Topping's new audio interface "E2X2" quick preview.

[Cuckoo Studio]

This is not a formal review, just because I recently completed its measurements and saw that many people in the forum were curious. Since I have it, it's not a bad idea to share it. I, like everyone else, am looking forward to Amir's official review of it!

Its input and output parameters are not as 'impressive' considering there are already many products in this field that have reached their limits on the ASR forum. However, as a sound card designed for functional creativity, and with an estimated price of around $130, it is quite good.

LINE OUT

LINE IN

300Ω load with GAIN active:
106mW （THD+N＜1%）
38mW（best）

Headphone preamp output can down to 1Ω

32Ω load with GAIN active:
615mW （THD+N＜1%）
350mW（best）

EIN：-130dBu A(150Ω)
It could be better if my multimeter were more accurate and if I could control the temperature properly

 

[Music1969]

A little better than my Focusrite Gen3 - nothing significant

What did you use to measure @Cuckoo Studio ?

 

[Cuckoo Studio]

A little better than my Focusrite Gen3 - nothing significant

What did you use to measure @Cuckoo Studio ?

cosmos E1DA+cosmos APU and a selfmade 150Ω XLR resistor

 

[Sokel]

That's less than average as it seems,decent thought.
I hope the sacrificed performance is in favor of stability, and rock solid operation.

@Cuckoo Studio,how easy is to press the 48V button by accident?Does it work instantly or takes some seconds to activate?

 

[Cuckoo Studio]

That's less than average as it seems,decent thought.
I hope the sacrificed performance is in favor of stability, and rock solid operation.

@Cuckoo Studio,how easy is to press the 48V button by accident?Does it work instantly or takes some seconds to activate?

48V is quick to start without any delay. There were some minor stability issues with the initial V1.00 driver, but I took the time to provide feedback to them, and they resolved all the issues I raised within 2 weeks. They are receptive and prompt in making improvements, which is a very positive sign. different people may have different use cases. If you encounter new issues that I haven't experienced before, perhaps try communicating with them. The problem should be resolved.

 

[JSmith]

cosmos E1DA

May I ask, which one?

128-129db(A) grade-A, 127+/-1db(A) grade-B, <126db(A) grade-C

JSmith

 

[Cuckoo Studio]

May I ask, which one?



JSmith

I use grade-A

 

[Sokel]

May I ask, which one?



JSmith

I would argue too about the E1DA against E1DAiso as it first seemed that noise is the problem but in line out measurement there's distortion too.
(funny thing is that it has identical noise profile and distortion as my ancient E-MU when measuring it in loopback)

 

[half_dog]

Firstly, thanks @Cuckoo Studio to take your time to do those measurements and to publish here. I'm of them who was curious about it and waiting a unit to arrive . As you said, it does not impress by its numbers but I think it is okay for its functionality and the price offered.

Sorry if I'm asking too much but could you measure its mic input frequency response at low gain and high gain? Some "entry level" interfaces have different curves in those conditions (maybe to reduce noise), I think it could have this behavior.

 

[restorer-john]

I think it looks pretty decent and I like the LED meters. However, it is abundantly clear the Topping guys have never digitally recorded anything live in their life. Or used any recorder, I would say.

Why do I say that? Fixed linear spaced 6dB graduations! Around 0dB, the graduations should be 2dB or even less, and way more segments. Start at -60dB, do 20, 10, 8 ,4 and 2dB jumps.

Look at a DAT recorder for inspiration:

The new Focusrite Scarlett 2i2v4 looks to be a killer product and the Topping would have to be half the price or less IMO.

 

[Sokel]

What's funny and odd is that trying the same REW settings for EMU the former looks much more forgiving than Multitone Analyzer (let's not forget that Multitone is compared straight with an AP by the RME engineer and was found spot-on,multiple times)

Can someone explain this?

 

[AnalogSteph]

@Cuckoo Studio - since the Cosmos ADC can record at up to 384 kHz, looking at out-of-band noise on the output might be interesting. (We all know how nasty the CS4272 DAC can be.)
A zoomed-in frequency response of both input and output (0.2 dB per division max) is another thing I like to see, to check for digital filter periodic ripple and such (Ideally do this at 48 kHz, and if significant ripple found, repeat at 192).
Finally, the frequency response and a THD sweep at maximum input gain would be interesting.

Another (no longer) user of the E2X2 has reported noise briefly appearing in the input at high gains while adjusting input level, can you repeat that?

A little better than my Focusrite Gen3 - nothing significant

The headphone amp is massively better though. The Scarlett gen 3's is muy potato, literally among the worst in class.

However, it is abundantly clear the Topping guys have never digitally recorded anything live in their life. Or used any recorder, I would say.

Why do I say that? Fixed linear spaced 6dB graduations! Around 0dB, the graduations should be 2dB or even less, and way more segments.

It is not uncommon to record with 18 or at least 12 dB worth of headroom these days (at 110+ dB of dynamic range, you can afford it), so things aren't as tight as they used to be, but point taken.

Can someone explain this?

Have you checked whether absolute levels are the same? If REW and Multitone aren't using the same sound API, there may be some funny business going on there. In the olden days it would not be unusual to see shared mode APIs (MME, DirectSound) 3 dB lower than kernel streaming, presumably to account for mixing of multiple sources as well as resampling.

Oh, I think I see what the problem is, or a problem anyway - the THD+N in REW is either A-weighted (without saying) or wrong, as THD+N < THD is a physical impossibility. (@JohnPM ?) Otherwise they are in agreement within 0.3 dB regarding N and 0.8 dB regarding THD, which in itself seems close enough. Adding both manually, I am getting -103.8 dB THD+N for REW and -103.2 dB for Multitone, which ironically matches neither program's displayed THD+N result. There may be some differences in how they differentiate between HD and N, or S and HD+N, perhaps windowing related. (@pkane ?)

BTW - the traditionally used test frequency is 997 Hz instead of 1 kHz straight, presumably because it is not closely related to either 44.1 or 48 kHz. That would spread out the quantization noise.

 

[Sokel]

Have you checked whether absolute levels are the same? If REW and Multitone aren't using the same sound API, there may be some funny business going on there. In the olden days it would not be unusual to see shared mode APIs (MME, DirectSound) 3 dB lower than kernel streaming, presumably to account for mixing of multiple sources as well as resampling.

Oh, I think I see what the problem is, or a problem anyway - the THD+N in REW is either A-weighted (without saying) or wrong, as THD+N < THD is a physical impossibility. (@JohnPM ?) Otherwise they are in agreement within 0.3 dB regarding N and 0.8 dB regarding THD, which in itself seems close enough. Adding both manually, I am getting -103.8 dB THD+N for REW and -103.2 dB for Multitone, which ironically matches neither program's displayed THD+N result. There may be some differences in how they differentiate between HD and N, or S and HD+N, perhaps windowing related. (@pkane ?)

BTW - the traditionally used test frequency is 997 Hz instead of 1 kHz straight, presumably because it is not closely related to either 44.1 or 48 kHz. That would spread out the quantization noise.

None of the above,it was simple actually,@KSTR found it (as I would expect).I just had to disable the "manual fundamental" setting (which I didn't even know it existed )

 

[AnalogSteph]

None of the above,it was simple actually,@KSTR found it (as I would expect).I just had to disable the "manual fundamental" setting (which I didn't even know it existed )

RTA Window

Manual fundamental​

A value for the fundamental level may be entered for use in harmonic distortion calculations when the fundamental is attenuated by a notch filter. The level is in dBFS and should take account of any gain introduced after the notch.

Well, that would screw with the results, of course. The question is, how did it get enabled in the first place?

 

[Sokel]

RTA Window

Well, that would screw with the results, of course. The question is, how did it get enabled in the first place?

Screw is a light word,if you check Multitone thread you'll see (ANALOG) results down to -142db THD+N (!) which I achieved only by lowering the level (as is logical with the option enabled) and about 2-3 people there trying to see what's so wrong!

As of how it got enabled,I have no clue,I never touched anything in that window except for the bandwidth (I never measured this way before with REW,I always use MTA,I only use REW for the traditional room measurements and I thought I try to straight compare my ancient E-MU with the interface here)
If it wasn't for KSTR I would probably still searching what goes wrong.

On the other hand I now know the way people can "play" a measurement with REW and the tell signs (like low level for example as near 0db or couple of db lower results are as it should be,but go lower with the option enabled and you can turn a disaster into a top notch device)

 

[KSTR]

On the other hand I now know the way people can "play" a measurement with REW and the tell signs (like low level for example as near 0db or couple of db lower results are as it should be,but go lower with the option enabled and you can turn a disaster into a top notch device)

You can always fake a measurement, I've even seen photoshopped AP plots in documents from "professional" companies. Sometimes the cheaters make subtle errors which identifies the cheating. Only way to avoid that is being present in person and supervising the whole procedure.

I agree that the manual fundamental option and other stuff hidden in a RTA sub-panel of REW is prone to get overlooked.

 

[Sombreuil]

 

[lewdish]

The mic pre-amp section and build according to Julian Krause is still a little flawed for a high level interface but for entry level its fine, Topping still showing its chops in the HPA section though. I think its a safe entry point if they can get the reliability worked out, unlike hifi gear pro audio stuff gets tossed around a lot, people carry interfaces everywhere and Topping doesnt have a great record when it comes to survivability of their products.

Kinda wished E1DA would build a full interface w/ high performance HPA already~

 

[tiramisu]

This is quite exciting for those who actually use audio interfaces.

Topping entering the market will push out the bottom and force the top to improve or drop their prices.
Topping also puts out new hardware frequently so we can expect updates at a fresh pace rather than waiting 2 or 3 years for minor updates.
More inputs, inserts, and better user experience on volume controls are needed
but this is an excellent first-release that will compete directly with devices twice to four times the price.

Very good news for everyone who uses an audio interface daily.

 

[Zim]

Ooft if this has a measured low RTL then we have a winner here!