Why does cheap integrated sound have better latency than expensive audio interface..?

Configure and optimize you computer for Audio.
RELATED
PRODUCTS

Post

@diggler: it's not the onboard soundcard which is the bottleneck when adding plugins, that would be the CPU!
We are the KVR collective. Resistance is futile. You will be assimilated. Image
My MusicCalc is served over https!!

Post

AnalogGuy1 wrote: Kaine, Thanks for the link; I've not heard of the RTL Utility before, but after seeing how it works I agree that it is exactly what is needed to determine latency. However, my results don't agree with your predictions; I get a sub-10ms latency using sample buffers well over 32 samples using nothing more than my laptop's integrated sound card and Asio4All. Image
You've got a great on board then! I assure you not all chips are created equal. The other side of the equation is how well the driver set can handle the stacking of VSTi's (as Diggler asks about) and for that you would use it in conjunction with the DawBench suite to come up with the results that you see on the chart on the previous page... it's also handy to carry out a transparency test as well but that doesn't factor into that set of results on the chart.

As I said before, not everyone needs more than A4A and the onboard. But I have heard some really horrible on board and not all of them work fine with A4A. If your laptop does what you needs to a level you need then yeah, its all you need.

Post

Kaine wrote: The older legacy devices that they inherited from the pre-takeover company on the other hand could be patchy, but that came down to the controllers in use and what they had to speak to at the other end in the system. They could have fixed it by changing the hardware in use for sure, but then they would have had to support two lots of drivers and in the end it would have been more cost effective to run down the stock they inherited and develop a new range as they did.
Surely cost effective but very unethical (and also lying to customers). I think I still saw Fasttrack Pro selling perhaps 1-1.5 years ago. Windows 7 driver is dated to 05/2009 (Installation package claims that the driver is year 2010, but device manager most likely tells the truth). So in still 2011, perhaps even 2012 they were selling a product that has 1.5 - 2.5 year old drivers that have been confirmed to be very buggy and unstable from multiple sources and claiming in the product description that it is windows 7 compatible. This interface was first in the top-10 sold list for many months in a row so I guess they made a lot of money too. I have tried many different computers (desktop and laptop from 2006, laptop from 2010, desktop and laptop from 2012 and a few others), all with intel chipsets and USB host controllers and everyone has the same problems in win7-64bit.
The NI kit I've used on decent systems are trouble free. I've never had one glich with them and I've spent months on the road using them for demo purposes on a multitude of rigs. I'm not going to say they are flawless because for sure they won't be.
My Traktor Audio 2 is the most stabile interface I have owned, I haven't had even one problem with it. But I do read and hear complaints about komplete6.
Look at Apple. They build everything to spec. They fit a dedicated in spec FW controller and a dedicated USB controller etc... and they use the same one's in each range.
I have understood that Apple uses the USB controller that is integrated to Intel chipset as does most of modern PC laptops too. But probably the FW is better than in majority of PC laptops, or so I have heard.

This means that the firms in question can write drivers to speak to those onboard controllers and it all just works because the is no variation to screw it up.
I would really want to hear references if there is onboard USB controller in modern macs. I got the impression that Apple switched to USB3 because they were forced to (because intel integrated USB3 to its chipsets).
So all the sound card manufactures can do is design hardware that meets offical spec and hope it works. It isn't their fault if the laptop you bought isn't capable or working to an industry standard. That and we have to rely on a third party driver to make the OS work with shoddy controllers and whatever other random junk is in the average cheap pc these days I'm amazed any of them work to be honest.
well I did try many laptops, if you have the same problem in many different configurations then you can point finger to the audio interface vendor.

I think big problem in the audio industry seems to be that everybody seems to assume that only OS X users deserve good drivers. I do understand if you get trouble with exotic configuration like netbooks but if you have most of the same internals as a macbook but get trouble in windows (or even in Mac running bootcamp windows!), then the trouble is clearly in windows drivers.

Exactly, we're talking about a tiny company here. Creative has thousands of staff, where Natives you could probably count on your fingers and toes at a push.
If Native Instruments is considered a tiny company, how small are then companies like Focusrite, RME, Presonus, Motu, etc...? They seem to sell much smaller amounts than NI.

Post

golemus wrote: Surely cost effective but very unethical (and also lying to customers). I think I still saw Fasttrack Pro selling perhaps 1-1.5 years ago. Windows 7 driver is dated to 05/2009 (Installation package claims that the driver is year 2010, but device manager most likely tells the truth). So in still 2011, perhaps even 2012 they were selling a product that has 1.5 - 2.5 year old drivers that have been confirmed to be very buggy and unstable from multiple sources and claiming in the product description that it is windows 7 compatible. This interface was first in the top-10 sold list for many months in a row so I guess they made a lot of money too. I have tried many different computers (desktop and laptop from 2006, laptop from 2010, desktop and laptop from 2012 and a few others), all with intel chipsets and USB host controllers and everyone has the same problems in win7-64bit.
Avid took over 2 or 3 years back and then the new company took over 6 months back. I don't think it's an unethical situation in regards to lying to customers rather they took over and decided to can support for a load of hardware they were never going to build. I don't think people know how long it takes to shift a batch of stock through channel when it comes to audio kit... it doesn't really fly off the shelf!

Just as a relevent example, we got a heads up that the Delta series was due to go end of life about a year ago because none of the core components were available to the market. These units are still being sold due to various firms stocking up on them and the chances are the are still pallets of these things in various warehouses around the world. In fact a load of Delta 1010lt's turned up recently 6 months after we were told no more were available anywhere!

Now in that time frame the company has changed hands and the new company doesn't have access to any of the old engineers. So they can either spend R&D on building new interfaces not based around hardware that is now over a decade old, or they can spend time reverse engineering the drivers so they can update the hardware that is already out there that they haven't made any money on?

Whilst I can understand your frustration at the inevitable answer to that question that's business for you.

My Traktor Audio 2 is the most stabile interface I have owned, I haven't had even one problem with it. But I do read and hear complaints about komplete6.
They are based around the same controller chip! The have been some issues in regards to failure to power the phantom power enough (not overly suprising on a USB powered device) and some of the earlier units have had USB 3 compatabilty hissy fits, but they have updated and refined the units as they've gone along. I'm in no way claiming they have a 100% rate but they do seem to try!
I have understood that Apple uses the USB controller that is integrated to Intel chipset as does most of modern PC laptops too.
Yeah, except it only started going native on this generation of boards in the last 12 months, so anything previous to that USB3 wise is on a none native (NEC/Renesas) controller and that can be where some interfaces tend to trip up. As the are a number of laptops out there now with only one or two USB2 ports and a raft of usb 3's this can be problematic for some users.
But probably the FW is better than in majority of PC laptops, or so I have heard.
They use a Texas Instruments chip on a lot of models and nobody on the PC side does these days, so yeah that is right.
well I did try many laptops, if you have the same problem in many different configurations then you can point finger to the audio interface vendor.
Depends what the kit in question was you were trying. Your average DICE controller interface doesn't work on the vast majority of laptops but it's built to work with the TI controllers which do meet spec standards. They don't work with those cheap controllers I mentioned before that don't meet spec, is that TCKonnekts fault for not testing it with every 2cent controller coming out of Taiwan or the fault of the controller maker for cutting corners?
I think big problem in the audio industry seems to be that everybody seems to assume that only OS X users deserve good drivers. I do understand if you get trouble with exotic configuration like netbooks but if you have most of the same internals as a macbook but get trouble in windows (or even in Mac running bootcamp windows!), then the trouble is clearly in windows drivers.
Or it lays with windows. As get noted here any number of interfaces work fine with mac but fall apart on windows. Mac has a native engine that will handle the audio whilst windows requires the third party ASIO solution make it capable.

I do agree with what your saying but I'm just playing the devils advocate here in an attempt to show why the problems manifest as I do agree the situation is a bit of a joke at times. It always comes back to how predictable the end combination is through with these things.

With Mac's they are built with good quality parts which they can predict with ease what they are going to be and support them accordingly. The is less variation for them to deal with and it's easier to test all combinataions of kit.

PC they don't know what the end combination is going to be from the controller level upwards so to test ever combination is pretty much impossible. If you invest in a good system where you know good quality parts are being used and the right controller chips are in play you tend to do better with stability but you still can't predict everything.

Then when you factor in that most firms buy their controllers off the shelf with drivers written for them that they then tweak you have another layer of abstraction to deal with. As the Engine for the Mac is native you can just plug them in and get on with life, with the PC they have to write the ASIO drivers to make it run, and if they are relying on a third party firm to write and maintain the updates your going to run into problems... expecially if that firm is doing so for more than one company!

The are days I'm amazed anything works!
If Native Instruments is considered a tiny company, how small are then companies like Focusrite, RME, Presonus, Motu, etc...? They seem to sell much smaller amounts than NI.
NI however have a far, far broader product base. Yeah, perhaps they have a few more staff than RME/Presonous/Motu but those firms don't employ coders/artists/software testers etc. Just because a firm has a larger number of staff doesn't imply that any of them know anything about sound interface design and support.

Post

Kaine wrote:
golemus wrote: I have understood that Apple uses the USB controller that is integrated to Intel chipset as does most of modern PC laptops too.
Yeah, except it only started going native on this generation of boards in the last 12 months, so anything previous to that USB3 wise is on a none native (NEC/Renesas) controller and that can be where some interfaces tend to trip up. As the are a number of laptops out there now with only one or two USB2 ports and a raft of usb 3's this can be problematic for some users.
So you are saying that Macbooks 2011 and 2010 do have 3rd party USB host controller that is connected to the physical USB ports and the intel USB-controller (that is integrated to the chipset) is not used at all in these machines (except for perhaps integrated webcam)....?

Post

golemus wrote: So you are saying that Macbooks 2011 and 2010 do have 3rd party USB host controller that is connected to the physical USB ports and the intel USB-controller (that is integrated to the chipset) is not used at all in these machines (except for perhaps integrated webcam)....?
No, because USB 2 is in the Intel chipset spec, so it was... well part of the chipset on the Intel controller, USB 3 only went native on the Z77 chipset which came out last year so that was on an add on chip up until that point. So yeah I get your point about the chipset itself in those regards and yeah it should be standardized by this point to the level where everything just works on it.

When I talk about shoddy third party controller chips I'm talking about the one's handling the stuff that isn't on the main chipset spec. Some chipset specs only support 2 or 4 X native USB ports and some laptops might have 6 so the rest have to come from somewhere... this is why you sometimes see some ports on a laptop working fine and some not work at all with your interface of choice.

The other factors that can still come into play even when running off the native ports. Are they meeting spec in regards to getting enough power? Or is another component being prioritized higher up the chain moving the bottleneck elsewhere in the system?

Ok, another example the system has all sorts of bits in there working together but something must be given priority. This can be done via the bios or the design of the unit. Some systems will always give more realtime access to the gfx subsection or the WiFi controller will seize control of the CPU and not return it in enough time to allow other subsections real time access. Even if it's coming through the intel controller which is supposed to manage allow fair access where required, one poor driver or component elsewhere in the system can take control send it to pot.

Post

[quote="Kaine"]
When I talk about shoddy third party controller chips I'm talking about the one's handling the stuff that isn't on the main chipset spec. Some chipset specs only support 2 or 4 X native USB ports and some laptops might have 6 so the rest have to come from somewhere...

My laptop (lenovo thinkpad EDGE E530) has HM77, 4 USB ports (3x USB3), two integrated USB devices (webcam and fingerprint reader), so in total 6 USB devices and it does not have a 3rd party USB controller. All of them are under intel controller.

I think I saw one ivy bridge laptop with 3rd party USB controller (Renesas), I think it might have been Samsung laptop and I really didn't understand why it had Renesas controller as it had few USB-ports/devices.

Post

My theory is (tell me if I am wrong) that internal soundcard has always very high priority (higher than any other device) so if you set a low buffer it really doesn't matter. Always when soundcard needs something, anything else is interrupted immediately.

If it is like this then the driver quality doesn't even matter (or maybe bad drivers consume more CPU but you still will not get glitches).

But if it is a USB device it will never have the highest priority and there can always be problems. I just don't understand why there doesn't exist some kind of mechanism inside the OS (NOT BIOS) to specify that a certain device has higher priority than anything else in the system. If this was possible we would get rid of all glitch problems permanently as you could always set the highest priority for audio interface.

Maybe this will be possible with USB3 audio interface...?

Post

golemus wrote: I just don't understand why there doesn't exist some kind of mechanism inside the OS (NOT BIOS) to specify that a certain device has higher priority than anything else in the system. If this was possible we would get rid of all glitch problems permanently as you could always set the highest priority for audio interface.

Maybe this will be possible with USB3 audio interface...?
jackd sound server in linux lets the musician set maximum priority
to it, if desired, so no other process can interrupt audio. There are mac
and windows versions. qjackctl is it's gui.

The bus bandwidth access to cpu and ram, whether pci, or usb or firewire,
will determine how much data can pass, then, audio hardware designs, software
drivers, operating systems, and the speed of cpu/ram, determine the systems
potential.

Then, user knowledge of optimizations and workarounds come into play.
(If microsoft and intel never sat down in peace-pipe powow to agree on
usb2 implementation, don't hold your breath for a miraculous V3.
Usb2 was rushed out fast, to satisfy the immense emerging market
of pirated media. :x
Cheers

Post

golemus wrote: My laptop (lenovo thinkpad EDGE E530) has HM77, 4 USB ports (3x USB3), two integrated USB devices (webcam and fingerprint reader), so in total 6 USB devices and it does not have a 3rd party USB controller. All of them are under intel controller.

I think I saw one ivy bridge laptop with 3rd party USB controller (Renesas), I think it might have been Samsung laptop and I really didn't understand why it had Renesas controller as it had few USB-ports/devices.
The HM77/Z77 boards were the first ones to get USB 3 natively and came out around the start of last year, before that it was all third party chips.
golemus wrote:My theory is (tell me if I am wrong) that internal soundcard has always very high priority (higher than any other device) so if you set a low buffer it really doesn't matter. Always when soundcard needs something, anything else is interrupted immediately.

If it is like this then the driver quality doesn't even matter (or maybe bad drivers consume more CPU but you still will not get glitches).

But if it is a USB device it will never have the highest priority and there can always be problems. I just don't understand why there doesn't exist some kind of mechanism inside the OS (NOT BIOS) to specify that a certain device has higher priority than anything else in the system. If this was possible we would get rid of all glitch problems permanently as you could always set the highest priority for audio interface.

Maybe this will be possible with USB3 audio interface...?
Depends on how the bios/os allocates resources. I wouldn't be suprised at all however!

Post Reply

Return to “Computer Setup and System Configuration”