How thoroughly bizarre! (Interface comparison)

[drew_drummer]

So I was troubleshooting some issues with Presonus Studio One tonight, by using a different card - using my Digi 002 with the CoreAudio manager.

Logic reported it as having a 2.9ms roundtrip latency at 64 buffer size.

Compared to my Profire... the Profire gets 6.4ms roundtrip!

How strange!

If the Digi 002 had a Hi-Z input for guitars, I'd switch!

 

[AllanD]

I'm always looking for ways to reduce latency, 2.9 sounds pretty great :O.

 

[HeadCrusher]

If the Digi 002 had a Hi-Z input for guitars, I'd switch!

Aren't inputs 1-4 supposed to be Instrument level as well? At least that's what they are labeled...

 

[drew_drummer]

Hmm. Will double check... I'm pretty sure they're Mic or line-level only.

 

[Line666]

There should be a button next to the line ins for instrument level?

 

[HeadCrusher]

No button but still it clearly says "Inst" which I suppose means "Instrument".

 

[drew_drummer]

Hmm. Seems you're right, according the manual I just connect a guitar to one of those inputs and adjust the gain. I will give it a go.

 

[Line666]

No button but still it clearly says "Inst" which I suppose means "Instrument".

Ah yeah, I was looking at the phantom power button and putting two and two together and making five, my collective bad.

 

[Jind]

It's my understanding that the reported latency that you see within your DAW is what the device driver is reporting it's latency as (generally only it's playback (ASIO output buffer) and not necessarily accurate as a round-trip latency figure as true round-trip latency is the sum of the following items (I got this information from Jim Roseberry over at the Sonar forums):

* ASIO input buffer
* ASIO output buffer
* D/A and A/D converters
* Driver's safety buffer (this is typically hidden - not adjustable)

Note: Especially on older units, it's not uncommon for the driver's hidden safety buffer to literally double the round-trip latency.

I've always used "CEntrance Latency Checker" to check my latency as it does a true round-trip test and reports the results accurately, just make sure you use it correctly because while it will run internally over digital patching will give you a low reading that is not accurate as it does no conversion. The proper way to use it is to patch an analog out to an analog in and test your latency over that connection so it includes everything end-to-end (including D/A A/D). There are details of how to use it correctly in the PDF that comes with it.

Instructions: http://centrance.com/downloads/ltu/measurement_instructions.shtml

Application: http://centrance.com/downloads/ltu/

Hope that will help.

 

[drew_drummer]

Logic explicitly says 'Round-trip latency' when you go to the audio preferences.

The figures I gave 6.4ms and 2.9ms seems to be with all of that taken into account.

I think Logic has some clever way of figuring it out - or it just doubles the one-way latency and shows that maybe. As I said ... Logic. I'm on a Mac. That utility wont work on a Mac as its for PC only.

 

[Jind]

Logic explicitly says 'Round-trip latency' when you go to the audio preferences.

The figures I gave 6.4ms and 2.9ms seems to be with all of that taken into account.

I have no doubt that's what it says, I'm just relating the information that I've found when researching the very same thing reported in Sonar - it also states "round-trip", but as was explained to me - it's still trusting what the driver reports which in most cases is simply the playback results from the ASIO output buffer. These is no way for the results to be accurate without taking into account the analog world without physical patching. While that may be fine if everything you are doing is completely in the box, once you plug-in a guitar or microphone you are introducing the A/D D/A converters which also need to be measured.

Once again, this is how it was explained to me so I now use the Latency Checker to get my readings.

 

[drew_drummer]

Yeah, you're right. I was just baffled at the differences between the two reported figures. Could be due to one device being synchronous and one being asynchronous.

 

[Jind]

Yeah, you're right. I was just baffled at the differences between the two reported figures. Could be due to one device being synchronous and one being asynchronous.

Not sure if this could be it, but in the same thread over at the Sonar forums that I found the information I posted, it was noted that different devices have different "Safety Buffers" built in that can result in some wildly different reported latencies. The person I quoted - Jim Roseberry is one of the most knowledgeable guys I've meet online about hardware and has explained and answered many a question I've wondered about.

 

[drew_drummer]

Yup, I know Jim. Spoken to him lots of times, met him on the Reaper forums back when Reaper was my main host.

 

[MetalJonesy]

drew, one thing to bare in mind with using the 002 with the instrument level inputs,

i struggle to use EMG equipped guitars as they clip the input, and you can't turn the gain down enough on them, which also seems to be the case when using them with a DI box (on the mic input side of things), so when i HAVE to use the 002 to DI, i use an inline 20db attenuator

 

[drew_drummer]

Cheers Jonesy!

I tend to use the 002 for extra inputs via ADAT. Usually that's for drums, but I've never noticed needing a pad for my Bareknuckles - which are of course passives.

 

[aortizjr]

Soundonsound has started to post the figures from CEntrance Latency Checker now on interfaces they review which is awesome. If you want to use VST guitar plugs or Synths, you have to watch out for that. There are a few interfaces out there that are really unusable.