[HN Gopher] Apple's Pro Display XDR takes Thunderbolt 3 to its l...
___________________________________________________________________
Apple's Pro Display XDR takes Thunderbolt 3 to its limit
Author : WithinReason
Score : 197 points
Date : 2023-11-24 09:34 UTC (13 hours ago)
(HTM) web link (fabiensanglard.net)
(TXT) w3m dump (fabiensanglard.net)
| lispm wrote:
| Does the MacBook Air with M1 (or M2) actually support 10bpc on
| the 6k Display? It is never mentioned in the specs about the
| Macbook Air, AFAIK. My guess is that it does not. The M1 from
| 2020 also did only support 8bpc on the internal display.
|
| For the MacBook Pro with M1 Pro it is explicitly mentioned in the
| tech specs that it supports 6k with 10bpc.
|
| This page claims that even the M1 MacBook Air from 2020 supports
| 10bpc on the external 6k monitor: https://support.apple.com/en-
| us/HT210437
|
| But the actual technical spec page for the machine never says it,
| AFAICS.
|
| https://support.apple.com/kb/SP825?locale=en_US
|
| > Simultaneously supports full native resolution on the built-in
| display at millions of colors and: One external display with up
| to 6K resolution at 60Hz
|
| My impression was that Apple implemented Thunderbolt themselves
| (and got rid of external chips) and that at least the first
| generation M1 machines lacked the 10bpc feature in the GPU and/or
| the Thunderbolt part.
| dmitshur wrote:
| Do you have ideas for how to test it and find out definitively
| in person? I haven't found a way to confirm one way or another.
|
| The Graphics/Displays page in System Information unfortunately
| says nothing about colors, only resolution (https://github.com/
| shurcooL/home/assets/1924134/af4b19a3-b85...).
|
| When looking at a 16-bit PNG of a white to black gradient, I'm
| not able to visually spot any banding even when zooming in.
| It's fairly easy to spot steps when looking at a 8-bit PNG
| version of the same gradient. But the same happens on the
| built-in display despite it supposedly having only 8-bit color
| ("support for millions of colors").
| lispm wrote:
| Yeah, that's definitely my experience, too. It is difficult
| to actually get color depth information. Several other Macs,
| especially the older ones tell you there the color depth.
|
| You might want to try out SwitchResX and see what it says
| about the screen. In some menu, there is color depth
| information about the screen. But I don't know if it is
| actually accurate and where the application gets this info
| from.
|
| https://www.madrau.com
| chx wrote:
| It doesn't work like that, the 40gbps is actual bus bandwidth
| AFAIK.
| https://www.thunderbolttechnology.net/sites/default/files/Th...
| figure 7 says 5120 x 2880 @ 60Hz which requires 22.18gbit/s
| leaves 18gbps data bandwidth. (This figure is quite important
| because this is one of the two only "official" sources which
| admits the raw data transfer limit of TB3 is 22gbps, the other is
| Dell at https://www.dell.com/support/kbdoc/en-
| uk/000149848/thunderbo... otherwise you'll only see the 40gbps
| speed.)
|
| What happens rather is much simpler, the blog post forgot to set
| the calculator to 10 bit
| https://linustechtips.com/topic/729232-guide-to-display-cabl...
| if they did you'd see the data rate required is 38.20gbit/s so
| the bus is near full. USB C has separate 2.0 wires so unless you
| have DP 1.4 for DSC you can only use those for USB data, there's
| no space for anything else.
| korhojoa wrote:
| That's not really the full story, when this is about displays.
| I currently have a display that requires Displayport 2.1 to get
| everything out of it, and with DSC and DP 1.4, I can get a
| "theoretical bitrate" that is higher than what they would
| normally allow. (Of course, this doesn't change what the link
| can do, it just allows you to do more with the same bandwidth)
| On thunderbolt 3 systems, the link is on HBR3, which with DSC
| allows 120hz, 10bit at 7680x2160. No DSC support limits the
| refresh to 60hz and 8bit (as then you won't have the higher
| rates available anyway, since DSC is mandatory with newer
| standards).
|
| You can check
| https://tomverbeure.github.io/video_timings_calculator to see
| what is possible.
| brigade wrote:
| Thunderbolt 3 doesn't use those USB 2.0 wires; that was one of
| the big changes in USB4/TB4 to use them. Once you're in
| Thunderbolt 3 mode, all USB ports are provided by PCIe xHCI
| controllers by the TB3 device.
|
| In actuality, the mode where the XDR display consumes 38gbps of
| uncompressed display bandwidth is an Apple-only mode requiring
| special Titan Ridge firmware that aggregates 6 lanes of HBR3.
| Contrary to the article, Alpine Ridge _does not_ support 6k,
| which specifically is why the iMac Pro doesn 't support 6k
| output. It requiring special firmware is also why this mode
| _only_ works when directly connected to the Mac.
|
| But yeah, it's annoying that everyone throws bandwidth numbers
| around without mentioning or even thinking if it's link rate or
| data rate.
| mark_l_watson wrote:
| I have one of these. My Dad died last August (just before his 102
| birthday) and he had bought himself a Pro Display XDR about 6
| months before that. My brother is a PC person and I am a Mac
| person so I got the Pro Display.
|
| It is an amazing display and I love it, but I would never buy one
| for myself. It is obviously fine for programming, but for me it
| really stands out as something for consuming entertainment, even
| though I only get 4K content. It is capable of, I think, 7K with
| the right computer and has 10 bit color depth. When my Dad first
| bought it, I used it to play Apple Arcade games on my iPad Pro -
| that was fairly spectacular.
|
| EDIT: my Dad had a Black Magic video camera that I think had 8K
| resolution, and so he had a lot of fun with his setup.
| Spooky23 wrote:
| I'm sorry for your loss, but I have to say that your post made
| me smile. How awesome that you and your brother got to enjoy
| dad at 101 being able to nerd out with video and high end
| displays.
| mark_l_watson wrote:
| Thanks!
| Teever wrote:
| Wow it would have been neat to talk to the kind 102 year old
| person who is buying this kind of hardware. I'd love to know
| what he thought about the progress of technology and how he
| felt it had impacted society.
| dbspin wrote:
| Your 101 year old father was out shooting on an Ursa 12K? What
| a guy.
| xattt wrote:
| The frail elderly are a very prominent group in society
| because their needs are so great. Robust "old-old" adults
| tend to blend in because they are inconspicuous and they go
| on about their business.
|
| I think we're going to see a "silver tsunami" of robust
| elderly persons as millennials and Gen Xers age simply
| because healthy lifestyle activities that had been a part of
| their lives.
|
| I.e. don't buy into Acorn Stairlift and Lifealert futures.
| FirmwareBurner wrote:
| _> robust elderly persons as millennials and Gen Xers age
| simply because healthy lifestyle activities that had been a
| part of their lives._
|
| You're probably leaving out the issues around teflon,
| microplastics and antibiotics poisoning all food, air and
| water, general increased stress and anxiety about the wars,
| economy, job market, environment, debt and unaffordable
| rent/housing, the loneliness epidemic plaguing the west,
| which have already tanked their/our sperm count so we can't
| be too sure they'll/we'll see much healthier retirements if
| these keep piling up.
|
| Those with solid careers in tech in developed countries
| yeah sure, they'll be fine and happy, most likely retired
| early, house and debt paid off and focused on enjoying
| their hobbies instead of working the 9-5 grind. The rest,
| not so much.
| graphe wrote:
| Just because you're old doesn't mean you're rich. China
| and HK has old ladies working out on the streets.
| FirmwareBurner wrote:
| That was kind of my point. If you wanna enjoy your old
| age you also need to be somewhat wealthy or at least
| financially very table.
|
| The old people working in the streets till they drop in
| China and Korea do it because they have no wealth to rest
| on, not because they enjoy doing that kind of work so
| much.
|
| It's doable to be young and poor, but being old and poor
| sucks.
| graphe wrote:
| He said robust old people. The Chinese old ladies are
| robust and impoverished.
| data-ottawa wrote:
| This 101 year old lived through a world war with
| rationing, would have been born into the great
| depression, saw the rise and fall of Nazi Germany, the
| Cold War and Cuban Missile Crisis, the Kennedy
| Assassination, the Nixon years, the oil crises and
| recessions, Gulf wars, 9/11...
|
| On the healthcare front there was the proliferation of
| lead (in paint, toys, fuel, everything), smog from cars
| and coal burning, toxic fertilizers, the rise and fall of
| smoking, the discovery of HIV, Polio outbreaks, things
| like the Cuyahoga river fire (where rivers were so
| polluted they literally caught fire every couple
| decades). The mining town my family lived in would just
| throw the arsenic and mine tailings into the lakes
| because they figured it couldn't hurt them there, and
| that was a common thing to do at that time.
|
| Gen X and Millennials are not the only generations who
| have faced adversity. It's a rough moment now for sure,
| but it's not unique. We shouldn't fall into baseless
| optimism but also don't shouldn't neglect human strength
| and creativity. We have new problems, and we have new
| tools.
| FirmwareBurner wrote:
| _> This 101 year old lived through a world war with
| rationing, would have been born into the great
| depression, saw the rise and fall of Nazi Germany, the
| Cold War and Cuban Missile Crisis, the Kennedy
| Assassination, the Nixon years, the oil crises and
| recessions, Gulf wars, 9/11..._
|
| Sure, not stealing his thunder, but that's how selection
| bias works. Not everyone got to live to 101 despite maybe
| even living healthier lives. I know people in their 40's
| who already died of cancer. Life can always throw you a
| curb ball.
|
| _> Gen X and Millennials are not the only generations
| who have faced adversity._
|
| Fair point.
| data-ottawa wrote:
| Living to 101 is definitely not representative.
|
| My point really was that second one.
|
| As I said, it's a rough time right now, we're going
| through a lot. But we passed environmental reforms
| before, we removed lead from gas, we invented vaccines,
| we set standards for chemicals, we've cured a few people
| of HIV, there is good to find out there.
| xattt wrote:
| Yes, there is obesity and yes, there will be long COVID,
| but the health-positive initiatives (more women actually
| encouraged to work out; men not perceiving weightlifting
| as gay; marathons are a normal thing now; herpes zoster
| vaccines helping with long-term immunity against a
| probable cause of Alzheimer's dementia; cigarette smoking
| as socially unacceptable behaviour) will tip the scales
| in favour of longevity towards making it to 100-120.
| philsnow wrote:
| > We shouldn't fall into baseless optimism but also don't
| shouldn't neglect human strength and creativity. We have
| new problems, and we have new tools.
|
| Thank you for this comment, it helps to contextualize two
| moods that I have, as one who has struggled with
| depression (not currently, but off and on):
|
| When I'm in a low mood it's easy to see and dwell on the
| new problems and discount the efficacy of the new tools.
|
| When I'm in good spirits it's easy to see the new tools
| and (temporarily) forget about the new problems.
| macNchz wrote:
| > I think we're going to see a "silver tsunami" of robust
| elderly persons as millennials and Gen Xers age simply
| because healthy lifestyle activities that had been a part
| of their lives.
|
| > I.e. don't buy into Acorn Stairlift and Lifealert
| futures.
|
| Trends in obesity-which is a huge driver of poor health in
| America-don't seem to support this hypothesis. I think the
| great majority of health and wellness activity in recent
| years has been concentrated among people at the upper end
| of the socioeconomic scale, which also drives perception
| since companies will spend a lot on advertising to attract
| people with money. Things in this country look very
| different depending on how far you are from the nearest
| Whole Foods/Equinox/Soulcycle/Sweetgreen.
| MBCook wrote:
| Apple made a big deal about using it for video production and
| how it could replace extremely expensive reference monitors
| during its introduction, if I remember correctly.
|
| Through that lens it seems like a useful product.
|
| For everyone else it seems like a pretty amazing monitor if
| money doesn't matter. It's most useful quality is probably
| being 6k, so you have tons of screen real estate.
| easygenes wrote:
| It doesn't _really_ though. There was hope it would be a
| dual-layer LCD device that could, but alas we're stuck with
| $20k+ Sony monitors for that still.
| teaearlgraycold wrote:
| Hopefully I'll have my InfinityK display at 100 that can get
| passed on to my kids
| belugacat wrote:
| I do visual work (graphic/UI design, photography, video
| editing) along with programming and there is no display with
| the resolution and color fidelity of the XDR at its price
| point. I got one shortly after release, and if it stopped
| working today I'd buy another one in the amount of time it
| takes me to click "Submit" on the Apple Store. It's just that
| good.
|
| When I look at a high resolution scan of a large format
| negative on it, it feels like looking at it directly on a light
| table. It's insane. My only complaint is the local dimming,
| which shows its limit when you're doing fine white on black
| linework in a dark room. Hopefully we'll get a pro OLED display
| of that quality in the next decade which will solve that one
| issue.
|
| The only other piece of hardware I've spent money on that comes
| close of giving me the same satisfaction is my Happy Hacking
| Keyboard, which I've used for over a decade now and I hope I
| will keep using until I cannot use computers altogether anymore
| (I have a few spares just in case).
| mark_l_watson wrote:
| Thank for that. I used to be into photography and I did just
| once try shooting raw images with my Canon and view/edit.
| nickpeterson wrote:
| Not to presuppose anything intimate, but if your dad was 101
| buying that monitor, he was basically buying it for his kids as
| much as himself ;) sounds like he was into neat stuff!
| kylehotchkiss wrote:
| Your dad was shopping for cutting edge Apple tech at 101!
| That's super cool. I aspire to be like him and never lose my
| sense of wonder about tech. Sorry for your loss, but I also am
| happy to hear you got to enjoy many years with him.
| amluto wrote:
| Is Thunderbolt 3 like the rest of the USB-C protocols in that
| dedicated pins are used for USB2? If so, the bandwidth doesn't
| need to add up -- the cable carries the Thunderbolt protocol and
| USB2 separately.
| Moto7451 wrote:
| https://www.etechnophiles.com/thunderbolt-pinout-1-2-3-4/
|
| Based on the D+/- pairs I believe you're correct. On Titan
| Ridge and later add in cards USB 2 is expected to be passed
| through via a motherboard header as the controllers lack a
| dedicated USB 2 chip.
| tedunangst wrote:
| Would people be happy if Apple refused to support incremental
| improvements? Sorry, the sticker says Thunderbolt 3, you can't
| use fast USB devices here. You have to wait for the next product
| cycle to get a new sticker.
| xoa wrote:
| I still desperately wish the industry had been able to push
| through to an optical data + plain power interconnect standard
| ages ago instead of it falling apart. It's so pleasant to deal
| with fiber. The same OS2 or OM3 I installed over a decade ago for
| 10 Gbps is still fine for 40 or 100 Gbps. Cost was a complaint at
| one point, yet even without the enormous economies of scale a
| general standard would bring price and performance lines have
| ended up converging anyway. 40G SR (so still good to 150m)
| modules are now at $40 or less, even 100G is less than $100.
| Putting that onto simpler fiber instead of MTP/MPO with SFP56
| remains much more expensive, a 50G-SR is still like $280, but
| that appears to primarily be a product of it being very new and
| not yet scaled, not that it couldn't have been years ago. And
| that still then runs to 100m with duplex LC.
|
| Meanwhile, Apple wants $70 a pop for a single, 1m Thunderbolt 4
| cable, and that's only been increasing. What will Thunderbolt 5
| be? Corning and I think one other briefly did optical Thunderbolt
| cables, but those were $500-1000. Whereas premade quality duplex
| OM4 with helical steel armor runs more like $2.20/meter.
|
| Feels like we somehow ended up in a yet another technology path
| dependent evolution path where decisions that saved a bit at the
| time have then imposed major costs forever more :(. Man it'd be
| so cool to just be able to run a screen and input boxes hundreds
| of feet away from a workstation for $60, or have an $8-12 cable
| be good for decades of evolution in bandwidth barring regular
| wear (and when it's that cheap who cares even if it breaks after
| 5 years?), or be able to route displays/TVs and PCIe and whatever
| else around like any other networking with no compromises. Sigh.
| BiteCode_dev wrote:
| Fiber doesn't bend well, which for a cable that will be moved
| regularly by ordinary customers, is game over.
| kllrnohj wrote:
| It bends just fine. There's tons of fiber cables for
| "ordinary users" including HDMI, DisplayPort, Thunderbolt, VR
| link cables, and, of course, toslink. You just end up with
| the extra cost and complexity of having transceivers built
| into the cable which is a waste especially at these speeds
| where copper is clearly a limitation.
| gregsadetsky wrote:
| I didn't know about fiber HDMI, DisplayPort or
| Thunderbolt...!
|
| Are these used for very long runs - as in a video source to
| a projector (via hdmi) hundreds of meters away, like in a
| stadium..?
|
| Thank you!
| mlyle wrote:
| 8m or so is the limit for HDMI passive cables being 100%
| reliable.
|
| If you want to go 15-30m--- using a video source from
| another part of your house, or to drive a projector in
| the middle of a classroom--- and you buy an "active
| cable", odds are it's fiber optic inside.
|
| If you want to go hundreds of meters away, you'll get a
| purpose-built box that uses your own optical cables
| instead of a cable that hides the optical transceivers
| inside.
| tuetuopay wrote:
| Those are used in multiple cases:
|
| - long runs that are not that long. hdmi does not like
| long cables _at all_. even an overhead projector in a
| classroom requires super expensive cables (ever wondered
| why it 's still mostly vga?)
|
| - packing multiple displays in a single cable. fiber is
| so thin that with trunk cables you get _a lot_ of strands
| in a single cable, capable of running a lot of displays
|
| - just getting a really thin cable that can be run in
| existing conduits or hard to reach places
|
| I did buy such a cable for home for the third reason, to
| run from the PC in the office to the TV in the living
| room. It runs on a bog standard OM3 MPO cable. The
| specific one I got comes from HeyOptics and their website
| already showcases a few usecases [1]. (not affiliated,
| just a great product that just works)
|
| As for stadiums and more generally broadcast video,
| they're using SDI instead of HDMI. Those are indeed most
| of the time fiber, both for range and weight (think of
| the cameraman running along the terrain during a sport
| event, and their long tail of cables). When they use HDMI
| it's more in the control room.
|
| [1]: https://www.heyoptics.net/products/8k-hdmi-mpo-
| optical-cable
|
| (edit: fixed list formatting, I always forget this is not
| markdown)
| macNchz wrote:
| I have a 50 foot fiber HDMI cable to get 4k/120hz signal
| from the PC in my home office to the TV in the living
| room. Works great!
| bartvk wrote:
| That's quite amazing. How thick is it? For metric folks,
| that's more than 15 meters!
| macNchz wrote:
| It's fairly thin, quite a bit thinner than most of my
| traditional copper cables I think. Apparently it's
| nominally 15m, they also have a 20m one:
| https://www.monoprice.com/product?p_id=43328
| kllrnohj wrote:
| Copper thunderbolt cables are only good for 2-3m, after
| which you need active cables or repeaters and fiber is
| easily the best option at that point.
|
| HDMI and DisplayPort are good for a bit longer than
| Thunderbolt over copper, but not by all that much. HDMI
| 2.1 can only go to around 3m as well now.
|
| So we're in a world where "long" is a mere 5 meters / 15
| feet. This is why so many VR headsets are using fiber
| cables - it has to be long to enable the movement and
| logistics of connecting a PC to someone freestanding in a
| room, but modern video signals are just too hard to drive
| over copper at that not really that long distances.
| xoa wrote:
| > _Fiber doesn 't bend well_
|
| Eh? Typical min bend radius is like 10x diameter in static
| conditions, so for a 2mm fiber cable that'd be 20mm or ~0.8".
| That's not an issue at all with consumer usage, and if you're
| really worried is trivially solved by just building the thing
| up to the level of thickness Thunderbolt cables have already.
| That's why Corning for example advertised their old TB
| optical cables as "zero bend radius" [0], adding more
| polymer/armor around a fiber cable so that someone can do
| whatever with it and it naturally won't go out of spec isn't
| a big deal.
|
| FWIW, anecdotally high performance copper doesn't like
| abnormal use either. 10g USB-C cables are cheap enough that I
| tried taking one and using a vice grip to actually _really_
| squash thing at an angle, and it didn 't like that at all in
| terms of working reliably afterwards. I doubt DP, HDMI and
| the like would do better, and those are plenty thick too. It
| never comes up though in normal use. And when cables are
| _cheap_ vs stupid pricey making a mistake no longer is such a
| big deal. If you use an $8 cable a bit too hard and it stops
| working, well grab another one out of the drawer. If your $70
| cable breaks that 's a touch worse.
|
| ----
|
| 0: https://www.corning.com/microsites/coc/ocbc/Documents/CNT-
| 00...
| lostlogin wrote:
| I'm with you and want more fibre, but people are brutal on
| hardware.
|
| I've been involved in the replacement of two fibre optic
| cables at work, used for pulse oximetry in an MR scanner.
| It was very expensive, twice.
|
| People force a bend and destroy them. It doesn't matter
| what you tell people, they break them.
| zamadatix wrote:
| Was in healthcare IT as well: https://www.fs.com/products
| /41028.html?attribute=35025&id=61... for a couple of
| dollars you can get 2 paths of crush protected bend
| insensitive high speed fiber that barely degrades signal
| quality when you knot it up. To get around the
| limitations of the LC connector and wall jacks being
| regularly obliterated by equipment/tables being moved
| around use recessed jacks, that way they just pass by
| instead of break the connector (applies to rj45 as well).
| We had great success with this approach for a couple
| thousand locations.
|
| Of course, the real problem 95% of the time is really
| around the device not the technology the device uses.
| There is a misalignment of incentives on who can work on
| the device, whether standard parts are used, and whether
| the approved parts are the $2 type solution that will
| cost $800 in on site contractor fees to replace again or
| the $4 solution which actually stands a chance to being
| used. Once the device is bought nobody is going to be
| incented or allowed to do anything but fix it to status
| quo. It's one reason I had to get out of healthcare IT -
| it was more often the system getting in the way of what
| patients and nurses needed to do than the actual
| technology itself so solving things from a technology
| perspective felt like running on a treadmill and going
| nowhere.
| eurekin wrote:
| I'd go like: "third time, I'm wrapping it in a kiddy pool
| noodle" for their embarassment
| zamadatix wrote:
| In my experience they wouldn't even mind or be
| embarrassed, so long as said solution sounded like it
| would stay out of the way and let them use the device
| more conveniently. To them that's the whole point of IT:
| make it less painful to use the devices how they want to
| use them to do their job. IT isn't the savior that
| designs things they want, it's the cost center that makes
| the things they have to use less annoying to use.
| jasomill wrote:
| Until someone figures out how to supply bus power over
| optical fiber, I don't see it as a viable alternative for
| typical consumer peripheral I/O applications, no matter how
| inexpensive and robust the cables are.
| wtallis wrote:
| The most common use cases for the kind of long range high
| bandwidth connectivity that fiber is good for are
| networking and displays, neither of which typically carry
| power in consumer use cases. USB cannot be replaced by a
| purely optical connection, but Ethernet and DisplayPort
| certainly can, and tunneling USB alongside DisplayPort to
| split back out at the monitor doesn't present any power
| delivery challenges.
| vinay_ys wrote:
| We have FTTH everywhere in India and seeing how fiber is
| installed in all kinds of nooks and crannies, I can tell you
| that fiber is more resilient than you think. And frankly, I
| have experienced far less Internet outages with FTTH than I
| used to with twisted copper pairs (ADSL2+).
| dagmx wrote:
| How thick are those cables versus standard consumer cables?
| InvaderFizz wrote:
| Most GPON is quite thin. Thinner and tougher than
| standard fiber patch cables.
| chrischen wrote:
| The Meta Quest link cable is optical and way more than 1m
| long.
| solarkraft wrote:
| Fiber cables are often more flexible than copper ones by
| virtue of just being a lot thinner.
| lxgr wrote:
| Which exacerbates the problem of users easily exceeding the
| minimum allowable bending radius.
| jauntywundrkind wrote:
| USB is still (as far as I know) a mess and that sucks.
|
| But DisplayPort has great optical cables available! I was early
| in & they were very cheap then (no one trusted them yet), 50m
| v1.4 for $60. Turns out to be vastly longer than I needed, &
| got a 20m cable for gaming from the roof. There's even DP 2.1
| cables now.
|
| A USB4 with optical transport would be divine.
| lxgr wrote:
| There are optical USB-C cables available! The standard
| explicitly supports that use case.
| Aurornis wrote:
| > It's so pleasant to deal with fiber.
|
| Consumers are absolutely brutal to cables. Minimum bend radius
| of copper cables get violated all the time, but copper has a
| decent chance of continuing to work if bent back. Not so with
| fiber. Fiber would only work with some heavy armor.
|
| Consumer fiber was actually tried in the past: Optical TOSLINK
| was briefly popular in the audio world. It was cheap relative
| to what you'd need for modern high speed fiber, but even that
| was too much to win out over copper. It faded away.
|
| Even within professionally maintained data centers, direct
| attach copper is often preferred over fiber interconnects when
| it's possible to get away with it.
|
| Fiber is great when called for, but copper wins in practicality
| when you can get away with it. It's been proven over and over
| again across industries.
|
| > Meanwhile, Apple wants $70 a pop for a single, 1m Thunderbolt
| 4 cable, and that's only been increasing.
|
| You picked the absolute most expensive cable as your benchmark.
| Look anywhere else and prices are lower and decreasing.
|
| Also, you are wrong about Apple cable prices increasing.
| They've actually dropped the price of their longer cables.
|
| But again, look to the overall market and you'll see prices
| going down.
| FirmwareBurner wrote:
| _> Consumer fiber was actually tried in the past: [...] It
| faded away._
|
| Not everywhere. It's still not as popular as copper but it's
| not extinct.
| ender341341 wrote:
| to be fair TOSLINK was a particularly bad standard and offers
| lower audio quality to similar or cheaper analogue devices.
| It's whole selling point was basically "it's fiber optic so
| that makes it better" while not actually being better.
| RF_Savage wrote:
| No ground loops is a feature.
| yummypaint wrote:
| It's literally the exact same digital signal sent over
| copper S/PDIF. The only meaningful advantage it offers is
| electrical isolation.
| xnzakg wrote:
| Didn't stop manufacturers from making gold plated TOSLINK
| connectors though.
| pdpi wrote:
| I wish I could dismiss that as a joke, but I have in fact
| seen some of those.
| pdntspa wrote:
| It fools the same people who somehow think toslink audio
| quality is worse over a bitperfect link
| ender341341 wrote:
| No one's saying it didn't transmit the data bit perfect,
| but it re-encoded the media at a relatively low bitrate
| with a decently crappy codec which is where the crappy
| audio came from.
|
| EDIT: Looking closer at the wiki (it's been forever since
| I saw it used), it's when it's being used for surround
| sound that it compresses it super poorly, for stereo it's
| just PCM and sounds fine.
| ender341341 wrote:
| S/PDIF is also crappy in comparison to analogue audio of
| the time.
|
| EDIT: Looking closer at the wiki (it's been forever since
| I saw it used), it's when it's being used for surround
| sound that it compresses it super poorly, for stereo it's
| just PCM and sounds fine.
| atoav wrote:
| Lower audio quality? Nope. Toslink sends the actual bits
| that make up you audio digitally over the cable. If the
| receiving device does a shit job at converting the signal
| to analog, this is barely the fault of the standard.
| ender341341 wrote:
| The codec they transmitted over the wire (regardless of
| the wire media) was crappy, not the wire itself, but in
| this case those imply the same thing as as far as I'm
| aware there was only the one format that it'd transmit
| over that connector.
| neckro23 wrote:
| What codec? It's PCM.
|
| You could also use that PCM stream to transmit Dolby
| Digital or DTS, but that's up to the device, not TOSLINK.
| ender341341 wrote:
| okay, looking closer at the wiki (it's been forever since
| I actually tried it), for stereo it's full bandwidth,
| it's when it's surround sound it's super compressed.
| atoav wrote:
| [delayed]
| toast0 wrote:
| I've used spdif in coax and fiber (toslink) to transport
| audio from tv (atsc1) and dvd where you're just
| bitstreaming the data from the antenna or the disc. It's
| also fine for 2-channel PCM.
|
| Dolbly Digital (ac-3), dts, and 2-channel PCM are fine
| for what they are. More channels in PCM would be nicer,
| as would other newer higher bandwidth, lossless codecs,
| but as a unidirectional signal, it's hard to add support
| for more stuff.
| conradfr wrote:
| It's still used to link audio interfaces with ADAT.
| vlan0 wrote:
| Have you played with the newest bend insensitive fiber? You
| can wrap it around a pencil and an otdr shows no or minimal
| loss. But I think you're correct about damage from average
| consumers. They'd destroy the ferrule in no time.
| Aurornis wrote:
| The failure mode with consumers is usually pulling cables
| against a sharp edge. Even a pencil has a larger radius
| than the edge of your desk or the corner of your computer
| case. You can buffer the radius by putting a thick jacket
| around the fiber, which is about the only thing that works.
|
| But fiber still isn't a free lunch. The longest high speed
| cables already use fiber internally, but they're expensive
| because they need extra optics and transceivers inside.
| Moving the extra fiber hardware into the laptop and client
| device would make the cables cheaper but make the hardware
| more expensive. That's not a trade off that most people
| would take as most people don't actually need long cable
| runs.
|
| The above comment was trying to use Apple premium cables as
| the reference for being too expensive, but opening Amazon
| shows plenty of Thunderbolt 4 cables in the $20-$30 range
| from other vendors. It's really hard to imagine a scenario
| where forcing every cable to be a combination of fiber and
| copper would make things cheaper for us consumers.
| _a_a_a_ wrote:
| I've never even heard of a minimum bend radius, and I care
| about my kit so I do tend to read the docs that come with it.
| Where are such details given, because I'm now curious,
| thanks.
| adhesive_wombat wrote:
| I don't know about all cables, but Cat 6 has a generally
| accepted minimum radius of 4 times cable width and there's
| some standards to back that up.
|
| You can also find datasheets for industrial cable that
| specify it for fixed installations and repeated flex
| applications: https://www.sab-
| kablo.com/fileadmin/user_upload/pdf/catalog_... (notably
| this has 5 times the diameter, so it wouldn't pass some
| standards)
|
| Consumer cables, not just Ethernet, probably do have such
| specifications when produced in bulk, but the manufacturer
| that turns reels of 1000 metres of raw cable might not
| include it in the end manual (just like they usually don't
| include the frequency response curves in there).
| Someone wrote:
| Would high speed optical work for connections that require
| regular unplugging?
|
| https://en.wikipedia.org/wiki/TOSLINK and
| https://en.wikipedia.org/wiki/ADAT_Lightpipe are limited to
| about a megabyte/second, if I'm doing my math right.
| https://en.wikipedia.org/wiki/MADI#Sampling_frequency gets
| higher, at around 12 megabytes/second. That still is at least a
| factor of 100 away from this cable.
|
| I think that's because it isn't easy to make a plug that works
| reliably for a gigabytes per second fiber cable.
| lwkl wrote:
| You would need a new connector because dust can break the
| signal and a cable that needs a dust cap sounds like a
| horrible consumer product.
| gatkinso wrote:
| This is my dream. To those worried about bending, I would be
| really curious to try one of these with levels of sheathing to
| see how they hold up in reality.
| spicyjpeg wrote:
| Alec from Technology Connections covered the topic of consumer
| optical standards - or rather, the lack thereof - in a video
| that's now 4 years old:
|
| https://youtube.com/watch?v=CwZdur1Pi3M
|
| It mostly boils down to the reason others already mentioned:
| fragility, cost (optical transceivers have only recently become
| cheap enough to enable the use case of optical cables that
| behave like copper ones) and the requirement for copper wires
| anyway in order to carry power in addition to data.
| mmastrac wrote:
| Normally I find Fabien's articles to be a lot clearer, but I came
| away from this one not understanding what it was really talking
| about.
|
| Does the monitor have XDR? Does it auto-negotiate extra bandwidth
| on newer MacBooks?
| BearOso wrote:
| XDR means nothing. It's an Apple buzzword.
|
| With DSC, you have at the minimum a fixed amount of bandwidth
| gain because blanking intervals can always be compressed, and
| DP is packetized without strict timing limitations. That should
| be enough to guarantee the additional bandwidth.
| Reason077 wrote:
| From the article:
|
| > _" But why is the 16-inch MacBook Pro able to run USB 3.1
| (10Gbps)?"_
|
| USB 3.1 Gen 1 is 5 Gbps, not 10 Gbps.
|
| (My Thunderbolt LG Ultrafine 4K also has USB 3.1 Gen 1 ports on
| it, so I should know...)
| jeroenhd wrote:
| I think they mean USB 3.2 Gen 2x1, which does support 10Gbps
| throughput.
|
| Apple's tech support page about their Thunderbolt 4 cable
| (https://support.apple.com/en-om/HT210997) also states "USB 3.1
| Gen 2 data-transfer speeds up to 10Gbps".
|
| I'm a little confused by Apple's machines wouldn't support USB
| 3.2 Gen 2x2 when they support USB 4 and Thunderbolt. I guess
| they couldn't figure out how to get faster USB out of their
| chipset? They list the same limitation on their iMac USB 3
| ports (https://support.apple.com/guide/imac/take-a-tour-imac-
| apd2e7...).
| Reason077 wrote:
| > _" I think they mean USB 3.2 Gen 2x1, which does support
| 10Gbps throughput."_
|
| They explicitly say USB 3.1 Gen 1 in the article, and the
| linked technical document [1]
|
| > _" I'm a little confused by Apple's machines wouldn't
| support USB 3.2 Gen 2x2 when they support USB 4 and
| Thunderbolt."_
|
| Yeah, it's weird, but Apple has never supported Gen 2x2,
| AFAIK. 2x2 means you have two 10 Gbps channels running in
| parallel on two sets of pins, but USB 4 does a similar thing
| to get 40 Gbps and they _do_ support that. Shrug.
|
| [1] https://fabiensanglard.net/xdr/Pro_Display_White_Paper_Fe
| b_2...
| londons_explore wrote:
| Can we please stop having displays be 'special'. I should be able
| to plug hundreds of displays into my laptop with any combination
| of USB hubs and have them all just work.
|
| Displays shouldn't need allocated bandwidth - they should give
| the best display quality possible given the available bandwidth.
|
| Nothing moving on the display - no bandwidth used. Just a little
| animated gif? Just a few kilobits used. Full screen HD video?
| Gigabits used. Gigabits not available? Frame rate and/or quality
| drop.
|
| This is the exact behaviour of plugging in hundreds of USB
| ethernet adaptors. Why should displays be different?
| andrewmcwatters wrote:
| The real reason? Because display protocols don't work that way.
| If they did, it would require that your displays retained state
| more than they do today.
| NavinF wrote:
| Besides the dynamic bandwidth allocation, you pretty much
| described how USB 4 and all versions of Thunderbolt works.
|
| Personally I'd hate to see my display bandwidth drop when I
| copy files to a flash drive so I'm glad no monitors support
| that part of your suggestion. Input lag = horrible UX.
| rbanffy wrote:
| > I'd hate to see my display bandwidth drop when I copy files
| to a flash drive
|
| Well... if you connect the flash drive to the monitor and not
| the computer that should be expected. OTOH, I'd love to have
| faster data rates when my monitor image doesn't change much
| (such as when I'm writing code or letting the computer move
| lots of data between external drives.
| layer8 wrote:
| You'd need an RDP-style protocol to do that, and at the cost of
| the increased latency this brings.
|
| And now think about what this will mean when you're actually
| doing RDP on top of it.
| rescbr wrote:
| You have a similar experience like you described above with
| those USB video adapters that don't run DisplayPort. They use
| lossy compression and - as a byproduct - have increased latency
| compared to a DP/HDMI output.
|
| Displays aren't special. It's just that they're moving huge
| amounts of data hundreds of times per second, so they use a
| specialized protocol to do so.
| mritun wrote:
| I upvoted to counteract the downvotes. The question is genuine
| and needs a thoughtful reply.
|
| OP, the tech exists however contrary to expectations, to have
| multiple displays attached with a bandwidth constrained
| connection, the display tends to have all the special bits in
| it (contrast it with your "please stop having displays be
| 'special")
|
| To support no bits moving when image is static, the display
| must incorporate a framebuffer and once you add franebuffer to
| the display, it stops being a dumb display. Eg. You can add
| smarts to it and expose higher level primitives for "display
| acceleration" and reduce the bandwidth required further... and
| very quickly the display is just a computer with memory and
| video accelerator (aka graphics card) connected with a cable.
|
| This is what RDP. Xdisplay and VNC accomplish. The basic
| complexity is not reduced but moved elsewhere. However the
| function gained is very useful so they exist and it's a
| competitive landscape!
| londons_explore wrote:
| > the display must incorporate a framebuffer
|
| I don't think there are any consumer displays sold today that
| don't include a framebuffer...
| kmeisthax wrote:
| Displays don't work this way. LCD panels do not have memory,
| they need to be periodically refreshed at least a few times a
| second[0], and so any sort of intraframe compression capability
| requires having extra memory in the controller and scaler. This
| was never done because displays have _always_ had allocated
| high bandwidth channels since the dawn of console televisions.
| If you needed to get intraframe compressed video into them, you
| plugged in a decoder box of some kind (e.g. your cable box, a
| computer with streaming software, etc) and that device would
| spit out the fully decompressed image.
|
| If you're just using normal HDMI or DisplayPort cables there's
| no bandwidth to share, and since that's the vast majority of
| display inputs[1], we haven't even had _inter_ frame
| compression until very recently when people started wanting
| extremely high resolution and high refresh rate displays. The
| only reason why this seems incongruous now is that we also want
| to shove this video data through a USB-C cable. USB shares
| bandwidth across multiple devices based on demand, so why
| doesn't video over USB? Why don't we just standardize "MPEG
| over DisplayPort" _today_ so we can shove a video wall of 4K
| displays over a single cable? Well, a few reasons...
|
| - Computer users expect lossless reproduction. Most lossy
| codecs only do well on pictures, and absolutely _murder_ text
| and graphics - which is what most computer users are actually
| watching. For various technical reasons involving an Nvidia
| driver bug[2], one of my three displays actually already uses a
| compressed input - specifically Miracast - and text 'twitches'
| every few seconds at every GOP[3] boundary. I hate it.
|
| - Lossy compression codecs add latency. This is not merely an
| artifact of the encoders being slow, some codecs also have
| 'algorithmic latency' - as in you get no output until you give
| it a minimum amount of input because the codec needs data to
| reference so it can remove redundancy.
|
| - Transient loss of USB bandwidth is extremely difficult to
| debug. Just as an example, Windows can't help me track down the
| rogue USB device in my setup that unplugs itself at 3PM sharp
| every day. It just says something about "the last USB device
| you plugged in", as in, "I can't explain what USB topology is
| so I'll throw all that bookkeeping work onto you". Now imagine
| that instead of a minor annoying pop-up, it's a device
| diverting bandwidth away from my display to itself. The OS
| developers aren't even going to flash a pop-up for that,
| they're just going to have the screen glitch out and hope for
| the best, in that I'll blame the display manufacturer rather
| than the computer. Neither party wants to have to deal with a
| tech enthusiast plugging in more displays than their cabling or
| USB topology can handle and not understanding what limits they
| hit.
|
| [0] Judging based off the minimum refresh rates in variable
| refresh rate mode
|
| [1] Desktop systems outright don't support USB-C video altmodes
| _at all_ , outside of special motherboards or add-in cards for
| Thunderbolt that give you DisplayPort injection almost by
| accident. The only monitors that support USB-C video are Apple
| displays; everyone other monitor company ships DisplayPort or
| HDMI. This is why Apple users have to carry around adapters and
| dongles all the time, to the point where Apple actually had to
| _walk back_ their "single cable future" and put HDMI _back in_
| their laptops.
|
| [2] I have three monitors, but only two optical DisplayPort
| cables to go to my other room. I used to use a DisplayPort MST
| hub to get my two smaller (1080p) monitors on the same cable,
| but for some reason my 1080Ti can't read the EDID data off one
| of the monitors when it's behind an MST hub. I worked around
| this with custom resolutions - i.e. manually inputting all the
| display timings - and it worked until a driver update last
| year. Now, if I ever plug in that second monitor to the hub,
| the mouse cursor (which I suspect is a hardware overlay) gets
| stuck on my primary monitor and shows corrupted image data,
| making my computer unusable.
|
| [3] Group of Pictures - the smallest seekable unit in an
| intraframe compressed video stream. Keyframes - i.e. frames
| that do not reference prior frames - are the start of a new GOP
| and take up significantly more bandwidth as a result.
| 3np wrote:
| > The only monitors that support USB-C video are Apple
| displays; everyone other monitor company ships DisplayPort or
| HDMI.
|
| Eizo has mass-marketed USB-C models.
| bdavbdav wrote:
| Samsung, dell...
| fleventynine wrote:
| > Let's remove 20% due to 8b/10 encoding
|
| Thunderbolt 3 uses 64b/66b encoding (unlike display port 1.2 alt
| mode), so there's more bandwidth left over for non-display
| protocols.
| dishsoap wrote:
| came here to comment this
| fabiensanglard wrote:
| Thank you for pointing out this mistake. So the 40 Gbps value
| advertised by Intel is not the pre-encoding rate but post-
| encoding rate. This means TB3 pre-encoding rate is 40.25gbps.
| Minus encoding we get post-encoding = 40Gbps (before TB
| headers).
|
| This leads me to question my USB maths. E.g.: USB 3.2 Gen 1x1
| advertised as 5,000 Mbps, but is that pre-encoding or post-
| encoding?
|
| Is it `5000 Gbps pre-encoding -> 8b/10b > 4000 Gbps` or `6250
| Gbps pre-encoding -> 8b/10b > 5000 Gbps`
|
| Same question for USB 3.2 Gen 2x1 which uses 128b/132b (I
| double checked this time :P!).
| rnantes wrote:
| Excited to upgrade to a Mac with Thunderbolt 5 which should allow
| for 6K at 120Hz with 10 bit color over a single cable. All this
| with an OLED panel is just about peak display for me.(though
| miniLED would be fine too)
| peebeebee wrote:
| Add a nice integrated iPhone quality webcam, and I think it
| would be endgame display. If there is one company that could do
| this, it would be Apple.
|
| Although I have an LG Ergo 4k 32inch at home and I think the
| 16:9 ratio is just not enough for having 2 windows side by
| side. You'd only get 1500pixels per window on retina 6K.
|
| I would've liked it to be a bit wider. Or maybe I'd go for a
| smaller 27 inch, and put my laptop screen as a secondary screen
| next to it.
| tambourine_man wrote:
| That Al Gore setup is really awesome, specially for 2007.
| mistersquid wrote:
| > That Al Gore setup is really awesome, specially for 2007.
|
| Yeah, the OP links to an image of Al Gore in front of 3
| displays with the caption
|
| _Climate central_ Gore in his Nashville home office. where he
| wrote his new book. Mind map software and huge Post-it notes
| help him order his thoughts [0]
|
| I did some medium-searching but couldn't come up with a name
| for the mind mapping software Gore used.
|
| Also interesting is how the 3 displays present in aesthetically
| pleasing way because they are identical (resolution, size,
| bezels, etc).
|
| [0] https://fabiensanglard.net/xdr/al.webp
| gatkinso wrote:
| I have one, its amazing obviously but the backlighting is not
| 100% perfect. Looking forward to the next one of course. I even
| got it to work with my PC via a [Belkin VR
| cable](https://www.belkin.com/support-
| article/?articleNum=316883). Even has USB support.
| knodi wrote:
| I'm really waiting on a new version of XDR to get one.
| huy-nguyen wrote:
| The article states the wrong resolution for the Apple display and
| it's an interesting mistake because these days there are actually
| 2 versions of 6K in consumer-marketed computer monitors: the one
| used by the Apple display (6016x3384) and the slightly larger one
| used by the Dell U3224KB 6K that came out earlier this year (6144
| x 3456). In fact, an interesting thing people found out when they
| use the Dell 6K display on Intel MacBook Pros running Mac OS
| between 10.15 and 13.6 is that the Mac cannot do Display Stream
| Compression at the Dell's native 6144 x 3456, hence the Mac can
| only drive the monitor at 30hz instead of 60hz. However, if they
| can fool the Mac into thinking the display is 6016x3384 (same as
| the Apple display), DSC magically works and they get 60hz on the
| Dell (at the expense of sacrificing some screen real estate).
| Apple must probably hardcode the 6016x3384 resolution somewhere
| in their OS code. Thankfully people report that this problem has
| been fixed as of Mac OS 14.1 but that bug existed for 4 years.
|
| Edit: this problem only seems to happen on Intel, not Apple
| silicon machines.
| sinfulprogeny wrote:
| > that the Mac cannot do Display Stream Compression at the
| Dell's native 6144 x 3456,
|
| Can't, or won't? M1 MacBook pros for some reason can't do 4k120
| over hdmi unless you buy a specific usbc-hdmi adapter and fool
| it into thinking it's displayport (or something like that, I'm
| paraphrasing. You can find info if you search for cablematters
| DDC 4k120 m1.)
| wtallis wrote:
| There's no "fool it into thinking it's displayport". What
| you're describing is having the Mac actually literally emit a
| DisplayPort signal, and a separate device converting that to
| an HDMI signal. The USB-C HDMI Alt mode standard was never
| implemented by any real products, and _all_ USB-C to HDMI
| converters are active adapters that consume DisplayPort
| signals and emit HDMI signals. Not all of those support HDMI
| 2.1, which introduced a drastically different signalling mode
| for HDMI in order to support much higher data rates (and also
| added display stream compression, further increasing the
| maximum resolution and refresh rate capabilities).
| thejazzman wrote:
| You're missing the point -- you have to use custom firmware
| on those adapters or Apple still only puts out 4k60
|
| I went deep on this last night shopping for a cable
| brigade wrote:
| That "some reason" is that a standard DP-to-HDMI 2.1 protocol
| converter can't negotiate beyond HDMI 2.0 link rates without
| the host computer knowing about and doing FRL training on the
| HDMI side. Completely unrelated to any limitations related to
| 6144 x 3456.
| eludwig wrote:
| I can verify that an M1-M3 Mac running Sonoma (14.1.1) the
| U3224KB supports 6144 x 3456 at 30bit (60Hz). Under Ventura
| this did not work. Seems fixed now.
| fabiensanglard wrote:
| Thank you for pointing the mistake. I had no idea, this is
| super interesting.
|
| I have fixed the article and added a footnote to this comment.
| pram wrote:
| It's not the thunderbolt controller totally, the Vega and GCN
| (RX580 etc) Radeon cards on older Macs don't support DSC. The 16
| inch MBP had RDNA cards (which do)
|
| For example the Vega MPX cards on the Mac Pro don't do it, but
| the 5700 etc cards will give you the added port bandwidth.
| bastard_op wrote:
| Apple's got nothing left to offer here, Thunderbolt died with
| their relation to intel. It's a corpse still running.
___________________________________________________________________
(page generated 2023-11-24 23:00 UTC)