• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are daily reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

I have a decent interface and want to measure some stuff to post here. What software do I need, and which tests should I run so things are comparable?

mike7877

Addicted to Fun and Learning
Joined
Aug 5, 2021
Messages
561
Likes
110
Like the title says! ^^

I have some things I'd like to measure and post here - one of them is the second version of the Denon x3700H! :D (the one with the alternate DAC!)

I'd like, when I'm done, for my results to be directly comparable to Amir's. I'd like to know which tests to run, and which software to use. And maybe some pointers like what to do to eliminate error, or if you know common mistakes people make when they're just starting measuring stuff (grounding tips maybe? I don't know, I have no experience lol)
(I do with electronics, just not measuring audio equipment)

The IMD test, for example - if I can't use the same software, I'd like to know things like which tones are used for the IMD test, 60Hz and 7kHz with 4:1 ratio?

I'm going to be using my RME Babyface Pro! You can see how it performs by looking at the review here of the RME Babyface Pro FS, and imagining it's a little worse with jitter, because it is, and that everything else is the same, because it is.

OK it's not exactly the same. FS version shaves 7 or 9 samples off latency (oooooooo lol), jitter's a bit better (well, a lot better, but it was good to begin with, FS is like Perfect re: jitter)

RME's released specifications of the RME Babyface Pro FS match Amir's measurements, and RME's Babyface Pro (non-FS) released specifications, for the most part, match the FS version. I made a picture with some notes comparing the two. Tell me what you think of what I posted, how useful my interface will be for measuring things with. For example, with 30dB gain, THD+n can be -100dB. Does this mean that if I turn up the gain to +30dB, that, effectively, I can measure down to -115dB THD+n on signals up to -30dBFS? If so, that'd be cool!

1698189816048.png



A long time ago I did some measurements, and its ADC section does seem to be good down to -108dB THD+n, maybe even better (sans gain, I never tried that). It was under ideal conditions (I believe I was using a shielded USB cable, and 9 or 12VDC power supply I made using an LM7809 or 7812, some nice, low ESR capacitors + some ceramics (on both sides lmao), run from 4x two-cell lead-acid batteries (of the gel variety) in series.

Anyway, if anyone has any thoughts, if there's a tutorial on some site you could link me to, some helpful pointers for me. Regarding electronics, I own a couple scopes, bench power supplies. I have spare parts to fix stuff - I've built some circuits, designed some simple stuff like power supplies a few years ago. I've fixed stuff over the years.
(also broke some stuff, but we all have lmao)

Once I get this working I'm going to measure everything!
 

staticV3

Master Contributor
Joined
Aug 29, 2019
Messages
6,999
Likes
11,088
Good software for audio measurements includes Room EQ Wizard, Virtins Multi Instrument, and Multitone Analyzer.

Which one(s) you choose doesn't really matter. I personally like REW the most.

Common mistakes are:
-common mode noise due to bad grounding (ground loops)
-measuring class d amps without appropriate switching noise filter (AUX-0025)
-leaving the ADC at default 16bit
-having your measurements be dominated by your signal generator or capture device, rather than the DUT
-using the Wasapi Shared or DirectSound API
-graphs with poorly chosen upper/lower limits on either axis
-unhelpful choice of y-axis unit (e.g. FFT graph with Vrms or dBFS)
-unlabled graphs with arbitrary output-, loading-, and capture configuration
 
OP
M

mike7877

Addicted to Fun and Learning
Joined
Aug 5, 2021
Messages
561
Likes
110
Good software for audio measurements includes Room EQ Wizard, Virtins Multi Instrument, and Multitone Analyzer.

Which one(s) you choose doesn't really matter. I personally like REW the most.

Common mistakes are:
-common mode noise due to bad grounding (ground loops)
-measuring class d amps without appropriate switching noise filter (AUX-0025)
-leaving the ADC at default 16bit
-having your measurements be dominated by your signal generator or capture device, rather than the DUT
-using the Wasapi Shared or DirectSound API
-graphs with poorly chosen upper/lower limits on either axis
-unhelpful choice of y-axis unit (e.g. FFT graph with Vrms or dBFS)
-unlabled graphs with arbitrary output-, loading-, and capture configuration

I'm wondering... You can see here

index.php

that the Topping DX3 Pro (sans plus, I think, which wasn't officially reviewed here on ASR (as far as I can tell)) was measured to be just a bit better than the Babyface Pro FS in IMD

I have the DX1 which is one of the things I'd like to test. I haven't come across measurements anywhere which aren't Topping's own. The measurements they put out on their website do tend to be within 2dB or so of any capable 3rd party measurer (from what I've seen, which is some but not a lot..), so I'm leaning towards trusting their 0.0003% and 0.0002% ratings (~-110dB and ~-114dB ) for THD+n.

They didn't release an IMD chart for the DX1 though, and after taking mine apart the other day and seeing
1698197768014.png

the op-amps chosen (in such tiny 1.7mm x 1.5mm x 0.3mm packages!) were some unknown Chinese manufactuer, I want to trust, but I can't trust what I can't see, and the fact that they release IMD for their other stuff, makes me want to verify...

Usually I'd guess the IMD is 7-10dB worse than THD+n (from what I see other DACs do), which if it's -100dB that's still pretty good, but if it's -90, I don't know if I wanna be keeping it when I could get the DX3 Pro + (which is -110) for a bit more, ya know?

In the DX1 there are two op-amps for the headphone jack out, and one for the RCA jacks. It's my speculation that they chose to use one side of each op-amp for the high current application (up to 100mA per channel), Topping's power ratings imply they limit to ~80mA, but I think if you tried to drive two 80mA loads with a decently high supply voltage too (+5.5, -5.5 I believe), the tiny 0.765mm3 volume op-amp (with no heatsink and not BGA for better heat dissipation but pins on either side) would start to cook itself real fast!

There is a 5v voltage booster+ filter on the USB power in, which drives four (or five, I'm not sure if there's a fifth...) supplies for the internal components. One for the DAC, three identical ones for each of the op-amps, and one for maybe the XMOS? I don't know. Anyway, there's all that stuff inside, and these really tiny obscure op-amps, so I want to verify... 'cause like I said, the DX3 Pro+, for like $200 instead of 100 for DX1, could be up to 20dB better in the IMD category if the DX1 isn't strong there.

So it looks like the Babyface Pro will be between -97 and -100dB for the range between -15dB and 0dB.
What would be the best reliable measurement it could do?

Maybe a hypothetical example is better, actually, to be more exact. Say there's an ADC with IMD of -100 at -10dB. Would its accuracy be enough to pick up -99 at -10dB? If not, what does the difference generally have to be for a measurement to be verifiable/accurate/repeatable?

Maybe I can only measure some aspects of the DX1. I'm really hoping that none of the tests I do on it are able to be resolved by my interface, but you never know!
Like I said in the OP, I'm also doing the DAC in a receiver (Denon x3700H) and it will definitely be good enough to do that one - probably 20dB room to spare!

I'd really like to dig down into how its amps work, and try driving its power amps through the pre-outs on the back. Not all of them work that way (the fronts don't - I think they're buffered with op-amps, same with the Zone2/rear heights) but most of them do. Subjectively it sounds better this way to me, and once upon a time I knew which IC was used in the receivers, but not anymore! I do vaguely remember reading its datasheet, and it's not exactly something I'd choose (or call transparent...).

Lots of work ahead hopefully! I'm going to install some software and play around with it a bit. Starting with REW!
 
OP
M

mike7877

Addicted to Fun and Learning
Joined
Aug 5, 2021
Messages
561
Likes
110
1698202288125.png



This is my loopback sound card measurement!
What's going on with it?
XLR cable from XLR out to XLR in! It's a decent cable...

I expected more consistent results, not this steadily rising crap! I don't see that on things here!

edit: I guess it's all done at one frequency, right? At 1kHz?
In that case, I have -108dB!
4dB better than spec, and just as good as the FS!
Just like I suspected!
Does this mean the ADC is -111 and the DAC is -111? Or how does it work?
 
Last edited:
OP
M

mike7877

Addicted to Fun and Learning
Joined
Aug 5, 2021
Messages
561
Likes
110
Get a Cosmos E1DA ADC from our own @IVX with all the extra added bits. Get a good 120dB SINAD or better DAC. REW is your best choice as it it gets better with every release.

I have a G5, but after using it 24/7 for 4 months I started worrying about its battery. Now I only use it for music - driving my amplifier and headphones, and when I'm out because stupid new phones like S22 have no jack. Next time I'm not getting a Samsung because they're anti-consumer and their OLED screens pulse at 240Hz for no reason, even when they're not dimmed, which will eventually give everyone eye strain and damage and I think they do it on purpose because it doesn't save power or improve image quality or NOTHIN at high brightness all the way down to like 20 nits.

The DX1 is the cheaper one that can stay on all the time because it's cheaper to replace if it breaks from being on all the time, and it won't need surgery every 6 months for a new battery (I can't just leave it expired and then on because then I can't take it out with me. Plus the batteries bulge and might expand and break the board inside)

edit: oh my god it's only like $250!
edit2: they should make one that goes in the other direction! How is an ADC so cheap that good? Was a new chip made? Did they stop hiding all the advancements they've been making over the last 30 years for some reason?
 
Last edited:

Blumlein 88

Grand Contributor
Forum Donor
Joined
Feb 23, 2016
Messages
20,225
Likes
36,399
I would just mimic what staticV3 said for software. I most often use Multitone (written by a member Pkane) or REW (the writer is also a member here). REW has versions for Mac, Windows and Linux. Multitone is Windows (can be made to work on Linux).


As for your measurments they will of course be limited by your interface vs what Amir does or someone with a Cosmos. So you can measure whether something is as good as your interface or worse. Some gear might be better and would be limited by your interface. Your interface is pretty good (I have the BabyFace Pro FS), but you cannot get results equal to Amir's or using the Cosmos. You'll find you can do nearly all the measurements that Amir does with this software. IMD can be done anyway you wish including how Amir does it. You might take a couple devices and work thru duplicating each measurement Amir does and you'll soon enough see how it all works.
 
OP
M

mike7877

Addicted to Fun and Learning
Joined
Aug 5, 2021
Messages
561
Likes
110
If you are interested in more accurate distortion measurements, you should use the Stepped Sine distortion measurement function of REW. It takes quite a bit longer, but for measurements of electronics, it is far more accurate.
View attachment 321037


Isn't what I posted a stepped sine? Or what did I do wrong? The one that shows my card's third harmonic is way louder than anything else... No?

edit: found it... it's running. Definitely not configured right yet, but the first run will be done in 4m54s
 
Last edited:
OP
M

mike7877

Addicted to Fun and Learning
Joined
Aug 5, 2021
Messages
561
Likes
110
This looks messed up compared to yours... lol
1698215667037.png
 
OP
M

mike7877

Addicted to Fun and Learning
Joined
Aug 5, 2021
Messages
561
Likes
110
1698215870784.png


I have these anomalous looking things... I don't know why the amplitude's all messed up at the top, or why the 60Hz 7kHz thing I ran and monitored turned out so very, very terribly...
 

Sokel

Master Contributor
Joined
Sep 8, 2021
Messages
5,533
Likes
5,450
I have these anomalous looking things... I don't know why the amplitude's all messed up at the top, or why the 60Hz 7kHz thing I ran and monitored turned out so very, very terribly...
It's a SPL scale,change it.
 
OP
M

mike7877

Addicted to Fun and Learning
Joined
Aug 5, 2021
Messages
561
Likes
110
It's a SPL scale,change it.

1698218804560.png


Still looks terrible lol. The squiggly line at the top only happened once, but the multitone looks like hell!
Something's definitely wrong there
 
OP
M

mike7877

Addicted to Fun and Learning
Joined
Aug 5, 2021
Messages
561
Likes
110
1698219158496.png



Ah, the multi-step. It was done too quietly, for one...

I think I'm starting to get the hang of how this works
 

Sokel

Master Contributor
Joined
Sep 8, 2021
Messages
5,533
Likes
5,450
View attachment 321059

Still looks terrible lol. The squiggly line at the top only happened once, but the multitone looks like hell!
Something's definitely wrong there
For this kind of measurements use RTA tab.
Use Blackman-Harris 7 window,spectrum mode,8 averages,128k FFT length,no smoothing and 87.5% maximum overlap.
 
OP
M

mike7877

Addicted to Fun and Learning
Joined
Aug 5, 2021
Messages
561
Likes
110
Here are third-party measurements of the DX1: https://www.l7audiolab.com/f/topping-dx1/

Very nice. I just got REW mostly figured out. Well, I'm getting charts for everything like I would expect to see, except IMD should've ended at -2dBFS, but stopped at -12. It was supposed to start at -80, but in that chart it started at -90, and I'm assuming the TD+n is -10 as well. Small problem in the scheme of things - I'll figure it out tomorrow.

Thanks for the measurements - I see my DAC/ADC interface wouldn't be good enough to measure it now lol.
The x3700H shall be interesting tomorrow. I had v1 and v2 of it, and v1 sounded brighter but also more fatiguing and, IMO, a little less clear with dialogue. So I settled on v2 before it was known the DAC replacement was supposed to be notably inferior. It'll be good to see what it is actually capable of.
 
Top Bottom