Advantages of Thunderbolt/USB3 audio interfaces

[risome]

I see problems rather than solutions here. There arnt any audio interfaces on either yet so its waaaaaay to soon to judge anything on them. Maybe in a few years. Firewire is all but dead, whats left isnt compatible. USB is doing ok.

Firewire all but dead who are you kidding bro.Every decent studio i have been in including my own use firewire audio interfaces

If you buy a new laptop/PC, youll have trouble with Firewire. Its being faded out. The remaining interfaces arnt the required Texas Instruments type.

I bought a MacBook pro as a machine for my DJ performance recently and i dont have any trouble with firewire

[penguinfromdeep]

+1

[tehlord]

Most importantly if you're mixing VST's ITB it'll make no difference whatsoever.














Probably.

[UltraJv]

I see problems rather than solutions here. There arnt any audio interfaces on either yet so its waaaaaay to soon to judge anything on them. Maybe in a few years. Firewire is all but dead, whats left isnt compatible. USB is doing ok.

Firewire all but dead who are you kidding bro.Every decent studio i have been in including my own use firewire audio interfaces

If you buy a new laptop/PC, youll have trouble with Firewire. Its being faded out. The remaining interfaces arnt the required Texas Instruments type.

I bought a MacBook pro as a machine for my DJ performance recently and i dont have any trouble with firewire

https://discussions.apple.com/thread/37 ... 0&tstart=0

[siriusbliss]

UAD charges for a big bloated Thunderbolt card while Apple drives TB with a single-chip onboard solution.
I doubt there's any other conversion going on except maybe they're using PCIe protocol, and most of the board is power components for driving the TB pipeline.
I don't really know, but that's what it looked like when I checked it out at NAMM.
I'd rather see a PCIe-to-Thunderbolt conversion box for my Soniccore Scope system where I can then also daisy-chain off an external drive or two for archiving or library fly-ins. TB's advantage is that you will be able to daisy chain TB hardware off one cable chain.

Nevertheless the future of Thunderbolt will be optical, which will be even faster than USB3 anyways, so there's at least another five years of evolution there.

G

[Throbert]

Present TB is NOT full speed and what Intel intended to release and will next yr. thats optical based Lightpeak (the real name for TB)
many devs are holding out for the real deal..
lightpeak just adds yet more bandwidth.

TB/LP i think finally brings the ability for a more "lego" based growth

Scott
ADK

+1

and FYI firewire is NOT dead yet...

Scott
ADK

Laptop TI firewire and PCI slots seem to be vanishing and you have to find a card for desktops with non TI firewire.

[jupiter8]

Is it just me or is it slightly ironic that people hail the Apollo as a good example why FW is dead and TB is the future when in fact it does have a FW port but do NOT have a TB port ?

[hibidy]

For the record I am not implying that anything is dead.

I was under the impression that drivers could be written for soundcards that could handle lower buffers with the new devices. That seems to be debatable.

I have no interest in apollo other than unless it is a superior sounding device. Frankly, I can see many issues arising (not unlike the UAD cards in some peoples systems).

I simply provided the link and the currently known information. (including that you you must buy an optional card for the card)

[golemus]

Thunderbolt: lets be clear on a few things here.
in of itself it CANT offer any lower latency than we already have for external or internal (32 buffer)
it can offer more bandwidth or sharing ability.

Are you sure about that it cant offere lower latency than current external soundcards....?

As far as I have understood the worst problem with USB (2.0) is "speak when sponen to" architecture or device polling. Here are some quotes from Wikipedia and other places:

USB 1.0, 1.1 and 2.0 use a "speak-when-spoken-to" protocol; peripherals cannot communicate with the host unless the host specifically requests communication. USB 3.0 allows for device-initiated communications towards the host. A FireWire device can communicate with any other node at any time, subject to network conditions.

Polling mode is replaced by asynchronous notices.

Continuous device polling is eliminated



http://en.wikipedia.org/wiki/Universal_Serial_Bus
http://www.everythingusb.com/superspeed ... provements






So my question is that these things 1: "USB3 allows device-initiated communications towards the host", 2: "polling mode is replaced by asynchronous notices", 3: "continouus device polling is eliminated". Will they bring USB3 to the same level with firewire in reliability of transfering audio or video streams...?

As I have understood USB2 audio interfaces can be reliable as far as your CPU doesn't approach 100%. When it is almost 100% it is much more likely for USB2 to have glitches than for firewire.

In case somebody haven't noticed majority of PCs that come to sale after April will have USB3 as it is now integrated to new intel chipsets. But only few of them have thunderbolt. In addition to Apple machines I have found only 4 thunderbolt devices, Lenovo Thinkpad Edge S430, Acer Aspire S5, MSI Z77 motherboard and some Intel motherboard.

[jcschild]

i am 100% sure that you CAN NOT get below 32 buffer. IF PCIe cards dont have it
(thats what TB goes thru is PCIe) then no.

at 32 buffer you are pretty much at whats called "real system time"
to get any lower you would need a dedicated OS and non interupts
(another words 100% proprietary system) and then it would not be much less latency. besides who needs less than 64 or 32 buffer? and why?

look at it this way max speed thru a pipe is say 200 MPH

thru a 1" pipe you get 500Gal per hour
thru a 2" pipe double that.
thru an 8" pipe 8x that.

note never is the water moving any faster just a whole lot more of it per second.

this is what TB does compared to USB/FW..

as far as the USB/FW protcols again its pointless
RME can already do 48 buffer on either FW or USB2..

its not any different when PCIe first came out.
there was absolutely no difference in benchmarks from PCI to PCIe
no lower buffer etc.
what it gave was again more bandwidth. (bigger pipe)

finally yrs later we now have this card
RME HDSPe MADI FX which is 192 i/o (@48k)
on a single card where it once took 3.
no lower latency just more I/O

Scott
ADK

[jupiter8]

i am 100% sure that you CAN NOT get below 32 buffer. IF PCIe cards dont have it
(thats what TB goes thru is PCIe) then no.

at 32 buffer you are pretty much at whats called "real system time"
to get any lower you would need a dedicated OS and non interupts
(another words 100% proprietary system) and then it would not be much less latency. besides who needs less than 64 or 32 buffer? and why?

look at it this way max speed thru a pipe is say 200 MPH

thru a 1" pipe you get 500Gal per hour
thru a 2" pipe double that.
thru an 8" pipe 8x that.

note never is the water moving any faster just a whole lot more of it per second.

this is what TB does compared to USB/FW..

as far as the USB/FW protcols again its pointless
RME can already do 48 buffer on either FW or USB2..

its not any different when PCIe first came out.
there was absolutely no difference in benchmarks from PCI to PCIe
no lower buffer etc.
what it gave was again more bandwidth. (bigger pipe)

finally yrs later we now have this card
RME HDSPe MADI FX which is 192 i/o (@48k)
on a single card where it once took 3.
no lower latency just more I/O

Scott
ADK

In this video he claims to be running the MADI FX with a 14 sample buffer.
Had to double check if i remembered that wrong but he's quite clear about it (plus you see it on the screen for brief moment).

[hibidy]

My god, that is one of the worst looking shirts I've ever seen.

[jcschild]

checking on that 14 buffer, Funny Matthias (guy in the video owner of RME) is who i had the lower than 32 buffer discussion some yrs back

[jcschild]

DOH totally spaced it was an Apple system apple buffering (core audio) and Windows buffering are different

just got confirmation..
in Windows the hard limit is 32 buffer (which is NOT to say Apple does lower buffer)
32 buffer is 1MS
48 samples at 48kHz = 1ms pretty much as low as you can go...

scott
ADK

[golemus]

i am 100% sure that you CAN NOT get below 32 buffer. IF PCIe cards dont have it
(thats what TB goes thru is PCIe) then no.

at 32 buffer you are pretty much at whats called "real system time"
to get any lower you would need a dedicated OS and non interupts
(another words 100% proprietary system) and then it would not be much less latency. besides who needs less than 64 or 32 buffer? and why?

First of all I think it is better to talk in terms of ms rather than buffer size because that is what really matters, how long it takes from pressing enter to the audio to be hearable. In 44100 sampling rate 32 buffer corresponds to 0.73ms of latency. When I set 32 buffer and 44100 sample rate in both of my newest audio interfaces (NI Traktor Audio 2 and NI Traktor S2) which are supposed to be one of the best of budget class interfaces I get total latency of 4.6ms with Traktor Audio 2 and 5.1ms with Traktor Control S2.

So what causes the rest of the latency and would this "the rest" be smaller with USB3 or thunderbolt interface? I really don't know what latency (total latency in ms, not just buffer size) is achievable with todays PCIe soundcards as the last internal soundcard I used actively was Creamware Pulsar from 1998.

And I must also say that I never got 32 buffer to work in any of my computers or any of my audio interfaces so I am really curious how you are able to do it (without using PCIe card or investing to RME). It really doesn't matter to me now, 256 buffer is enough for me currently although it would be nice to get even 128 to work, but I don't know if I start to play MIDI drums in the future then I might prefer a faster latency.



What kind of total latencies are achievable with current PCIe cards...? I mean the affordable ones, not some 5000eur monster...