Julian, thanks for doing this - very useful. I would be interested to see a chart showing the relationship of both the analog gain & digital conversion step. Thanks for making audio info more user friendly for the non-audio engineer crowd.
Hey Julian, good that you added this chart to your reviews. But can you also add the Hi-Z input specifications of the different interfaces you reviewed? (just like how headphone outputs differ in quality in audio interfaces, I wonder if Hi-Z affects recorded DI signal)
Great explanation Julian, thanks! Considering that nowadays it’s more an more podcasters who are looking at i.e. the SM7B and that podcast recording consoles are USB audio interfaces too, it would be really interesting to see how the Zoom Podtrak P4/P8 (70db “gain”), the Tascam Mixcast4 (66,5db “gain”) and the Røde Rodecaster Pro (55db “gain”) stack up against the other interfaces on your killer graph.
I would like to see a EIN vs gain graph. From this you can see when a higher gain will not make EIN substantially lower and the higher gain will just make the dynamic range lower.
Good as always!!! could you analize the Apolo UAD? It would be super nice to ser all elements you normally analize into that interfaces! (solo and twins) Thanks for your awesome vids always!
@@TheSideProject They make killer stuff for a great price! For budget stuff they are unbeatable as proven by the good results of their UMC HD series here.
The Rode AI-1 just recently came out with a firmware update to increase the gain. 1. Is that possible? I had always assumed it was hardware based. Someone did a test on Reddit, and determined the SNR was the same, however, which I assume is good. Secondly, I assume because of that, the AI-1 would sit higher on the list now, yes? (The update came out after this video)
Fell for the marketed "65 dB" of gain with the zen go. It actually works out to 45dB in practice, so it's fortunate that the interface makes it easy to add digital gain in real time or it's be no good for live applications with lower sensitivity mics.
Hi Julian, great work on this :D Since interface manufacturers do not have a universal standard specification for the minimum and absolute gain on their interfaces, do you think it will be a good idea to *cover this aspect in your interface reviews?
When you look at the gain graphs of various audio interfaces you've compiled, it seems like a lower max input level and a higher gain range are better. This is true if you're recording the source directly into your computer using the interface's preamp. However, if you plan to use external gear such as a compressor in between, you should be mindful that having a higher max input level is advantageous. The reason for this is that a max input level minus 4 dBU allows for true headroom in analog signals.
Cheers! But - how come you didn't mention the Dynamic Range details and added it to the calculation? As far as I remember, it says something about the quality of a device, isn't it?
I hate it when i cant compare two of the same specifications, without having a degree, The last few weeks i read up on the subject of audio equipment and i thought iam now capable to compare DACs Interfaces, Preamps and Microfones. Its not very beginner friendly, but i will try my best ^^ One question: How can i find out, without measurements, how much voltages is supplied by audio interfaces on the "XLR & 6.3mm combo port" when phantom power is turned off ? Icant find any information on this, neither for my umc202hd nor a different audio interface. If i had to guess, id said its 5V coming from the USB port that gets converted to ~48V
Nice video! ✨ Thanks a lot. Have you already made a specific video about ad/da convertion? I might have missed it.. If not, i would find that a interesting topic for a future video! :) best regards
It's weird, there's many people on reddit claiming the iD4 MKII and iD14 MK2 have substantially more mic input gain, yet your measurements show that it's actually the Scarlett 2i2 that wins out that comparison (by one 1db, but still).
Hey Julian, could you try and get yourself a unit of the new MOTU M2/M4 which have different internals in them and see if you can spot a difference in performance?
I have a MOTU M4 of the new units being made in China. As far as the interface itself, everything seems to work as intended with the original units, the mic pre-amps can power an SM7b/SM57/RE-20 to clipping levels. As far as the DAC inside the interface I believe it is an upgrade over the original components, as it's now the same chip that is in the Ultralite Mk5. I unfortunately don't have any measurement equipment in order to test these things out, though. Would just be a good video update since they're no longer making the M2/M4 that he reviewed.
Thats helps me a lot. As i bought TLM 102 and scarlett can't handle it i had to found this video to understand why it's so quiet (small waveform) at HALF OF GAIN. Big thanks! Maybe you can recommend me something for my TLM?
I don't get how anyone who is concerned about mic gain would bother to even consider interface "total gain" to be an attribute worth comparing when there are so many fantastic hardware preamps out there.
So the less capable 2V full scale interface will produce louder digital output for the same 1V input than then a 4V full scale device. That's counterintuitive, but very good to keep in mind - louder does not mean better.
I have the MOTU M2 and the MTrack Solo and according to these graphs, the MTrack Solo has a higher maximum gain than the MOTU M2 - but ----- from my own personal tests the MTrack Solo couldn't even really properly drive an SM7B or an XM8500 whereas the MOTU M2 could do it very clean. There's a stepping issue with the Mtrack Solo which makes it jump quite a few decibels towards maximum gain range. But overall I find myself, with my own personal tests, finding the MOTU M2 has more gain range and more capability in driving an SM7B as compared to the MTrack Solo -- but that confuses me because according to the graphs the MTrack Solo has better gain than the ID14 Mark 2, MOTU M2 but imo those interfaces on my experiements far outperformed the MTrack Solo.
I think because of the step in the gain of the M-Track Solo it seems that it has less gain because you often stay below this step. But if you fully max out the gain on the M-Track you will notice that it records a stronger digital signal than the M2. The measurements for my graph are always done at max gain and the M-Track has a lot of gain avaailable in the last few degrees of the gain dial.
How does this affects when recording guitars with hot pickups? When I record guitar with active pickups I always get about -1, -2 or close to 0 dbFS which leaves a very little headroom and when adding more layers of plugins it starts clipping. to get a lower (like -6dbFS) input should I get an interface with low "gain" or high "gain" ? in my case lower "gain" is better, am I correct?
I susspect you are missing something, because using an SM 57 to record pan flute & quena, I only need 1/4 gain to maax out with a Presonus 24C, despite your claim thai it is a poor interface. The 24C gain is all pos when using XLR input but with line in, the gain is control drops the whole range to provide some attenuation at the expense of reducing max amplification. This fuction is not obvious in the specs. Anyway, I just want to add that the 0 VU guide holds pretty well, especially for programs with good dynamic expression where 0VU mean will naturally run very close to 0 dBfs peaks. It's only not the case when content has had the dynamic expression compressed out off it. That of cause may be to please studio execs that show up fot 2 mins or to fight wind & traffic noise on the go, but it suck for listening to music in any meaningful way.
There's another way to explain "Full Level Scale", etc. There is a maximum voltage which is a physical limitation and set's the end of one's dynamic range. So in the speaker, there is a voltage which would extend the magnet to its full projection. Any voltage greater than that and the speaker might break. So it's not like someone picks 1 volt or 4 volts. The electrical engineer defines it through the physical characteristics of their circuit. Put another way, let's say we measured gain from 0; that is our reference system is based 0 volts. So no voltage is 0 and any higher voltage is greater, we can give them any numbers. 0,1,2,3, to as much voltage as our source gives out. Let's say we have a dynamic range of 0 to 454 volts. How can we know our output source can handle it? We can't? We MUST expect that no system can take unlimited voltage but must WORK BACKWARDS from its highest voltage. "Full Level Scale" is another word for the maximum voltage the signal is referenced to (again, can't reference to 0 because that is really 0 to infinity, in a sense). Because we reference to the MAXIMUM level all numbers below it are NEGATIVE. When you move your signal from one system to another the electronics match the maximum value of the source to its maximum value, up or down. If you didn't do that, one could damage the other. When we're measuring sound we use positive dbs because a sound can go physically higher. Like "Big Bang" high ;) In short, in electronics we work backwards from our voltages. In describing physical phenomenon, we can set our measurements and let it run. So the larger our negative db the more dynamic range we have between our strongest signal and weakest. But as you've pointed out in other videos, a large range in itself does not equal better quality. I see dynamic range misused in photography ALL THE TIME. People don't get, measuring dynamic range basked on a gray card only tells you the dynamic range of grays, NOT colors, and it doesn't take into account noise. So much fantasy.
I'm glad I went in with the UMC 404HD 4 years ago. I wasn't sure what I was getting and only bought because it was a great deal for its price, but now I realize that it really was the best option for me at that moment. Nice graphs btw
Hi Julian Do you have a measured calculation for a Roland Rubix 22. I bought an audient 1d4 mk11 and the gain is about half of my previous rubix 22. the seller says there is nothing wrong with it? this was my calculation - audient - Maximum input level 12DB - Gain range 58 DB = - 48 DBu rubix - Maximum input level 2 DB - Gain range 48DB = - 46 DBu by the the audient should be louder not quieter?
I have both of these and I was using m-audio one first but I hear better when singing with the Scarlett 2i2 …… and is it me or on the Red Scarlett there are some laser line and dots that (I’m still playing with these) but if we get the perfect adjustment we get a awesome result, but I’m still trying to figure about do I use the second xlr input when the mic is on the first and turn at max level and the big silver monitoring where it needs to be set and last the volume button, does it affect the level of your recording or it’s only for headphones ???? And the 48v for phantom power red button do I use it even if the mic don’t need it lol sorry but it’s been a year now I’m learning all by myself but need some answer by a pro and you seem to be this guy lol ✌️ and you videos are really helpful to understand this confusing world of audio interfaces hehe
Like many things the answer is... it depends. An interface with a low max input is actually a help (the first half of Julian's gain equation here) in getting more signal into the device and getting you closer to 0dbfs. While that might be a good thing if you have low gain mics and low level sources it can be a problem if you have high gain mics and/or loud sources because you may experience clipping in the analog stages of the preamp before getting even near digital clipping (Odbfs). Manufactures have to make a decision of the balance between max input, amount of preamp gain, and cost when designing a device. If the mic input has more overall gain (Julian's combined figure) that would benefit someone recording a podcast with an SM7b (the actual gain of the micpreamp matters as well of course) but that same micpreamp if it has a low max input might easily overload on someone recording say a drum kit with a large diaphragm condenser mic or sometimes really loud sources even with dynamic mics. What is "best" comes down to cost, what exactly you record, and what mics you use to record with. Knowing what you are doing like matching high output mics with low level sources, low output mics with loud sources and knowing when to engage a pad (on a mic, micpreamp, or adding an in line one) will help you to get more into the sweet spot and get more out of any interface.
Oh, definitely, particularly if you are known for yelling (or singing very loudly) into your LDC microphone. Some of them have a sensitivity as high as -33 dBV/Pa without featuring a pad switch. If now the interface cannot take more than say -4 dBu (Behringer UMC202, 204, 404), that's the equivalent of 120.8 dB SPL if you do the math. Opera singers at 6" may peak at up to 125 dB SPL, and you may be singing closer than 6". Some LDCs can handle more than 140 dB SPL, most will make it to 126 dB (at 1% THD) at the very least. Now thankfully most interfaces will accept around 0 dBu at least - in fact values of +9 to +10 dBu are quite common these days. Maximum input level tends to attract more interest from guitar and bass players - humbucker pickups in particular are known for delivering particularly high output, although even those tend to peak at no more than 1 V peak... still enough to just about clip a Behringer UMC202 instrument input (-3 dBu max = 775 mV peak). The 1st gen Focusrite Scarletts were also known to be somewhat problematic in this regard, later ones can take higher levels and are generally considered fixed.
@@PileOfEmptyTapes thanks for explain, I am a electric guitar player and I am very curious how to record the guitar with the best quality(full, juicy, powerful ^_^). So that’s to say , higher max input level can record the sound better, but less “gain”. The lower max input level can get more “gain” but it would be clipping easily. Right?
I just want to add this here, because I think it's important context that is missing from the video. This type of "gain" is usually present in budget interfaces and is not always desirable. If you look at the chart, the interfaces with the least "gain" are among the most expensive. That's because this isn't really "gain" - it's actually "headroom". The more expensive audio interfaces have a ton of headroom - that M-Audio interface can only handle a 2v signal whereas the Focusrite interface could handle a signal that is twice as loud at 4v. In other words, you can send a LOUDER signal to the Focusrite. The thing about using audio interfaces with expensive microphones is that the onboard preamps on even the best interface can't compete with a real outboard mic preamp. If you're going to spend $400 on a microphone, you're wasting your money if you're just going to run it through the onboard preamp of an audio interface. BUT if you have a good interface you can turn the gain all the way down and have enough headroom to run a really loud signal from a good quality mic preamp. All of this is really expensive, which is why so many people would be better off buying an SM57 instead of the SM7B.
I learned something new from what you wrote. Everyone is assuming that every audio interface has the best preamps when that would make the whole interface a lot more expensive.
Hi Julian! Sorry if this question is off topic from the video but thought this was the best way to contact you. I have 28 ohms headphones and I recently got a great deal on an audient evo 4. After watching your videos you mentioned there could be an issue with low impedance headphones with this interface. Would you be able to explain what issues it causes when listening to something like game audio or music?
Hallo Julian, weißt Du zufälligerweise, welches Audio Interface sich für weibliche Stimmen eignet? Ich hatte vor 8 Jahren, als ich mein Mikro (sE Electronics USB2200a, USB/XLR-Hybrid) kaufte, da schon Problem, eines zu finden, das zu meiner Stimme passt. Da mein altes Mikro nicht auf Windows 10 per USB laufen wird, muss ich es leider per XLR anschließen, aber ich bin mir unsicher, welches Interface in meinem Fall am besten wäre. Schöne Grüße!
Die meisten Audio-Interfaces haben keinen hörbaren Einfluss auf den Klang. Sie verstärken das Mikrofonsignal und digitalisieren es. Den größten Einfluss auf den Klang haben die eigene Stimme (Tagesform, Stimmung, Trainingszustand, ...), das verwendete MIkrofon und die Raumakustik. Die analoge und digitale Technik sind heutzutage durchentwickelt. Klangliche Quantensprünge nicht mehr zu erwarten. Wenn Du mit deinem Mikro happy bist, sollte jedes Interface zwischen 100 und 200 Euro Dir helfen können.
From my experience with a Scarlett 2i2, UMC404HD, Motu M2, Audient iD14 and Apogee Duet, the best has been the Apogee, followed by the iD14 and surprisingly then the Behringer. With Motu M2, I could never get a clean signal with my guitar with SD Blackouts and EMGs 85/81 at max volume, even setting the level at minimum in the interface (it has't a pad button), I didn't like how it sounded, somekind of distorted and clipping with strong strums. 3 months later the unit died. Maybe I just got a bad unit. The Apogee Duet and the iD14 handle very well also the UMC404HD (without pad engaged) but this last adds a little coloration.
Do you have a website where you these type of graphs for preamps and microphones? If not, you can be the RTINGS of pro audio. They do the same thing as you but with TVs and sound bars and headphones. They take manufacturer numbers and show you the real numbers so you can see if there was a misrepresentation. And people can subscribe to you for a fee. Most of your clients will likely be people shopping for equipment and will only purchase one month, but pro musicians and enthusiasts will purchase the less expensive yearly subscription. You're welcome... My fee is 10%. I'm kidding of course
Hey, -3 dB is half the power and -6 dB is half the voltage. For a mic preamp and ADC we are talking about voltages and that's why the -6 dB statement is correct here. Cheers!
@@Thrashman138 I guess I understand why this guy is saying that this is wrong, because I was a bit confused when I first saw the video. As said in the video "dbfs" is a relative scale, which means that it doesn't mean any real value unless there is a reference point. So instead of comparing the signal levels in "dbfs", comparing the signal levels in "dbu" would make more sense. Let's say there is "Audio Interface A" with max input level as +4 dbu and "Audio Interface B" with max input level as +24dbu. If we record a +4dbu signal with both interfaces, "Audio Interface A" would show that the level is 0dbfs, and "Audio Interface B" would show the level is -20dbfs. Which means that "Audio Interface B" is capable of recording higher signal levels, which is generally preferred, since it offers more headroom. These days modern DAWs opperate with 32bit float point or 64bit, so increasing or decreasing gain with already recorded signals with out adding any artifacts is very easily achieved and the two signals mentioned above (0dbfs & -20dbfs) can be easily matched by just raising or lowering the signal gain in the DAW. I think this guy is saying that the video is wrong, because the video is comparing the gain range of audio interface's which really means almost nothing in real life. Almost every audio interfaces are capable of increasing the mic signal to it's maximum input level and what people should be seeing is not the gain or gain range but the actual maximum input level of an audio interface. Also, the video seems like it's saying that motu amplifies the signal better, but actually they are almost the same. - The max input level of focusrite is +9dbu and the signal is recorded at -11dbfs. This means that the recorded signal is -2dbu. - The max input level of motu is +1.5dbu and the signal is recorded at -3dbfs. This means that the recorded signal is -1.5dbu. the difference is +0.5 dbu which is negligible in real life. So, if it was up to me, I would pick the focusrite over motu since they both have almost identical amplifying capability, but focusrite being able to record higher signal levels.
Great info! One other thing I'd love to see you do would be a minimum "gain" comparison, too. For example, my Focusrite Scarlett 2Pre clips on its input with active guitar pickups and palm mutes even when the gain is set to minimum at "Instrument" level forcing me to use a DI box. Setting it to "Line" level or whatever they call it makes the signal sound strange. My Apogee Duet 2, however, takes the signal totally fine without the need of a DI box. It would be useful information for guitarists who may not have a DI box handy, I think.
I have something like this situation. I have a first generation Scarlett 2i2 (yes, it's alive!) and when I record a Squier electric guitar directly (then I use Amplitube), it sometimes overloads, no matter the gain level is all the way down. I ended up moving the pickups further away from the strings. I don't have a DI...
@@PabloUliver-Bass-Guitar beware the fact that changing your pickup height will change your sound! There are inexpensive DI that work great like the behringer ones. It would be better to use one of them and keep the pickup in it's proper position.
@@ferdinandobarisciano6934 Thanks a lot for the information! I'll have that in mind and as soon as I can I'll buy a DI. But, for now I have to do it my way. Thanks again!
So how does signal to noise ratio factor into this? My first thought watching this was that the usability of a higher input gain would have to be dependent on the noise floor of the system as a whole too, but how would you factor that into the gain calculations shown here?
Hey Julian, What an amazing video. Well explained. I see that I have chosen the right interface m-audio air 192/4 and other interfaces that I own are top too I see. Ok even though it not only has to with gain range only. But before I had an expensive Focusrite, it was not bad but not as good as I wanted so I sent it back and purchased the m-audio. Much better than the twice as expensive Focusrite. Danke sehr für deine Video.
Hey Julian, I'd like to hear your thoughts on a recent discovery I made thanks to your video. I just compared the outputs from my Rode AI-1 using patch 1.3.0 (latest standalone firmware) and patch 1.3.1 (updated with Rode Central*) and found that at 100% gain the gain range increased from 45 dB -> 60 dB. This seems to align with the updated website copy that claims as much. My understanding and methodology is very basic so I'm not sure what exactly changed, but I am seeing ~15 dB louder digital levels now. Hope this gives you a lead. Results before and after firmware update: (1 kHz test tone from a speaker, 3" from a Behringer XM8500, recorded into Audacity, measured with Youlean) 100% -36.6 LUFS -> -21.3 LUFS (+15.3 LU) 50% -56.7 LUFS -> -48.2 LUFS (+8.5 LU) 0% -80.9 LUFS -> -80.7 LUFS (within margin of error) Gain range: 44.3 LU -> 59.4 LU *For anybody having issues installing the update, Rode Central seems to close if there are no active sound devices in Windows 10. I had to plug my headphones into the motherboard instead of the interface for it to not exit prematurely during the firmware update. Otherwise it would drop out during the update, indicate a yellow signal level, and require power to be disconnected before it can be detected again.
I am curious… From my results, I get the impression, that unlike the signal, the EIN of the preamp is always relative to the full scale, no matter what the max Voltage may be. At the same time tho, the signal, being a certain voltage is relative to the maximum voltage. Is that accurate? And If so, would an EIN of -129 dBu and a max input level of 10dBu have the same noise in the DAW, compared to what I’d get with -128dBu EIN and 9dBu max input level?
Why can't the specs for every digital audio interface list the reference level? [Reference level: 0dBFS = +24dBu] How freakin' hard is it to put that on the box? Come on, why doesn't everybody do this?
To be fair, most of them actually do, especially on the output side or for fixed-level line inputs. It just isn't all that easy when reference level isn't fixed. For example, CS4272-based interfaces generally have analog gain control on the output, and on some of them reference level can effectively be set higher than maximum undistorted output level if cranked all the way up (e.g. Tascam US-2x2HR). In these situations one would wish for both to be specified. On the input side I think I've seen just about any conceivable combination of maximum input level, input sensitivity, and either gain range or min/max gain - any two of them will tell you what you need to know, just some math may be needed. This is a bit of an annoying situation for things like my online DIY microphone level calibration calculator, where for the best user experience I may be forced to provide input fields for all kinds of data, showing/hiding them as needed. The fact that some interfaces have the luxury of a gain display while in others with analog gain dials you are best off estimating your position from maximum is in no way helping.
I have a problem which is slightly off-topic but I'm desperate. I'm running a PC as a streamer (HP Deskmate, if it matters) which is outputting USB Audio2 over a USB cable to an Arcam irDAC, which at the moment is only functioning as a digital bridge to coaxial digital, instead of as a DAC. (This problem happened when its analog output stage was being used, also, so that's not it.) The coaxial digital output from the Arcam is connected to the coaxial digital input on a pair of Edifier s2000mk-iii powered speakers. The problem is this: Somewhere in the audio reproduction chain, the gain is being turned up way too high -- so much so that it introduces significant amounts of clipping -- and the signal is then being normalized back to the various slider-volume settings in the chain, so that turning down the volume only results in quieter distortion. I've tried to be scientific about where this is happening, and it's difficult to control for all the variables but I **think** it's somewhere in the initial conversion to USB Audio2. Can anyone offer any suggestions? There doesn't seem to be a control panel for that: I can access the Arcam's settings on my PC, and I can access the generic volume adjustment options for sound on the PC, but the necessary adjustment seems to be "in between" those two arenas of possible settings-changes. Any thoughts or reactions would be greatly appreciated.
7:52 I would really like a touch more information on the sensitive Mics - I occasionaly see comments about mics that need a bit more power. That might be useful for a noob like me.
