Scarlet 6i6 1st gen vs UMC404HD

stefanocps

New member
Hello everybody
I have a scarlet 616 1st gen which has always suited me for what I do, home recording on Cubase, 90% only daw, I rarely record a voice or a guitar. Apart from the mixer and the sw that generally I never really liked, thecard went well until I discovered that due to the old drivers (I suppose) the latency is not very low, I have to have the buffer at about 160 to have a 10-11m latency. I have never noticed for the keyboards it is fine, but I tried to lower it for direct monitoring of fx and you can hear them a bit ..
Now I can sell the scarlet for 120euro and for the same amount take the behringer for example. I don't want to add too much money so I don't think about others
what do you think about it?
 
I don't know anything about Scarlet's but my UMC202HD has served me well for many, many years now. The Midas preamps are good and it even records up to 192kHz (though I don't do that). However you may want to look into different ASIO drivers or something like ASIO4ALL to rule out "old drivers" as your issue before trying to sell/buy new hardware. That being said the Behringer UMCHD line is really good regardless of bang for your buck price. I'll be upgrading to the U-Phoria UMC1820 next time I'm looking to add new gear.
 
I don't know anything about Scarlet's but my UMC202HD has served me well for many, many years now. The Midas preamps are good and it even records up to 192kHz (though I don't do that). However you may want to look into different ASIO drivers or something like ASIO4ALL to rule out "old drivers" as your issue before trying to sell/buy new hardware. That being said the Behringer UMCHD line is really good regardless of bang for your buck price. I'll be upgrading to the U-Phoria UMC1820 next time I'm looking to add new gear.
no asio4all no! with a dedicated soundcard minimum is to have good latency driver!that's th eonly reason i would change the scarlett, though as i said, for main routine work it is ok
 
I don't think that buying another cheap interface will help your latency problem. You need to be looking at a more up-market interface like an RME or possibly MOTU. In my experience RME drivers can work at much lower buffer sizes than other USB drivers on my machine.
 
I don't think direct monitoring and effects go together. Direct monitoring bypasses the trip through the computer. Effects require a trip through the computer (except with certain interfaces that have onboard effects).
 
also i realize that the behringer is too big..so it seems only choice is m4 motu..i don t know about tascam or steinberg 4 I/O interfaces..how do they pervform?
 
All of these interfaces, regardless of price will perform similarly in regards to Latency. The interface is not your problem. It's either your driver, your PC specs or the FX you're using.
 
All of these interfaces, regardless of price will perform similarly in regards to Latency. The interface is not your problem. It's either your driver, your PC specs or the FX you're using.
i don't write the driver anyway!that is focusrite anyway! i have a ryzen 3700x with 32gb ram and gtx 1070...pc specs are fine...
 
Last edited:
What sample rate are you using? One advantage of using 88k or higher is that it cuts latency in half compared to 44k at a given buffer. Your computer should easily be able to handle that.

With that said, any plugins you are using for effects will have an adverse effect on latency.
 
Latency is a very complex subject and can be affected by many factors. I had a UMC204HD but unfortunately did not get around to testing it for latency. The interface is very good value in most respects, pre amps and converter quality is excellent but it is a bit 'odd' operationally (compared to be fair to my M4) It also has a very low line output capability which might cause issue with some setups. I would also expect its latency to be of middling performance.

Two interfaces in the midd price range stand out to me for almost state of art, low latency. The Native Instruments KA6 and the MOTU M2 or M4. The latter perhaps has slightly better pre amps and seems easier to work (I have a Mk1 KA6) The M2/4 has a 'loop thru' function that records ANY sound going through the PC. It can also run 'stand alone' as a pre amp/converter if powered from any USB 2.0 source including a $2 USB wall plug. That means you can just plug it in amywhere, plug in guitar say and rock.

Higher sample rates in theory give lower latency but in practice the extra work demanded of the CPU can sometimes negate that I understand?

Now, I know I say this often but DO ask at www.soundonsound.com forum Those guys are actually using a lot of this kit and know it first hand.

Dave.
 
People get really bent out of shape about latency, some claiming that even 5ms is a problem. My question is "do you stand closer than 5 ft from your amp when you play?" If you calculate the speed of sound, it moves at approximately 1 ft/ms. 10ms is like standing 10 ft from your amp. If you are on a stage and standing 15-20 ft from the drummer does it cause you to get off time? It was not unusual for our band to be spread out by more than 20 ft on a large stage.

Adding in effects in a pass through the DAW can cause delays to go up drastically. Some effects are worse than others, so you can't really make a blanket statement that interface A will or will not solve a problem compared to interface B. Unless the effects are done in the interface via DSP, it may only make a minor change. Decreasing latency by 2-3ms with a faster interface/driver won't help if the effects delay is 40 or 50ms.

BTW, the best case latency can easily be calculated as (buffer size/bit rate). A 256 byte buffer and 44.1K sample rate gives a one way latency of 0.0058 seconds or 5.8ms. It will transfer the sound data 172 times each second. Increasing to 96K changes that to 256/96,000 = 2.7ms latency. Remember, this is only one way, so you really have to double it for round trip. That doesn't include any delay from the DAW or plug-ins.

If your computer can handle it, using a 64bit buffer at 192K would give you a best case 0.33ms of latency one way. It will have to transfer the data 3000 times each second. That's probably overkill by any stretch, but you get the idea.
 
Back
Top