The No.1 Website for Pro Audio
 Search This Thread  Search This Forum  Search Reviews  Search Gear Database  Search Gear for sale  Search Gearslutz Go Advanced
Do we really need computers? Audio Interfaces
Old 18th September 2011
  #61
Lives for gear
 

Quote:
Originally Posted by popmann View Post
Read it every **** day STILL...right here. You can read how all digital EQs sound exactly the same...to every digital mixer nulling itself. Digital=digital. ....
I don't think that you've seen me say it, and it is about as true as the other popular ideas, like all amplifiers sound the same, or all $100 Chinese microphones equal or better the sound of all $10,000 German or Austrian microphones, or all German/Chinese $250 converters equal or better all other converters at any price. In the rest of the world, the guys who buy Civics are happy to have an affordable solution that does the job. They don't pretend that their Civic equals a Porsche. Here in the audio world, we're not happy unless we can say that our $1.99 solution is better than the $10,000 that someone else has.
Old 18th September 2011
  #62
Gear Nut
 

Quote:
Originally Posted by popmann View Post
A computer is a LOUSY place to record music with fidelity.
Sounds like you want a Studer then. All these other little boxes you're talking about ARE computers.
Quote:
Compared to analog it's mixer, EQ, and compression sounds bad (though some better than others)-and I don't mean a vintage Neve, although, I dod think in terms of revolution, that is one of the gold standards. I'm saying for this point, that it can be anything but the crappiest analog (and sometimes, honestly, the crappiest)...it's tracking latency is unforgivable compared to analog (see crappiest--a $100 Behringer line mixer handles my tracking feed now for my $4000 computer and my $3000 in preamps/tracking EQ),
Gross.
Quote:
and even ten year old digital hardware...plug in reverb is JUST barely getting to where it TOUCHES 10 year old digital hardware.
Which reverb plugins are you talking about anyway? IMO, hardware reverbs have all bit died for everything but live sound because convolution reverbs just sound soooo much better than anything else. Waves IR1 is what killed the rack reverb, because it so good.
Quote:
...and the significance of 10 years...BTW--is that's when most things stopped being developed for the home/small studio in digital hardware. It might actually be 8...but, somewhere in there...saw the last of the 56bit fixed all in ones (Akai DPS24)...
You should know that every x86 based 32 bit float audio engine actually has an 80 bit accumulator. That's like 500db of DR before it gets trimmed to the plain ol' 32-bit 192db.
Quote:
In fact, even in SOFTWARE...Software samplers played HUGE catch up with Gigastudio for YEARS--and in ways still never caught it, but in others surpassed it.
How do you mean?
Quote:

At this point...had software not distracted the industry...we should have perfect sample/modeled keyboards of everything that came before--price really just determining the feel and construction of the board itself. We should have a unit that is a 96 (if not 192k) throughout at 48 tracks--that has the fidelity of an SSL with a wall of modeled outboard.
Too many points to argue - but where are you planning on storing your 192khz 48 tracks? You'll need to hire a crew just to handle all the mic-stands too
Old 18th September 2011
  #63
I think, beginning in the 90s and continuing through much of this decade there has been a very large potential market for standalone recorders, a number of very feature-rich machines have been designed and brought to market -- at ever lower prices and always with more features. Such devices definitely appeal to those who might not feel comfortable working on a modular, computer-centric system.

That such devices remain a niche product probably results from a number of factors, but I suspect that the decision process for many is similar to my own: while I may admire the product engineering that can go into such a heavily designed product, having tinkered with them, I personally find the modular, PC-centric system to be not just far more flexible, but, for me, easier to do what I need to do. Before and after -- and even to some extent while -- I was using ADATs, I did a lot of editing on my projects. The ADAT work model made that inconvenient-to-impossible (there is no way to get backwards audio on an ADAT, for instance, something I used lots before and after my ADAT period) -- so when I created my first 8 channel DAW in '97, I went a little edit-crazy. But, you know, it was a continuation and extension of what I'd wanted to do even in my analog tape days.

Editing on the tiny built-in screens of table top recorders -- or even add-on video monitors -- is not nearly as easy as it is on a modern DAW. The editing features of my DAW, Sonar, are great. From the way a clip turns transparent if you click and move it over another clip -- making precision edits by lining up wave transients, etc, trivially easy -- to the way it automatically avoids abrupt signal transitions [cutting on a non-zero value], to the ability to automate not just track volumes but clip volumes or to have a completely separate FX setup on different clips in the same track, even as all the clips are still controlled by the track FX and automation, as well... I mean, show me a dedicated system with the flexibility of a modern DAW and I'll be impressed.
Old 18th September 2011
  #64
Gear Guru
 
UnderTow's Avatar
Quote:
Originally Posted by popmann View Post
Did anyone else read my posts as "An Akai DPS24 alone" is a better solution than a computer full of third party algorithms?
My apologies if I misunderstood you.

Quote:
The problem was they stopped developing (likely an expensive proposition) and the two big machines- the VS2480 and AkaiDPS24 kept getting new drive sizes...or the 2480 got a DVDR for backups...during product refresh cycles. Duh...and they wonder why people stopped buying. That's what I mean by no meaningful refresh. When you as a (digital) hardware company keep putting out old designs, of course you'll lose people--and you should.
I don't believe they had any choice. They could increase the specs and features but that would just make them even more expensive. Also hardware development costs are always higher than software development costs.

Let's not forget that the whole software DAW revolution is simply hitching a ride on the massive power growth of general purpose computers. Those computers are not being developed for audio. Developing dedicated hardware solutions for audio can not compete cost wise with the kind of R&D might that Intel and AMD yield.

The market was not distracted by software. The market just couldn't/can't resist the price of PC's.

Quote:
So, when we have this discussion today, when I say "hardware" people look at what a 10 year old design can do in terms of featureset, and compare it to a modern computer. Not fair. Compare it to you G3/233 with 256mb, which was the software choice then. Maybe a P2 300 with 128mb?
Well rather Intel PIII's and AMD Athalon XP's running at 1.53 GHz but yes the Mac side wasn't as lucky. ;-)

Quote:
That's your apples to apples comparison of hardware versus software. You want to work on THOSE machines instead of that hardware?
That isn't the right comparison IMO. What counts is what is available today and what might have been available in all-in-one solutions today. The all-in-one solution's power would not have scaled as fast as the native solutions. They were already not scaling back then (which is why they lost their market share). Now nearly all DSP solutions have been abandoned for native solutions except the very highest stuff like PT HD, Fairlight CC-1 etc.

We can like it or not but there isn't much we can do except demand high standards from the software developers. For instance I've stopped using Sonar because the audio engine gaps and stutters when you perform basic tasks. IMO that is absolutely unacceptable from a modern DAW. That company has it's priorities all wrong for my needs and wants.

Voting with our wallets (and complaining) is all we can really do.

Have you tried PreSonus Studio 1? By your description of what you want and need and it's reputation for rock solid performance and efficient engine, combined with a low-latency interface like RME's offerings, it might make you happy. :-)

Alistair
Old 18th September 2011
  #65
Gear Guru
 
UnderTow's Avatar
Quote:
Originally Posted by popmann View Post
Since you want to be literal...why does a relatively cheap analog mixer sound BETTER?
Because you want it to? ;-)

Quote:
I get that you're being an a$$ with the literal meaning of fidelity...is there a language barrier here?

You know what I mean.
I am not being an ass and no I do not know what you mean. If you don't mean to say what you wrote, don't write it that way!

Quote:
You may not agree with it, since it's subjective...
No it is not. Fidelity means something very specific and what I wrote is absolutely correct. If you mean that you personally prefer the sound of analogue distortion, then just write that instead of writing something that is absolutely not true.

And considering the context, where you are arguing in favour of all-in-one hardware solutions, your comments make no sense. Any DAW already beats the fidelity of any analogue console. An all-in-one hardware solution isn't going to beat it any more. Perfect summing all they way beyond human audibility can not be beaten.

Quote:
but, it's a stupid cop out to claim it's now perfect when I'd venture everyone here...or at least an overwhelming majority prefers a mix done on a big desk. Most would prefer things cut to 2".
Not me thank you. I don't actually like analogue mixers! Including high-end ones!

Alistair
Old 18th September 2011
  #66
Gear Guru
 
UnderTow's Avatar
Quote:
Originally Posted by DaveDaveDave View Post
You should know that every x86 based 32 bit float audio engine actually has an 80 bit accumulator. That's like 500db of DR before it gets trimmed to the plain ol' 32-bit 192db.
32 bit float has 1536 dB of dynamic range. Anyway that Akai didn't have a 56 bit mixer. It had four 56 bit bus effects! That doesn't mean the mixer was running at 56 bits. (And if it did, that would have been a bad engineering design right there).

Alistair
Old 18th September 2011
  #67
also alot of those all in one solutions don't and can't do all of what they say they do well without exploding the price beyond the consumers wallet.


the aw1600 had quite a few options it came with. it had a sampler, some mastering options, e.q. effects and things, were they any good? not particularly. i almost always bypassed those functions for outboard solutions that performed much better with alot of flexibility.

the sampler, was a quick loop, and to be honest i never even tried it. i think i read somewhere in the manual the sample time and it was useless, but i may be wrong.

the effects were o.k. but flipping through the menu's and actually controlling those effects was at times a pain.

the e.q. was pretty limited from what i remember.

the mastering options were all named something really stupid and didn't actually explain what it was doing.

for a system to do all in one operation at a quality competitive to the computer without being a computer would be mighty expensive and probably not possible.
Old 18th September 2011
  #68
Gear Nut
 

Quote:
Originally Posted by popmann View Post
You know what the most "missing the point" argument that always comes up here is? "anything digital is a computer". Which has already been said here in less broad terms.
The math that makes the sound is the same on a custom Korg DSP, an SSE enabled Intel CPU, a GPU running Cuda code, a TigerSHARC...etc,etc. The only difference is cost and latency. Cost matters a lot to most folks, and latency matters as much to me as it does to you, I think. As for the form factor of dedicated hardware or generic PC's, follow the money: who will pay for this super duper hardware that brings the latency gap under 1.6ms?

Quote:

My issue with recording is that the ONLY thing a computer has "advanced" in a decade is:

Content editing capability.
Track count
Sample rate Rez as it relates to track count
I think you're selling the PC based DAW short here. How about automatic delay compensation, including outboard gear? Bouncing mixes straight to market on the internet? Networked audio an multimachine processing like VST Bridge and Logic Nodes?

Quote:
Very soon we should see the idea that you record and mix in a consumer OS as "Luddite" behavior.
So you want to see more TASCAM X-48 style machines in the world?
So, will this specialized OS be a tweaked OS or something from scratch?
Old 18th September 2011
  #69
Lives for gear
 
popmann's Avatar
Sorry...I forgot where I was for moment.

Disregard.
Old 18th September 2011
  #70
Lives for gear
 

This thread has been muddied up with 2 different conversations.

The OP original post (I assume) was motivated by a frustration and hassle with computers. Probably the typical complaints such as DAW complexity, endless software updates, incompatibilities, form factor, etc. This is one distinct conversation.

The other conversation (put forth by popmann) is the issue of sound quality. (E.g. the Akai DPS24 + outboard gear is sonically superior to today's computer DAWs.) This is a separate conversation that's not related to the OP's question at all. On top of that, this is one of those subjective debates that would never get resolved.

As for popmann's thesis: all I can say is that we're not all a bunch of clueless producers that have no idea what those old all-in-one tabletop recorders could do. We lived through that era. Some of us actually bought them. We know what they sound like and what their quality level is. IMO, the computer DAW (even one 10 years ago with Sonar & Mackie preamps & Lucid ADDA) sounded better. It's a subjective opinion, as I said earlier.
Old 18th September 2011
  #71
Lives for gear
 
sam c's Avatar
 

Quote:
Originally Posted by popmann View Post
Sorry...I forgot where I was for moment.

Disregard.
You seem like a very unhappy person....
Old 18th September 2011
  #72
Lives for gear
 

Quote:
Originally Posted by Jason West View Post
The OP original post (I assume) was motivated by a frustration and hassle with computers. Probably the typical complaints such as DAW complexity, endless software updates, incompatibilities, form factor, etc.
This is only an issue when the user can't leave the damned thing alone. Most pros who depend upon their machines to make a living don't make even minor changes without considering the ramifications. The machine that I brought from my studio is three years old or so, and it had no changes other than the normal updates to the audio application, with the exception of an aborted attempt to install a drum software, which was later removed. (Then this year, I messed up.)

Brad Blackwood (four recordings in the pop charts today) is still using version 7 of the same app I use, and I'm on version 11.5. (THAT'S being careful! ...but I'm talking to him......) Build a machine, set it up, and forget about adding the latest and greatest freeware, cheapware, demo, or even drivers unless you KNOW that you need them and don't try to be the first guy on the block, let someone else cry a while. If it ain't broke, don't fix it!
Old 18th September 2011
  #73
Lives for gear
 
dotl's Avatar
 

Do we really need cars?
Old 18th September 2011
  #74
Gear Maniac
 

"Do we really need computers?"

I tried feeding my handwritten score to my hypothetical 16 track Tascam machine with a cd burner and no sound came out.
Old 18th September 2011
  #75
Lives for gear
 

Quote:
Originally Posted by dotl View Post
Do we really need cars?
No we don't. The only reason we buy cars instead of horses is because of the horseshoe manufacturers were too slow to adopt space-age composite plastics or nanotechnolkogy carbon fiber. The saddle manufacturers also didn't retrofit the harness with large cupholders to accomodate the 64oz big gulp supersize drinks from McDonald's and Burger King.

Today's car drivers have tunnel vision with their technology because they've forgotten that horses can handle uneven terrain and move across unpaved roads much better than cars. As a bonus, they can also wade through 3 feet of water whereas most cars would stall unless it was a Humvee or jacked-up monster truck.

Lastly, you look really cool when you ride it. Was the Marlboro Man riding a horse or driving a Honda Civic when he was advertising cigarettes?
Old 18th September 2011
  #76
Gear Guru
 
UnderTow's Avatar
You forgot to mention that cars are only for the talentless or those driving the talentless! ;-)

Alistair
Old 18th September 2011
  #77
Lives for gear
 
TAFKAT's Avatar
 

Quote:
Originally Posted by Jason West View Post
The one area that dedicated hardware really shows its advantage is near-zero latency capability. If you want or need latency of less than 1.5ms ... dedicated DSP chips such as Metric Halo or the cue mix architecture on Pro Tools Native/TDM can deliver that type of performance. Even the next 10 years of progress on general purpose computers with liquid-nitrogen cooled 4.0GHz 50-cores may not be able to match that low latency.
Not really getting your point here.

Anything that has been capable of ASIO Direct Monitoring has had the same Low Latency Monitoring capability you speak of even before the current tech of the DSP powered monitoring mixers that most if not all audio interfaces now employ. I have been directly monitoring straight from AD thru to DA since around 1997/98 with Echo/RME cards.

Direct Monitoring Round Trip Latency on current systems with high quality interfaces with tight AD/DA converters can be as low as 28 samples ( .63 ms ) at 44.1 , we don't have to wait 10 years, its already here.

Not sure what the vanity FX via the DSP on some of the current interfaces is adding in sample latency , but monitoring via FX Plugins which adds the latency of the sample buffer/safety buffers will be above the listed 1.5 ms , even for the best of the best. However when placed in context of a working environment, its more than acceptable and workable IMO.

Old 18th September 2011
  #78
Lives for gear
 

Quote:
Originally Posted by TAFKAT View Post
Direct Monitoring Round Trip Latency on current systems with high quality interfaces with tight AD/DA converters can be as low as 28 samples ( .63 ms ) at 44.1 , we don't have to wait 10 years, its already here.
What level of workload stress can you put on the box and still have 0.63ms latency? Also, about year did technology evolve to deliver that latency spec?

The scenarios I'm thinking of is fx for vocals without comb filtering, and guitar amp sims with tactile responsiveness indistinguishable from a real amp.

For example, with vocals, the Metric Halo guys have DSP effects within the unit so the latency can stay low. It negates any issue with high buffer sizes (128+ samples) on the host side to handle cue mixes with heavy workloads. RME TotalMix doesn't include fx so I wasn't thinking of that.

As for guitar amp modelers, the latency gap is one of the reasons why a product like Fractal Axe-Fx can exist. Yes, that box has very good algorithms, but it's also got lower latency than any computer with Native Ins Guitar Rig, IK Multimedia Amplitube, Waves GTR, etc. I can't prove this but I don't think Fractal's mathematics kung fu (algorithms) is any better than the math wizardry at those other 3 software companies. If his mathematics is truly better, it won't remain that way for long because it's almost impossible to stay ahead of the competition on that aspect alone. Fractal has to work on within TigerSharc chip's 4 GFLOPS while a modern Intel i7 has around 100 GFLOPS. The remaining hardware advantage is the low latency which contributes to the "feel" of the box and therefore the overall perception of sound quality. I think it's very possible that a dedicated hardware guitar amp sim will keep this advantage for the next 10 years. Maybe a future Thunderbolt will change this equation. I don't know.

I've run guitar amp sims through my Apogee Symphony (supposedly one of the lowest latency interfaces). However, its latency (and therefore playing responsiveness) is worse than Fractal. UnderTow mentioned that the HD OMNI is very fast. Maybe someone can comment about running amp sims through it in comparison with Fractal.
Old 18th September 2011
  #79
Lives for gear
 
Sounddesigner's Avatar
 

Quote:
Originally Posted by TAFKAT View Post
Not really getting your point here.

Anything that has been capable of ASIO Direct Monitoring has had the same Low Latency Monitoring capability you speak of even before the current tech of the DSP powered monitoring mixers that most if not all audio interfaces now employ. I have been directly monitoring straight from AD thru to DA since around 1997/98 with Echo/RME cards.

Direct Monitoring Round Trip Latency on current systems with high quality interfaces with tight AD/DA converters can be as low as 28 samples ( .63 ms ) at 44.1 , we don't have to wait 10 years, its already here.

Not sure what the vanity FX via the DSP on some of the current interfaces is adding in sample latency , but monitoring via FX Plugins which adds the latency of the sample buffer/safety buffers will be above the listed 1.5 ms , even for the best of the best. However when placed in context of a working environment, its more than acceptable and workable IMO.

Actually you've only made his point. Those interfaces you refer to with direct monitoring e.g. RME, MOTU, etc are using dedicated dsp because of the flaws of the computer. His point was that dsp is needed for that direct monitoring since that's computers flawed area. Those interfaces are considered Native but the moment you use Cue mix, Totalmix, etc it is by definition dsp processing and it illustrates the need for the dsp/hybrid system due to the fundamental flaws of computers.

The RME, MOTU, etc direct monitoring enviroments ARE THE EXACT same concepts as Metric Halo and Sonic Core SCOPE (wich are labeled dsp platforms) just simpler versions that meet far fewer needs, and those simpler enviroments are less desired cause there ain't much to them. So his point still stands
"The one area that dedicated hardware really shows its advantage is near-zero latency capability. If you want or need latency of less than 1.5ms ... dedicated DSP chips such as Metric Halo or the cue mix architecture on Pro Tools Native/TDM can deliver that type of performance. Even the next 10 years of progress on general purpose computers with liquid-nitrogen cooled 4.0GHz 50-cores may not be able to match that low latency."

Computers can't get as low of latency as some dedicated dsp systems neither can they remain low regardless of load like a dedicated dsp system. Neither will you harness your computers power to a greater extent without dsp's (multi-client drivers, etc).

EDITED
Old 18th September 2011
  #80
Quote:
Originally Posted by Jason West View Post
[...] guitar amp sims with tactile responsiveness indistinguishable from a real amp.

[...]
I don't get that from dedicated DSP boxes like the POD XT one of my pals left over here a few years back. I mean, on the clean settings, it's not bad, pretty easy to adjust to. But stick some FX in that chain and all of a sudden you're talking about sticky fingers latency, a real playing in mud feel. Of course, still a lot better than through-the-box sims via my FW box.

But even if time isn't an issue (and it always seems to be for me -- hard enough for me to play as it is, I guess heh ) the lack of feel is always notable, even if the sound of a given patch is intriguing.

For that matter, it's struck me in recent thinking that the extremely dynamic impedance relationship of a passive guitar p/u or p/u matrix and a conventional amp is not likely to be mirrored by the input section of the guitar DSP. Maybe I'm wrong. But I have to say that the thing that really 'un-impresses' me about the POD XT (admittedly, not a recent box) is the lack of feel, the sense that I'm playing into something that is not responding to me or my dynamics.
Old 19th September 2011
  #81
Lives for gear
 
TAFKAT's Avatar
 

Quote:
Originally Posted by Jason West View Post
What level of workload stress can you put on the box and still have 0.63ms latency? Also, about year did technology evolve to deliver that latency spec?
That latency value is achieved independent of the sample buffer setting, so you can place what ever load the system is capable of at the respective buffer settings.

Quote:
For example, with vocals, the Metric Halo guys have DSP effects within the unit so the latency can stay low. It negates any issue with high buffer sizes (128+ samples) on the host side to handle cue mixes with heavy workloads. RME TotalMix doesn't include fx so I wasn't thinking of that.
Thats what I referred to as Vanity FX , not entirely sure what latency the DSP is adding there, but the on-board DSP is not a priority for many, of course YMMV.

Quote:
As for guitar amp modelers, the latency gap is one of the reasons why a product like Fractal Axe-Fx can exist.
Can't really comment about the unit, but you are drifting off topic as the unit is a dedicated unit for one thing and one thing only , not really that relevant to the current discussion IMO.

Quote:
Maybe a future Thunderbolt will change this equation. I don't know.
Why would it ?

Thunderbolt is simply external PCIe , the current PCIe interfaces are not exhausting the bandwidth of PCIe , why would Thunderbolt improve anything in regards to latency , its simply a more convenient interconnect ?

Quote:
I've run guitar amp sims through my Apogee Symphony (supposedly one of the lowest latency interfaces). However, its latency (and therefore playing responsiveness) is worse than Fractal.
What is the actual RTL of the Fractal and I'll be able to answer that very easily for you.

Also what do you believe is the highest latency that is acceptable for real time playing ?

I have volumes of information in this exact area on my other thread here regarding audio interface low latency performance, based on Win7 but the principle will be the same. RTL is the sample buffer/safety buffer + the units AD/DA .


Quote:
Originally Posted by Sounddesigner
Actually you've only made his point.
No I haven't !

His first statement and the main point was that the advantage of dedicated hardware ( as opposed to computers ) is the low latency monitoring.

My point was that Low Latency Monitoring has been available on computer based DAW's for well over a decade , whether that is achieved via a dedicated DSP on the interface is irreverent , the point stands that the supposed advantage of the dedicated units over computers is not valid IMO !

BTW: I don't need you to school me in what direct monitoring environments are, there wasn't anything in what I posted where I was confused !

Quote:
Computers can't get as low of latency as some dedicated dsp systems neither can they remain low regardless of load like a dedicated dsp system. Neither will you harness your computers power to a greater extent without dsp's (multi-client drivers, etc).
What DSP system are you referring to ?

For Native DAW's - DSP Hybrid solutions using the UAD2 for example are in fact detrimental to Low Latency Performance where the added arbitration when using higher loads of the DSP cards themselves impose substantial stress on the potential native overhead available. The whole premise of being able to use the DSP to alleviate the loads off the CPU's to reserve them for native processing doesn't hold true when running at low to moderate latency
Old 19th September 2011
  #82
Lives for gear
 
Sounddesigner's Avatar
 

Quote:
Originally Posted by TAFKAT View Post
No I haven't !

His first statement and the main point was that the advantage of dedicated hardware ( as opposed to computers ) is the low latency monitoring.

My point was that Low Latency Monitoring has been available on computer based DAW's for well over a decade , whether that is achieved via a dedicated DSP on the interface is irreverent , the point stands that the supposed advantage of the dedicated units over computers is not valid IMO !
You've made his point simply because cue-mix, total mix, etc ARE DEDICATED DSP ENVIROMENTS that are needed for direct monitoring due to the fundamental flaws of computers. It's only Native when the processing is occuring Natively on the Computer when it occurs on dsp's a dedicated dsp enviroment is being used and it's used cause it's needed. Whether that enviroment is called RME, Metric Halo, MOTU, SCOPE,Pro tools, etc it's all dedicated dsp that's needed and the exact same concept with most. The name of them and the year they came along to help computers achieve direct monitoring is irrelevent only thing that matters in this context is dedicated dsp is used and needed for ultra-low-latency due to computers inabilities.

Quote:
Originally Posted by TAFKAT View Post
BTW: I don't need you to school me in what direct monitoring environments are, there wasn't anything in what I posted where I was confused !
Not trying to be offensive but your comments seem to suggest you don't consider totalmix, cue mix, etc as dsp processing when in reality they are the definition of dedicated dsp enviroments wich are needed for direct monitoring as Jason West stated. The definition of dedicated dsp is fairly simple and easy to understand.

Quote:
Originally Posted by TAFKAT View Post
What DSP system are you referring to ?

For Native DAW's - DSP Hybrid solutions using the UAD2 for example are in fact detrimental to Low Latency Performance where the added arbitration when using higher loads of the DSP cards themselves impose substantial overheads on the potential native overhead available. The whole premise of being able to use the DSP to alleviate the loads off the CPU's to reserve them for native processing doesn't hold true when running at low to moderate latency
I'm refering to REALtime dsp Platforms (SCOPE, Metric Halo, Kyma, Soundscape, etc) not NON-realtime such as Uad, and Liquidmix. SCOPE and Metric Halo have around 1.5ms roundtrip latency at 44khz (less latency at higher samplerates) and only strenghthen your computer's system not tax the Asio resources like uad simply cause their plugins don't have to be ran in the Native enviroment as VST but are ran in the dedicated dsp enviroment like MOTU's Totalmix fx.
Old 19th September 2011
  #83
Lives for gear
 

Quote:
Originally Posted by TAFKAT View Post
Can't really comment about the unit, but you are drifting off topic as the unit is a dedicated unit for one thing and one thing only , not really that relevant to the current discussion IMO.
The context was the advantage of a dedicated hardware unit of some type. With my particular definition , I include the onboard DSP chips of Metric Halo, RME.

Quote:
Why would it ?

Thunderbolt is simply external PCIe ,
Because technology evolves. It won't be Thunderbolt v1 but maybe Thunderbolt v3 or v5 or whatever ("a future Thunderbolt") along with whatever advances happen on the cpu & memory & io bus to transport data back and forth. I won't be so bold to say that it's impossible for Intel/Windows/Mac architecture to match Metric Halo & RME within 10 years but I certainly wouldn't be surprised if it doesn't happen. Low latency appears to be very difficult obstacle to overcome. Part of it is also caused by the inherent audio stack architecture of MS Windows and Mac OS X operating systems that don't have ultra-tight realtime requirements. DSP software code is not hindered with this software legacy.

Quote:
What is the actual RTL of the Fractal and I'll be able to answer that very easily for you.
They don't publish that spec so I don't have any concrete quantification of it. All I know is that it certainly feels better.

Quote:
Also what do you believe is the highest latency that is acceptable for real time playing ?
Not sure. Probably 1ms or less for a real cabinet speaker pushing air. With headphones clamped on the ears, probably 0.5ms or less.

Quote:
His first statement and the main point was that the advantage of dedicated hardware ( as opposed to computers ) is the low latency monitoring.
Sorry. To be clear, I meant "dedicated hardware" to include the DSP chips with "ASIO Direct Monitoring" along with the standalone tabletop recording studios. I didn't recognize the phrase "ASIO Direct Monitoring" at first so I thought it was shorthand for the signal going round trip through the actual host cpu. When I say "computers" I meant computers without augmentation of any DSP -- in other words "computer" as a synonym for the digital audio to go through the Intel cpu x86 instruction set with host DAW plugin processing.


Quote:
My point was that Low Latency Monitoring has been available on computer based DAW's for well over a decade , whether that is achieved via a dedicated DSP on the interface is irreverent , the point stands that the supposed advantage of the dedicated units over computers is not valid IMO !
The dedicated hardware units can include fx with less latency than computers (software input monitoring). The "vanity fx" as you call it is absolutely part of this equation to consider parity of host cpu vs hardware dsp.
Old 19th September 2011
  #84
Gear Maniac
 

The question "do we really need computers" remains a great question!

You can use a Metric Halo ULN8 as your preamp and converter, for monitoring your take, for mixing your song with processing, as your studio monitor controller, and for basic mastering. It is a seamless thought thru closed digital hardware system, and mobile. You use a standalone hard drive.

In this setup what is "the computer" for?

The personal computer is being used for drawing the waveform, editing and riding faders. Yet the tools (mouse/keyboard) are wrong for this job.

What I want is a 20"W x 10"H super-sensitive low-friction touch screen loaded with a DAW that lays flat on my mix desk, with lines out to my MH and HD. Touch edit/ride everything...no screen in my line of sight, no email, no net, no GS, no mouse, no keyboard, no "computer".
Old 19th September 2011
  #85
Deleted User
Guest
Radar with outboard.

You guys have no imagination ; P
Old 19th September 2011
  #86
Lives for gear
 
woodhenge's Avatar
 

Quote:
Originally Posted by Jason West View Post
Not sure. Probably 1ms or less for a real cabinet speaker pushing air. With headphones clamped on the ears, probably 0.5ms or less.
1ms? Seriously? Do you often play on stage with your ear 1 foot away from your guitar cab? I'd hope not...

Realistically speaking, you're probably used to dealing with acoustic latencies much higher than this and just don't realize it. Obviously, when using a guitar sim, you want to keep it as low as possible, but when you think about it in comparison with using a live amp on stage you're actually probably more used to 10-15ms of natural latency anyway.
Old 19th September 2011
  #87
Gear Maniac
 

Radar is cool but that's still using a clunky fixed physical interface, think Rader w/ iPad interface...something intuitive that can be played like a musical instrument.
Old 19th September 2011
  #88
Gear Guru
 
UnderTow's Avatar
Quote:
Originally Posted by Jason West View Post
I won't be so bold to say that it's impossible for Intel/Windows/Mac architecture to match Metric Halo & RME within 10 years but I certainly wouldn't be surprised if it doesn't happen. Low latency appears to be very difficult obstacle to overcome. Part of it is also caused by the inherent audio stack architecture of MS Windows and Mac OS X operating systems that don't have ultra-tight realtime requirements. DSP software code is not hindered with this software legacy.
See my comment about HD¦Native and the new Avid converters. We are already there today! Unless people are going to claim that PT HD TDM has too much latency...

Alistair
Old 19th September 2011
  #89
Lives for gear
 
TAFKAT's Avatar
 

Quote:
Originally Posted by Jason West View Post
Sorry. To be clear, I meant "dedicated hardware" to include the DSP chips with "ASIO Direct Monitoring" along with the standalone tabletop recording studios. ....
Hey Jason,

Actually I should be the one apologising as I misinterpreted what you meant by dedicated hardware as meaning the standalone multi track recorders that were being debated prior to your post. After rereading your post I realised that is not what you meant at all.

Fault is entirely mine for skimming . To be honest the debate leading up was giving me a headache so I skipped past a large chunk of it and then wrongly assumed that was what you were referring to.. :-(

Apology goes out to Sound Designer as well for the above misunderstanding..

I will respond directly to some of the other points tho.

Quote:
When I say "computers" I meant computers without augmentation of any DSP -- in other words "computer" as a synonym for the digital audio to go through the Intel cpu x86 instruction set with host DAW plugin processing.
Crystal clear now.. :-)

Quote:
Because technology evolves. It won't be Thunderbolt v1 but maybe Thunderbolt v3 or v5 or whatever ("a future Thunderbolt") along with whatever advances happen on the cpu & memory & io bus to transport data back and forth. I won't be so bold to say that it's impossible for Intel/Windows/Mac architecture to match Metric Halo & RME within 10 years but I certainly wouldn't be surprised if it doesn't happen. Low latency appears to be very difficult obstacle to overcome. Part of it is also caused by the inherent audio stack architecture of MS Windows and Mac OS X operating systems that don't have ultra-tight realtime requirements. DSP software code is not hindered with this software legacy.
No matter what future protocol , be it Thunderbolt/Lightpeak or any other interconnect , my point still stands that the current technologies are nowhere near being exhausted so the faster interconnects will not give us lower latency. As you noted the underlying variables are the Operating systems and the respective associated driver protocols . However we really aren't that far off now, my only hope is that the future O.S's don't hinder the performance we are achieving now.

Quote:
They don't publish that spec so I don't have any concrete quantification of it. All I know is that it certainly feels better.
Not sure how you are utilising the Fractal , but if you are using its USB interface capability that should be pretty easy to work out the I/O and RTL latencies depending on what DAW host you use. The Host should report the values. If you are interconnecting via its Digital Out via your main interface, you will still be dealing with at least DA latency of the audio interface as well as the additional latency introduced depending on your physical monitoring being headphone/speakers and your position to them.

Which leads me to this

Quote:
Not sure. Probably 1ms or less for a real cabinet speaker pushing air. With headphones clamped on the ears, probably 0.5ms or less.
Really ?

The finite speed of sound in air even at arm length will be about 3ms , so unless you have your ear right on the speaker cone you won't have anything close to 1 ms. Also remember when dealing with Monitoring within the DAW, latency attributed to the finite speed sound/distance is added to the latency from the respective DAW monitoring, whether it be hardware/software.

Quote:
Originally Posted by Sounddesigner
Not trying to be offensive but your comments seem to suggest you don't consider totalmix, cue mix, etc as dsp processing when in reality they are the definition of dedicated dsp enviroments wich are needed for direct monitoring as Jason West stated.
I wasn't dismissing the DSP aspect at all, my point was skewed because I misinterpreted what Jason has been referring to as dedicated hardware.

Quote:
I'm refering to REALtime dsp Platforms (SCOPE, Metric Halo, Kyma, Soundscape, etc) not NON-realtime such as Uad, and Liquidmix. SCOPE and Metric Halo have around 1.5ms roundtrip latency at 44khz (less latency at higher samplerates) and only strenghthen your computer's system not tax the Asio resources like uad simply cause their plugins don't have to be ran in the Native enviroment as VST but are ran in the dedicated dsp enviroment like MOTU's Totalmix fx.
Gotcha... :-)

The difference with those environments ( * not sure about Kyma , don't know enough to form an opinion *) except for SSL MX4/Soundscape , is that the dedicated DSP is not directly integrated into the Host DAW which many if not most end users have a preferance for. Thats not to say that they are not useful or beneficial , only that having to deal with 2 totally separate processing and mixing environments from experience is less desirable. You can also add the SSL MX4 to the later when not used in the Soundscape Host environment.

Anyhooo,

What was the subject again, oh yeh, do we need computers .. ?

Well my opinion maybe taken as biased being I am a professional DAW builder, so I'll reserve comment... :-)

Old 19th September 2011
  #90
Deleted User
Guest
Quote:
Originally Posted by HiFi Yeah View Post
Radar is cool but that's still using a clunky fixed physical interface, think Rader w/ iPad interface...something intuitive that can be played like a musical instrument.
Clunky?

What. That you you use your ear and not your eyes?

I'm not trying to be a dick, but there isn't anything clunky about radar other than the "flash" of wAveforms with earlier versions.

With a radarII system And an Eventide And lexicon you should be able to shame these "computer only" zombies fairly easily.

Even a tascam 8 track with Eventide, lexicon and neve/API would challenge a lipstick and hairspray DAW. No?
Post Reply

Welcome to the Gearslutz Pro Audio Community!

Registration benefits include:
  • The ability to reply to and create new discussions
  • Access to members-only giveaways & competitions
  • Interact with VIP industry experts in our guest Q&As
  • Access to members-only sub forum discussions
  • Access to members-only Chat Room
  • Get INSTANT ACCESS to the world's best private pro audio Classifieds for only USD $20/year
  • Promote your eBay auctions and Reverb.com listings for free
  • Remove this message!
You need an account to post a reply. Create a username and password below and an account will be created and your post entered.


 
 
Slide to join now Processing…
Thread Tools
Search this Thread
Search this Thread:

Advanced Search
Forum Jump
Forum Jump