The future of Firewire???

[BrianF]

HI guys,

I'm a lurker here, but have been reading for quite awhile and learned many great tips from you all. Anyways, I'm doing a research paper on the future of FW and how it relates to products/recording in the MI industry.

I was hoping some of you might be able to give me a hand. I already understand the basics, and history and big players in the game and who makes great FW products, but maybe some of you have an idea as to what new types of interfaces, or connections you'd like to see, or if you seen recent documentation from Summer NAMM, if FW will be obsolete in a few years being fully replaced by USB 2.0 as the standard.

Are we making any progress with FW devices? is FW800 worth the time and money for companies to invest in? Any newer FW technologies out there that anyone is familiar with? thanks in advance guys!

[soulstudios]

Originally Posted by BrianF

HI guys,

I'm a lurker here, but have been reading for quite awhile and learned many great tips from you all. Anyways, I'm doing a research paper on the future of FW and how it relates to products/recording in the MI industry.

I was hoping some of you might be able to give me a hand. I already understand the basics, and history and big players in the game and who makes great FW products, but maybe some of you have an idea as to what new types of interfaces, or connections you'd like to see, or if you seen recent documentation from Summer NAMM, if FW will be obsolete in a few years being fully replaced by USB 2.0 as the standard.

Are we making any progress with FW devices? is FW800 worth the time and money for companies to invest in? Any newer FW technologies out there that anyone is familiar with? thanks in advance guys!

When even FW400 (1394a) isn't supported properly by Windows XP after sp2 (the microsoft patches don't work - you have to revert your drivers to sp1-level - see details), I have serious doubts about Microsoft's willingness to support firewire, given that they have stakes in the competing technology (usb).
And apparently FW800 (1394b) is broken in Vista. Not that anybody uses that (Vista, I mean - not FW800).
Which is a shame, because firewire is a faster technology - though USB 2.0 has higher 'data-rates', it's effective bandwidth in practice is much smaller.
Cheers,
m@

[BrianF]

Thanks for the input..
I've also heard about that fix, however, there are two sides to that rumor, I've also spoken with those who have said its completely false that sp2 runs a much slower fw bus than sp1.
I havent found a way to verify the issue. so i take it fw800 is practically dead for pro-audio use?

[soulstudios]

Originally Posted by BrianF

Thanks for the input..
I've also heard about that fix, however, there are two sides to that rumor, I've also spoken with those who have said its completely false that sp2 runs a much slower fw bus than sp1.
I havent found a way to verify the issue. so i take it fw800 is practically dead for pro-audio use?

Check out the links at the bottom of here.
It's not false, because Microsoft themselves put out a patch. Only, the patch doesn't work as well as reversion.
I'd trust the word of reputable firewire product manufacturers, such as RME, and the word of microsoft, over whatever your friend told you.

Since noone needs FW800 bandwidth for audio AFAIK, FW is very much alive, unless the standardised base samplerate increases to 192khz - then we'd need more. That's my opinion.
m@

[Reptil]

Originally Posted by BrianF

...snip...
I havent found a way to verify the issue. so i take it fw800 is practically dead for pro-audio use?

dunno about PCs but it works like a charm on macs.

[staudio]

Firewire is a great technology. The main problem people seem to have with it is when they daisy chain too many devices or are using too many devices on a single Firewire bus. Computer manufacturers (namely APPLE) need to start building their computers with more than one Firewire bus in order to really offer all the convenience and speed that the protocol can offer.

Firewire 800 is a big leap in convenience for me over 400, mostly because backups and transfers are faster. I also do notice a slight increase in available track count when using 800.

I want locking connectors in the next revision.

[soulstudios]

Originally Posted by staudio

Firewire 800 is a big leap in convenience for me over 400, mostly because backups and transfers are faster. I also do notice a slight increase in available track count when using 800.

Sounds like you're talking in terms of using firewire hard drives -
Not a good look, if you're wanting speed, the internal computer bus is always far faster. Of course if you're on a laptop you may not have much choice.
m@

[Francis Vaughan]

Firewire does seem to be somewhat marginalised - not for technical reasons - but commercial. History shows that Apple were perhaps too greedy for just a few moments too long, and by making it worth Intel's while to create a competitor they assured its marginalisation.

Firewire has a number of significant technical attributes that make it significantly superior to USB. Support of isochronous channels, peer to peer transfer, much more sensible protocols in general, and so on. These advantages make it better for audio in many ways, and allow it to perform better than USB in real terms. FW400 is noticeably faster than USB 2.0 (at 480Mb/s) due to the significantly greater overheads in USB.

Both are trying to run for the next generation. FW will go to 3.2Gb/s in the current roadmap. USB 3.0 is a bit of an unknown quantity still, it will be backward compatible, use more wires for the higher speed parts, and go to 4.8Gb/s, but also include some protocol changes. But would not appear to include any changes to the basic rationale of operation. So still no FW like performance guarantees or peer to peer operation.

Competitors for FW will be USB for commodity uses, and PCI Express for specialist high performance. The ability of PCIe to bring out a single lane (such as with an ExpressCard) means that some performance critical uses currently the province of FW may migrate. This clearly includes multichannel digital audio converters. Proprietary interfaces (like Apogee's Symphony) will probably be the first to move to PCIe, but FW could find serious competition here. However the inability to daisy chain easily may reduce the uptake.

FW can reticulate over CAT6 cable, and there is some suggestion you could run FW though a building on fixed wiring. Exactly why is another question.

So, the landscape will be USB 3.0, FW 3200, PCIe. Each with advantages and disadvantages. I don't see USB making any inroads into high quality audio, the protocol is basically unsuited, even with the higher speeds. Firewire retains the best fit for purpose here, until you need studio levels of interface - where PCIe will probably take over.

Microsoft has a rotten reputation with FW. Just look back to the SGI Windows workstations. MS promised to support them, and these were one of the first workstations with FW, but MS never did. SGI bet the company on these machines, and MS's lack of support nearly killed them. (As a reward MS hired the SGI CEO.)

[BrianF]

Cool guys! Thanks for the input.

Just to pick your brain further, if we don't see FW800 interfaces to be necessary move in the near future of the audio industry, does anyone foresee any changes in or trends happening within FW400?

It seems that the best thing we can hope for would just be better quality products, with stable drivers, but perhaps there's nothing product wise worth noting since most of the FW audio products will be limited to the same features that the best fw400 interfaces ahve now (8-16 inputs and/or 2-4 ADAT ports, decent work clock). Is FW400 going to be realistic for anything other than audio interfaces and storage (obviously).

But is this going to be enough within say 5-8 years from now?

[d1rtynyc]

Originally Posted by staudio

Computer manufacturers (namely APPLE) need to start building their computers with more than one Firewire bus in order to really offer all the convenience and speed that the protocol can offer..

That is SO true!

[soulstudios]

Originally Posted by Francis Vaughan

Firewire does seem to be somewhat marginalised - not for technical reasons - but commercial. History shows that Apple were perhaps too greedy for just a few moments too long, and by making it worth Intel's while to create a competitor they assured its marginalisation.

Technically, I'd disagree with that - there are major problems with support of firewire on chipsets, to the extent that many manufacturers recommend specific makes (e.g. Texas instruments) to get correct connection to the device-in-question. SO either the spec itself is deficient (not specific enough) or the hardware manufacturers are to blame. At any rate, I've never seen anybody have the same kinds of make-specific problems on USB 2.
Course, I've never had a problem with firewire, so...
m@

[Francis Vaughan]

Sort of chicken and egg here. Apple wanted a per chip tithe for all FW products - at the time when USB was only USB 1.0 and intended as nothing more than a simple low bandwidth bus for things like keyboards and mice. Intel balked at this, and eventually decided to invest in USB 2.0 and to bring about a true competitor. Had Intel not decided to do this they would have swung behind FW and the current issues would probably never happened. When the number of products are fewer, the volumes lower, the money available less, one would assume mistakes occur.

The exact nature of the current problem - TI verses Augere/Lucent is just plain odd. Chips get made with flaws all the time. Usually there are software workarounds. However the nature of the FW arbitration process is such that it is deep in the phy, and it looks as if a software workaround is difficult at best. Whether this was a design flaw, a specification flaw, or something that should have been caught in one of the early testing phases, hard to judge. Such answer will probably never leak out of the companies concerned. But I don't think this has much to do with the lack of FW interfaces as standard on Windows PCs.

USB is intrinsically a much simpler interface - it places much more work onto the host CPU. So software workarounds are much easier to use when the hardware is flawed. But this is part of the difference in design goals. USB won't work without a host computer (USB-to-go simply lets you put the host controller computer in the peripheral - so it can be a surrogate for the host.) Firewire can run host-less, and does not place load on the CPU - being able to DMA end to end. FW is always going to be more complex than USB. Just part of the tradeoffs in design.

[theblotted]

more food for thought on the matter:

FW will always have better continuous thru-put than USB.

...but FW800 in practical application, not as reliable. i don't know how many 800 ports i've seen getting fried in usage.

USB still dominates the market, because Windows/PC machines dominate the market. for example, as ironic as it is, they make a lot of FW & Mac's in Taiwan; but you'll rarely see anybody using the FW protocol. same in the China market, imagine how many USB vs. FW users there are. this is because of relative lack of Mac users over there in Asia.

this can be a concern for mobile musician/engineers.. FW also has a physical size limitation - notice how there are USB thumb drives, but you never see FW thumb drives. this is because you need particular components like the famed TI chips to operate FW. these are too big to fit, so the smallest FW devices are usually 2.5" external HDs...

and somebody mentioned FW400 is enough for audio application. i wonder how many tracks is the poster doing; because there's a reason why location and studio recordings of over 120 channels will never use FW400. instead they'll uses proprietary protocol such as Digilinks, but even at that we're hitting 32 channel limit per cable (but then again format being a priority than thru-put, another can of worms).

protocol such as SATAII and even SCSI/PCI is still recommended as a larger bandwidth stream for ensuring large track counts to Hard Drives.

---

at this point i just want FW3200 to come around so backing up session files at the end of the day will go much quicker

[triez]

The irony is that USB is a Microsoft invention that was completely dead in the water, until Apple brought out the iMac on which USB was the ONLY connection interface. Apple actually launched USB into public acceptance.

USB 2 is no substitute for Firewire for audio, USB sends data in packets while Firewire streams continuously. You can send 72 channels of Audio down a Firewire cable, I've personally never seen more than 4 down USB.

[Francis Vaughan]

Originally Posted by triez

USB sends data in packets while Firewire streams continuously. You can send 72 channels of Audio down a Firewire cable, I've personally never seen more than 4 down USB.

This isn't quite right. FW is packetised. FW runs a cycle clock of 8kHz, and any member of the bus that wants to send something on the bus can request a time slice to do it. Isochronous devices (which all audio devices tend to be) are able to pre-book the slices they need, and thus are always guaraneteed that they can send. So in a sense they stream - but only in that they are pre-booked and do not need to negotiate (and maybe fail) to get the slice needed. This indeed is a huge win over USB which provides no such mechanism. But inside the time slice, a packet is what is sent.

(It is interesting to notice that the 8kHz cycle clock is the same frequency as used on SDH and SONET for long haul data telco level data communications. Much the same problems are being solved in similar ways.)

[JH4music]

I personally hope that firewire bites the dust. I have had nothing but problems with it ever since it first started showing up on macs. I haven't used any FW800 devices yet, so no opinion there.

FW400 uses cheap cables and a bad connector design. A bad cable can cause the data and power wires to short, which usually ruins the audio interface and the computer's firewire port(generally a non-replaceable portion of the logic board).

This article has some good information on this subject:

WiebeTech Micro Storage Solutions -White Papers

Ever since Leopard, it seems like most of the firewire audio interface manufacturers are having nothing but problems getting drivers to work reliably, despite their claims that they are manufacturing their devices to apple's specs.

I think that apple should ship all of their computers with an expresscard slot instead of firewire ports. I really regret buying a 24" imac, I have had nothing but problems trying to get a reliable audio interface to work with it, my only options are FW or USB.

[triez]

Originally Posted by JH4music

I personally hope that firewire bites the dust. I have had nothing but problems with it ever since it first started showing up on macs. I haven't used any FW800 devices yet, so no opinion there.

FW400 uses cheap cables and a bad connector design. A bad cable can cause the data and power wires to short, which usually ruins the audio interface and the computer's firewire port(generally a non-replaceable portion of the logic board).

This article has some good information on this subject:

WiebeTech Micro Storage Solutions -White Papers

Ever since Leopard, it seems like most of the firewire audio interface manufacturers are having nothing but problems getting drivers to work reliably, despite their claims that they are manufacturing their devices to apple's specs.

I think that apple should ship all of their computers with an expresscard slot instead of firewire ports. I really regret buying a 24" imac, I have had nothing but problems trying to get a reliable audio interface to work with it, my only options are FW or USB.

I am using an an Apogee Ensemble with my 24' iMac and it works perfectly without any issues. I will take your advice to heart though and not do any hot plugging of Firewire, small price to pay for such a great system.

[triez]

Originally Posted by Francis Vaughan

This isn't quite right. FW is packetised. FW runs a cycle clock of 8kHz, and any member of the bus that wants to send something on the bus can request a time slice to do it. Isochronous devices (which all audio devices tend to be) are able to pre-book the slices they need, and thus are always guaraneteed that they can send. So in a sense they stream - but only in that they are pre-booked and do not need to negotiate (and maybe fail) to get the slice needed. This indeed is a huge win over USB which provides no such mechanism. But inside the time slice, a packet is what is sent.

(It is interesting to notice that the 8kHz cycle clock is the same frequency as used on SDH and SONET for long haul data telco level data communications. Much the same problems are being solved in similar ways.)

I didn't know that.... But how does a system with an 8 meg clock shift data at 400Mb per second?

[triez]

Originally Posted by
This article has some good information on this subject:

[url=http://www.wiebetech.com/whitepapers/FireWirePortFailures.php

WiebeTech Micro Storage Solutions -White Papers[/url]

Ever since Leopard, it seems like most of the firewire audio interface manufacturers are having nothing but problems getting drivers to work reliably, despite their claims that they are manufacturing their devices to apple's specs.

I think that apple should ship all of their computers with an expresscard slot instead of firewire ports. I really regret buying a 24" imac, I have had nothing but problems trying to get a reliable audio interface to work with it, my only options are FW or USB.

I read the White Paper you linked to, it says that FW port failure is not too common... From reading it through it would seem that by ensuring that you have good quality FW cables and avoid dodgy things such as no-name Chinese enclosures you are unlikely to have a problem.

I hope you get your issues sorted out, I use the Ensemble via FW on an iMac, a G5 tower and a G4 laptop with no problems at all. If FW dies as you wish then you had better get used to USB audio interfaces with 4 channels, or buying desktop computers with PCI slots.

[Francis Vaughan]

Originally Posted by triez

I didn't know that.... But how does a system with an 8 meg clock shift data at 400Mb per second?

It it an 8kHz (not MHz) clock - but this isn't the data clock - it is the clock that starts a new bus cycle. Clearly the data rate is much higher, and inside any 125 microsecond period a lot of data transfers can occur, and a lot of data moved.

What is also interesting is that the 8kHz clock is retained across all FW speeds - but of course with higher speeds the data rate is higher.

When the bus reconfigures a cycle master is selected - and it can be any device that wants to be the master. This allows FW audio converters to request that they become the sample rate master clock. The cycle clock can then be used as a clock distribution system to any other converter as well. But the ability of a FW audio device to retain the sample clock inside itself is a big win for audio quality.