 27MB/s (20:1) watching paint dry.

When BMD tests cards for the various codecs, they test them shooting a complex scene so people may report that a card worked fine (watching paint dry on a wall or clouds in a blue sky) while another person says that same card failed them (as they shot a blooms on a cherry tree).

Many people will just shoot ‘fixed bitrate’ rather than ‘constant quality’ so they have reliable results. It will take some time to learn what we can do with Q0 and Q5, but it think it’s a concept worth pursuing. Since I usually shoot 30 fps, I should be fine, but at 60 fps and higher the risk of failure is there.

Sent from my iPhone using Tapatalk

Rick. Good info to know. So Q5 could achieve better than 8:1 quality but potentially much lower than 12:1 as well.
Like others, if I were shooting green screen VFX shots, or complex moving/detail shots, Q0 seems like the way to go. For my sports stuff, I believe Q5 should be fine, want to maximize storage and little less worried about the overall frame to frame quality.

As for storage, I agree with you 100%. I definitely want to invest in some durable CFAST cards eventually. I just dont like the ridiculous cost of them right now. Question, are they durable because the material used inside of them ensures they can withstand the day to day rugged use of the cards (e.g. not only in camera high speed recording, ambient temperatures, etc, but being put in/out of camera/card reader possibly several times a day for month after month)? Or is it more about the housing..e.g. water resistant, more durable to drops and such?

As you said, my main concern with cheap SD cards is even if they record.. when I pull it out and put it in my reader... will it still be good or can it get corrupted or crap out? I want that durability assurance of the video more than anything else. So like you ideally I want a large CFAST card, and a good UHS-II SD card (assuming it is similar durable like CFAST is?).

I think with BRAW, being able to have one of each, both at 256GB in size, should allow for most of anything I would record in a day. With the 500GB SSD as a backup, I cant imagine needing much more even shooting at 3:1 or Q0.

However...
If you have some, can you try and let us know? If you don’t have a few already, I’m not sure if I’d be buying very fragile microSD cards instead of an already fragile SD card and risking a problem with precious footage.

I have Samsung SSD for "big recording" i want cheap SD only for some traveling and testing arround city.
I know SanDisk Extreme Pro 128GB with 170MB/s read is best solution at this price range. But they make exact same card in two sizes. But I don't know if micro version could have less IOPs performance or something like that, or if they have just different plastic outside and inside are same with same real speeds. If microSD will perform same as a same big version I will buy micro.

Perhaps you can look for IOPs on Tom’s Hardware? Sites like that are getting to be more video aware and skeptical of manufacturers’ claims. BMD writes in 128KB blocks so that’s what you want to look for on a hardware evaluation site. The vendor’s claims may be true for a single block size, but that size won’t be 128KB. Could be 4K or 1MB.


Sent from my iPhone using Tapatalk

Perhaps you can look for IOPs on Tom’s Hardware? Sites like that are getting to be more video aware and skeptical of manufacturers’ claims. BMD writes in 128KB blocks so that’s what you want to look for on a hardware evaluation site. The vendor’s claims may be true for a single block size, but that size won’t be 128KB. Could be 4K or 1MB.

Ok thanks, I didn't know about size of blocks...

btw I found now test with blackmagic speed test and 170MB/s version can write 10 FPS of DNG or Prores HQ at 25FPS.. I think official BM testt will use correct block sizes, so card look pretty usable for Sunday shoting.

Troska, the AJA disk speed test is also very good as it can display the history of the data rates. When a card has no problem writing data, it traces a fairly plain line at the expected data rate. When it struggles, the data rate trace looks more like a mountain range than a flat plain.


Sent from my iPhone using Tapatalk

If you aim it at the same scene for a few seconds, the camera will give you a quite decent estimation of the remaining time on the card. This way you can calculate which data rate Q0 or Q5 is using in that situation.

Good catch, Uli.

In practice, moving the camera with marginal media/excessive data rates does appear to force frame drops. This happens reliably, again and again. It may not make sense, but it's so.

Okay, John. I had found that with very early testing of the URSA Mini when I was investigating the Lossless compression of CinemaDNG. I could get a variety of Lossless ratios from around 1.2x to 1.3x and my recollection of those tests was that camera motion could give me slightly larger DNG files sizes. I haven’t given it much thought the last couple of years.

Now that BRAW will be one file to the operating system and doesn’t have a mathematically lossless compression, it isn’t as easy to determine, but Uli is also correct that the amount of ‘detail’ in an image is a driver of the Q0/Q5 bitrate.


Sent from my iPhone using Tapatalk

Lexar Professional 2000x UHS-II SDXC SD CARD.

Can confirm that the Blackmagic raw has a continuous shooting of up to 60 frames in all format.
Blackmagic raw Q0 60 fps
Blackmagic raw Q5 60 fps
Blackmagic raw 3.1 60 fps
Blackmagic raw 5.1 60 fps
Blackmagic raw 8.1 60 fps
Blackmagic raw 12.1 60 fps

Thanks, pulalis. Do you know the file size and duration of the Q0 test? I’m curious about the bitrate achieved.


Sent from my iPhone using Tapatalk

[quote="rick.lang"]Thanks, pulalis. Do you know the file size and duration of the Q0 test? I’m curious about the bitrate achieved.


Welcome Rick.

Blackmagic raw Q0 4k 60 fps test. 23 minutes. 118 gb

Hi, is BMPCC4K supports SanDisk 512GB Extreme PRO CFast 2.0 (SDCFSP-512G-A46D) card? I see 256GB card version on recommendation list, but not 512 GB.

Samsung EVO Plus 128GB 100 MB/s microSDXC USD $29

Blackmagic raw 12.1 60 fps
8.1 25 fps
5.1 25 fps
3.1 25 fps
Q0 25 fps
Q5 25 fps

I can confirm that it works.

I was able to record 12:1 4K DCI raw in 24p continuously on my SanDisk Extreme Pro 95 MB/s 128GB card. When switching to 60p it dropped frames after about 5-10 seconds but that didn't surprise me.

Do you format your cards in camera ? I always do.

If you format on the PC, it could assign a storage block size that isn't optimal for the camera, which in turn could reduce write performance, causing your card to drop frames at 60p?

Yes, I always format in camera.

Don't be fooled with card descriptions saying "up to".
This is fairly good info
https://www.sdcard.org/developers/overview/speed_class/

V rating shows minimum guaranteed speed which is key element. A lot also depends how well camera is designed and how big buffer it has to 'smooth' any spikes in needed speed.
Example: card saying up to 60MB but rated V30 will be most likely better then up to 90MB with V10 rating.

Yes, I always format in camera.

Does your card have a V30 label on it? Mine does, that could be the difference.

Thanks, pulalis. Do you know the file size and duration of the Q0 test? I’m curious about the bitrate achieved.

Welcome Rick.

Blackmagic raw Q0 4k 60 fps test. 23 minutes. 118 gb

85.5MB/s or the equivalent of 6.4:1 compression so that scene is quite close to the upper boundary of Q5. That’s making Q0 look more feasible. I thought Q0 had a minimum bitrate of 110MB/s or 5:1. Wonder what’s going on with the discrepancy between that chart showing the bitrate ranges of BRAW and your results?

Edit
Ignore this post of mine as it’s wrong.


Sent from my iPhone using Tapatalk

Thanks, pulalis. Do you know the file size and duration of the Q0 test? I’m curious about the bitrate achieved.

Welcome Rick.

Blackmagic raw Q0 4k 60 fps test. 23 minutes. 118 gb

85.5MB/s or the equivalent of 6.4:1 compression so that scene is quite close to the upper boundary of Q5. That’s making Q0 look more feasible. I thought Q0 had a minimum bitrate of 110MB/s or 5:1. Wonder what’s going on with the discrepancy between that chart showing the bitrate ranges of BRAW and your results?


The real time I was shooting is 9 minutes. The clip length is 60 FPS for 23 minutes

So that explains my discrepancy. Q0 was recorded about 207MB/s equivalent to BRAW 3.9:1 (since we are shooting at 60 fps in the BMPCC4K). The low end Q0 bitrate is 163MB/s at 60 fps equivalent to 5:1 compression. The high end Q0 bitrates 406MB/s at 60 fps equivalent to 2:1 compression. Given there was a fair amount of detail in your frame, I’m encouraged to shoot Q0 again.

Makes perfect sense now.

Edit
My earlier post was getting mixed up with BRAW on the 4.6K sensor at 30 fps. Hope I’ve got it accurately stated now! My figures assume that about 9.5 minutes were recorded at 60 fps and that accounts for playback time of 23 minutes at 25 fps.

Would be nice to have the BRAW bitrate comparison chart for each sensor as part of the camera manual rather than the just the webpage description based on the 4.6K sensor.


Sent from my iPhone using Tapatalk

So can a $100 256GB V30 card handle Q5 60fps continuously? What about 12:1?

For some reason, Q5 60fps was recording for 8 minutes on my cheap 10MB/s card, but 12:1 was not handling more than 8 seconds or so.

Question.. how often would you shoot continuous 60fps? I would think you only do it for shots you have planned for slow motion? What would be another reason to shoot an entire video in 60fps? Would that also apply to 120fps, 240, etc?

Question.. (not sure it was answered before) if BRAW is what makes 300fps (in HD) possible on the Usra Mini G2, why are we stuck to 60fps BRAW on BMPCC4K but 120fps using ProRes? I would like to think the hardware in the BMPCC4K is not that much older/slower than the new G2.. though given the price differences (besides all the other features/etc the G2 has), I would hope they have better hardware, especially to handle S35 sensor data. Still, I was hoping we would see the 120FPS HD still for BRAW, or even more FPS since BRAW is faster processing than ProRes. Curious if that is a future update, or maybe that is part of the 1.3 update and BMPCC4K just doesnt have it yet?

So assuming I shoot Q5 4K30 most of the time, but want to shoot Q0 4K60 from time to time, I mean I will most likely use my T5 for that, but I like the idea of having an SD card for backup (or possibly for primary Q5 acquisition). What SDXC card within the $100 range can do that? Primarily Q5 4K30 will be my primary acquisition (for sports stuff) or 4K24 for anything needing a little more cinematic feel to it.

Question.. (not sure it was answered before) if BRAW is what makes 300fps (in HD) possible on the Usra Mini G2, why are we stuck to 60fps BRAW on BMPCC4K but 120fps using ProRes?

120 fps BRAW is there. You probably forgot to switch the resolution over to 1920x1080....

Hmm.. I only tried it once prior, and dont recall.. but on my menu, I choose BRAW Q0 or Q5, resolution 1920x1080. Project frame rate only goes to 60fps. I do NOT have off speed recording enabled. If I enable that, then off speed frame rate allows me to go to 120. Is that the only way to choose it? I find it odd that the project frame rate is stuck at 60 though.

Just shot some video with this.. and load it in Resolve, and it shows up as 60fps. I press HFR and I see in the top left it says 120/60. What am I missing here? When I choose HFR the amount of recording time shows half of what it did prior.. so I assume it is putting me in 120fps mode.

Also, Window Sensor is ALWAYS on. No way to disable it. Even when I select Ultra HD, the window sensor is disabled but on. It seems only in DCI mode does Window Sensor disable and off. Tried it with all qualities selected too.

For some reason, Q5 60fps was recording for 8 minutes on my cheap 10MB/s card, but 12:1 was not handling more than 8 seconds or so.

Q5 can run in some situations under 10 MB/s, but 12:1 is always at least 27 MB/s at 4K, that is why. (talking about 24fps)

Well I just dropped $40 on a 128GB Sandisk Extreme Pro. Figured many are shooting 4K60 at 12:1 on it for several seconds, lets see what it does. If I can fit an hour of Q5 or so on it, I would be very happy, mostly at 4K30, but some 4K60 and 2K120.

OK..got the 128GB Extreme Pro in just now, tried it out.

Shoot, I was hoping to be able to hit Q0 30fps on 4K, but it only does about 3 seconds. I can record Q5 and 12:1 60fps, though I only went about 15 seconds before I stopped in both. 8:1 only goes 4 seconds at 4K30 (DCI), so it looks like this card is only good for Q5 and 12:1. Which honestly I would almost always shoot in anyway. I cant imagine using Q0 or 3:1 for anything but green screen work or I guess ultra detail shots? Not even sure what would constitute those.

Toshiba Exceria Pro - N501 SD Memory Card UHS-II U3 Class 10

Blackmagic raw 4k 12. 1- 60 fps
Blackmagic raw 4k 8.1 - 60 fps
Blackmagic raw 4k 5.1 - 60 fps
Blackmagic raw 4k 3.1 - 60 fps
Blackmagic raw 4k Q0 - 60 fps

Were tested.. This card is very good..

The Toshiba can't record braw for me, even at 12:1 24p