Stories
Slash Boxes
Comments

SoylentNews is people

posted by martyb on Wednesday May 20 2020, @07:40PM   Printer-friendly
from the put-your-whole-system-on-a-tiny-chip dept.

SD cards hop on the PCIe 4.0 bus to hit 4GB/s with version 8.0 of storage spec:

As outlined in a whitepaper [PDF] this month, the new spec will let existing SD Express and microSD Express cards employ PCIe 4.0 and NVMe to deliver a top speed data transfer speed of [3938 MB/s].

While the new spec is backwards-compatible, the latest top speed will only come with a card reader capable of connecting to the extra row of pins present on SD Express cards that support dual PCIe lanes.

[...] The good news is that SD Express and microSD Express cards can still get to 1970 MB/s on a device with a single PCIe 4.x lane under version 8 of the specification, and SD Express can get there with a pair of 3.x lanes. Which is rather faster than many SSDs and, as SD Express can climb to 128TB on a single card, a rather tasty storage option.

Also at The Verge, PetaPixel, and Yahoo! Finance.


Original Submission

 
This discussion has been archived. No new comments can be posted.
Display Options Threshold/Breakthrough Mark All as Read Mark All as Unread
The Fine Print: The following comments are owned by whoever posted them. We are not responsible for them in any way.
  • (Score: 5, Informative) by takyon on Wednesday May 20 2020, @08:19PM (3 children)

    by takyon (881) <takyonNO@SPAMsoylentnews.org> on Wednesday May 20 2020, @08:19PM (#997084) Journal

    Uncompressed video is a lot higher than that. Resolution could be much higher, e.g. 8K or 360-degree video. But good luck sustaining those speeds with QLC NAND. This is more about future proofing and maybe making smaller transfers faster.

    If you look at page 5 of the white paper, you can see some new pins are needed to support two PCIe lanes on the full SD card. But they can get to 1969 MB/s with PCIe 4.0 and no new pins. So why not?

    There are some other advantages of SD Express mentioned in the white paper, but it looks like they are not specific to these two new speeds.

    The legacy SD interface up to UHS-I consists of six single ended signals that may operate either in 3.3v or 1.8v logic levels. The 1.8v signaling allows lower power consumption versus the 3.3v. The differential interfaces introduced by SD UHS-II/III interface (0.14-0.28v) as well as the PCIe interface (0.25-1.2v), introduced with SD Express, allows significant lower voltage swing that reduces the consumed power contributed by the PHY during data transfer. Since the transfer rate is faster, the bottom line is that the new high speed interface will be much more energy efficient than legacy SD UHS-I interface. This is due to the fact that a system may transfer the same amount of data in shorter time and stay longer time in standby mode (see Figure 14).

    --
    [SIG] 10/28/2017: Soylent Upgrade v14 [soylentnews.org]
    Starting Score:    1  point
    Moderation   +3  
       Informative=3, Total=3
    Extra 'Informative' Modifier   0  
    Karma-Bonus Modifier   +1  

    Total Score:   5  
  • (Score: 2) by bzipitidoo on Thursday May 21 2020, @01:29PM (2 children)

    by bzipitidoo (4388) on Thursday May 21 2020, @01:29PM (#997363) Journal

    Raw video? Outside of movie studios, who uses raw video? It's too big. Just 1 minute of HDTV quality (1080p), uncompressed, at 30 frames per second, and assuming true color (24 bits per pixel), no audio, is roughly 11G of data. A 90 minute movie would require 1T of storage.

    Crank that up to 8K at 60 frames per second, and 48 bit "Deep Color", and the storage requirement is 64x more. Or, 128x more for 120fps. Throw in a 2nd camera for stereoscopic vision for 3D video, and that's another doubling. And I understand there's talk of going up to 16K resolutions some day. Lossless video compression (lossless H.265) can reduce storage requirements to very roughly 10%. 10% of 256T is still an awful lot of data.

    • (Score: 2) by takyon on Thursday May 21 2020, @01:43PM

      by takyon (881) <takyonNO@SPAMsoylentnews.org> on Thursday May 21 2020, @01:43PM (#997368) Journal

      I would have to do a bit more research to see what storage standards cinematic 6K+ cameras are using these days. Maybe this helps make SD more relevant to that market, maybe not.

      For filming with drones or action cameras (GoPro), SD Express or microSD Express is relevant. Less weight helps, less power usage helps. FPS can be very high for recording slow-mo shots. Apparently, GoPro does up to 240 FPS 720p/1080p, I could have sworn there was a 720 FPS model out there.

      Based on my crapculations, the ultimate target for 360-degree video cameras should be around 32K resolution [soylentnews.org], with VR headsets displaying 16K resolution (1/4 of the sphere with very wide/tall FOV).

      --
      [SIG] 10/28/2017: Soylent Upgrade v14 [soylentnews.org]
    • (Score: 2) by takyon on Thursday May 21 2020, @03:03PM

      by takyon (881) <takyonNO@SPAMsoylentnews.org> on Thursday May 21 2020, @03:03PM (#997400) Journal

      H.265 tiers [wikipedia.org] include up to full 8K (8,192×4,320) @ 120 Hz.

      H.266 [wikipedia.org] will have support for 16K, 360-degree videos (not sure what that entails), and 48-bit deep color (16 bpc):

      In October 2015, the MPEG and VCEG formed the Joint Video Exploration Team (JVET) to evaluate available compression technologies and study the requirements for a next-generation video compression standard. The new algorithms should have 30-50% better compression rate for the same perceptual quality, with support for lossless and subjectively lossless compression. It should support resolutions from 4K to 16K as well as 360° videos. VVC should support YCbCr 4:4:4, 4:2:2 and 4:2:0 with 10 to 16 bits per component, BT.2100 wide color gamut and high dynamic range (HDR) of more than 16 stops (with peak brightness of 1000, 4000 and 10000 nits), auxiliary channels (for depth, transparency, etc.), variable and fractional frame rates from 0 to 120 Hz, scalable video coding for temporal (frame rate), spatial (resolution), SNR, color gamut and dynamic range differences, stereo/multiview coding, panoramic formats, and still picture coding. Encoding complexity of several times (up to ten times) that of HEVC is expected, depending on the quality of the encoding algorithm (which is outside the scope of the standard). The decoding complexity is expected to be about twice that of HEVC.

      On top of everything you mentioned, there is High Dynamic Range, which may or may not eat into that color depth. I think it depends on the standard.

      AV1 looks similar to H.265. Nobody knows what AV2 will add yet, but it will probably be more of the same.

      --
      [SIG] 10/28/2017: Soylent Upgrade v14 [soylentnews.org]