The No.1 Website for Pro Audio
Don't Call it a Shootout - Audio Interfaces Compared
Old 29th September 2013
  #31
I'm assembling SoundCloud upload for another set, since these files are bigger WAV files. This time, source is WAV file playback from Reaper, using several DAC that includes Realtek onboard line out, and a mid- to high-end 2ch DAC. A caveat is, I have to capture the line outs from those DACs using another ADC to create the WAV files. That DAC-ADC looping could cancel the quality difference between DACs. But at least, listeners monitoring chain is kept constant, for a given listener, and the evaluation is blinded. The capturing ADC is Sound Device USBpre2, which is considered as 'pretty good' by many people, so there shouldn't be a too much criticism about this method.
Old 29th September 2013
  #32
Lives for gear
 

Quote:
Originally Posted by Masaaki View Post
I'm assembling SoundCloud upload for another set, since these files are bigger WAV files. This time, source is WAV file playback from Reaper, using several DAC that includes Realtek onboard line out, and a mid- to high-end 2ch DAC. A caveat is, I have to capture the line outs from those DACs using another ADC to create the WAV files. That DAC-ADC looping could cancel the quality difference between DACs. But at least, listeners monitoring chain is kept constant, for a given listener, and the evaluation is blinded. The capturing ADC is Sound Device USBpre2, which is considered as 'pretty good' by many people, so there shouldn't be a too much criticism about this method.
how much do you think is the influence of different clocks in converters ?
Old 29th September 2013
  #33
Quote:
Originally Posted by lllubi View Post
how much do you think is the influence of different clocks in converters ?
umm... I'm not a believer of impact of better clock on sound quality....except for my 2ch DAC which has 'Superclock' button (which upsamples to 192kHz with less jitter clocking). When I press this switch, the sound appears to widen (Did I say 'appeared'?). I wish I can make that button on/off, and capture the difference with USBpre2, and that's something I want to do with the next test.
Old 30th September 2013
  #34
Good stuff, Masaaki. I had intended to get some more recordings done this weekend, but my day job work crept into my weekend. Still plan to review your existing uploads with good 'phones too.

Anyone else with some gear, some interest, and some spare time (now that's the hard part, isn't it?) is also welcome to jump in with recordings of their own. Please just share what hardware, settings, source material, and software you're using where applicable.
Old 30th September 2013
  #35
Lives for gear
 

Quote:
Originally Posted by Masaaki View Post
umm... I'm not a believer of impact of better clock on sound quality....except for my 2ch DAC which has 'Superclock' button (which upsamples to 192kHz with less jitter clocking). When I press this switch, the sound appears to widen (Did I say 'appeared'?). I wish I can make that button on/off, and capture the difference with USBpre2, and that's something I want to do with the next test.

would like to know if and how far your ADC´s drift away from your DAC with the Superclock
would be nice to see if at 48khz and 120 BPM the difference from one sixteenth to the next would be indeed 6000 samples on the ADCs
Old 2nd October 2013
  #36
Masaaki, I listened to your uploads. Several times.

All three of the C recordings seem less bright than the A and B- the "symphonic c" sounded almost muddy.

I couldn't decide which I liked better between A and B. They each felt a little different, it seems? But I'm not astute enough to be able to characterize it. I would take either chain over what was going on in C though.
Old 2nd October 2013
  #37
Quote:
Originally Posted by UltimateOutsider View Post
Masaaki, I listened to your uploads. Several times.

All three of the C recordings seem less bright than the A and B- the "symphonic c" sounded almost muddy.

I couldn't decide which I liked better between A and B. They each felt a little different, it seems? But I'm not astute enough to be able to characterize it. I would take either chain over what was going on in C though.
Thanks for checking.
PM Sent.
Old 2nd October 2013
  #38
These links below are new ones.

This time, it's a comparison of DACs. Some of you may not agree with the method, but it could help your buying decision.

Signal chain was:
(Computer A, playback WAV file 24bit/44.1kHz Reaper) -> Various DACs ->Line out>Balanced (or unbalanced) cables>Line IN>USBpre2 > Computer B, record as 24bit/44.1kHz WAV file, Reaper)

Once again, I won't name what kind of DACs were used, but you could guess what was included from my previous posts....PM will be sent once you post your listening impression here, to make things a little more interesting, but also keep the blind nature of the test.

Since the recorded WAV files are about 100MB each, I enabled download on SoundCloud. I included as various types of music clips as possible, male/female vocal, clean/OD guitars, acoustic guitar, etc.

https://soundcloud.com/masaaki-slc/da-converter-a
https://soundcloud.com/masaaki-slc/da-converter-b
https://soundcloud.com/masaaki-slc/da-converter-c
https://soundcloud.com/masaaki-slc/da-converter-d

EDIT: In the signal chain, I used SD USBpre2's line input, which actually doesn't bypass preamp stage. I took advantage of it, so that I can adjust gain from different sources. The level matching was done purely on the USBpre2's gain knob, somewhere -3dB at the loudest part (around the end of the file, solo bass clip).

EDIT2: Some may wonder why names of converters are also blinded. Main reason is to remove psychological impact on the hearing. When we know 'one of them is Realtek' for example, we look for one that particularly sounds bad. At that point, our brain is already 'tuned' to find out a bad sounding one. Many shootouts are done in a way like "One is Apogee XX and another is Behringer ZZ, but which one is which is blinded". Such tests are better than non-blinded tests, but still have psychological impact on the listening by 'informed judgment' like "Apogee should sound better than Behringer, night and day" that comes either from spec sheet, previous experience, or hearsay.
Old 5th October 2013
  #39
Quote:
Originally Posted by GearSeekerSander View Post
I just listened to the files for a few minutes zapping thru them. Tbh the files are very distorted and have bad stereo image / too much noise so its hard to say something. My feelings tend to B since i could stand hearing it longer than the other files. All seem to be not sabre like to me... or at least not ES9018. A+D seem to be the worst if there is a possibility to seperate worse from worse.
Thanks for listening. PM sent.
Old 7th October 2013
  #40
Quote:
Originally Posted by Masaaki View Post
Once again, I won't name what kind of DACs were used, but you could guess what was included from my previous posts....PM will be sent once you post your listening impression here, to make things a little more interesting, but also keep the blind nature of the test.
Hmm, I think I liked A best overall, followed by B. C and D didn't seem to be as crisp as the other two, but I only had this impression during the passages that were just acoustic guitar. They seemed more equally matched during the parts with a complete arrangement.
Old 7th October 2013
  #41
Thanks for listening. PM sent.

These are all short passage from CD audio, so they are all somewhat compressed and mastered for loudness, except maybe the classical one. So, elements to differentiate the quality of converters may be lost in either 1) source itself, 2) DAC stage, 3) ADC stage, or 4) Listener's monitoring stage before reaching to the listener's ears. If we call this 'a test with limitation', sure it is, but this is probably what we can do just by providing WAV files. More than this should be done straight from DAC and high end monitoring with proper ABX test gear, and double blind testers and listeners. Playback files should also be something that has full frequency range without post-FX processing.
Old 7th October 2013
  #42
More ADC comparisons

Here's probably my last set of ADC comparisons using multitimbral synths as a source. I think I finally tackled the recording levels problem. There were two cases (one line level entry, one preamp entry) where limitations of the interface resulted in a non-matched level. The rest are consistent. This round uses a Yamaha SY85 playing a track with a fair amount of high-frequency content.

The interfaces under test: Access Virus TI, Focusrite Scarlett 2i2, MOTU PCIe-424+24I/O, RME Fireface UFX, Roland Duo-Capture EX, Steinberg UR22, Zoom R8, and my DAW PC’s built-in audio.

(The bolded interfaces do not have real preamps, so I only did line level tests with them.)

If you are into blind tests, you can go right here and see if you can identify which recordings are from which interface.

Line level: Index of /interfaces/SY85LineRaw
Preamp: Index of /interfaces/SY85PreRaw

Or, if you simply want to know which interface is which, check out the soundcloud playlists I made (the WAVs are also downloadable from there). Each uploaded track lists the name of the interface used to record it. Unlike my D-110 and Access sets, these SoundCloud playlists use the original, non-normalized recordings, although you can still find normalized versions by browsing around in here.

Line level: https://soundcloud.com/ultimateoutsi...y85-line-level
Preamp: https://soundcloud.com/ultimateoutsi...amp-comparison

All recordings are 44.1kHz, 24-bit, at 256 samples (or the closest buffer size, in the case of the Roland and the Focusrite, with their goofy control panels).

I'm considering re-doing my D-110 and Virus recordings, as I have a better way to get consistent recording volumes now.
Old 7th October 2013
  #43
Low Latency Performance Comparison (and other metrics)

Since I've had personal experience with USB interfaces that don't perform as well on USB 3 ports as they do on USB 2, I started a worksheet to track any differences in performance I find. I'm also tracking what the lowest selectable sample buffer sizes are, as well as the lowest USABLE sample buffer at 44.1kHz. Oh, and I also recorded a really quiet test tone to determine the full dynamic range each interface offers (from the gain knobs all the way down to all the way up).

A "usable" buffer size in my usage is one where you can make a 2-channel audio recording through the interface with zero unexpected pops/dropouts.

The worksheet is here, and a detailed description of the columns is here.

Here's a snapshot of the current results:


The RME, Steinberg, and the Zoom could basically take anything I could throw at them, on both USB 2 and 3. The Roland was completely unusable at its control panel's two lowest settings (so bad that some ASIO apps couldn't even communicate with it), even though four of the lowest settings all use a buffer size of 96. (The control panel is pretty weird.)

The Focusrite's lowest setting worked fine on USB 2, but not at all on USB 3- although the next buffer size up worked fine (albeit with the worst latency of the devices used). Another thing about the Focusrite- it ended up having a pretty wide gain range overall, but once the knobs are about 90% of the way turned, there's a big jump in volume, and it becomes very difficult to fine tune levels at that point. This happens on both channels, so I don't know if it's by design or a side-effect of their potentiometer components or whatnot. It's just not a linear volume ramp like with the others.
Old 14th October 2013
  #44
When doing some new interface test recordings early last week I discovered that the passive DI I'd been using for my "preamp" recordings is SUPER susceptible to EM interference. This pretty much invalidated any of the preamp recordings I'd made up to that point, because the DI had introduced some amount of hum that wasn't part of the source signal.

So I picked up a couple of Behringer DI100 active DIs (don't laugh- they're *ahem* strongly inspired by the heavily shielded BSS AR-133 DI and these things are built like tanks... they even look like little tanks), and they eliminated the hum.

Since I was going to redo all the preamp recordings anyway, I decided to put some other recent discoveries to use and redo all the Virus and D-110 line-level recordings as well. With only a couple exceptions which were due to individual interface hardware or config UI limitations, I was able to level-match all the line and preamp recordings at the source. Since the original recordings are more balanced now, I didn't see any point in having separate normalized versions of them, so I've deleted all those.

If you want to "blind listen" and see if you can tell the difference between an RME UFX and a computer's onboard internal audio, you can just browse the WAV folders directly (all files have randomly generated names):
To find out which interface is which, you can check out the SoundCloud playlists (each recording lists the interface used, and the recording names match the filenames in the WAV folders). The recordings are all download-enabled (44.1kHz/24-bit WAV):
The D-110 instrument is so quiet that I had the preamp engaged at least 33% on every interface anyway, so I didn't see much point in doing separate "preamp" recordings of it. In the actual preamp recordings of the other instruments, the signal passes from the synth through the active DIs, attenuated by 40db, and then into the target interface, where the preamp is engaged to make up the difference.

So the complete interface list now is:
  • Access Virus TI interface mode (only used in SY85 line test)
  • Avid Fast Track Duo (in all tests except SY85 line level- oops)
  • Focusrite Scarlett 2i2
  • MOTU 24I/O (no preamp recordings)
  • My DAW PC's Internal Audio (no preamp recordings)
  • RME Fireface UFX
  • Roland Duo-Capture EX
  • Steinberg UR22
  • Zoom R8
Old 28th October 2014
  #45
rjx
Lives for gear
 
rjx's Avatar
Great thread. Thanks! I'm surprised there wasn't more replies.
Topic:
Post Reply

Welcome to the Gearslutz Pro Audio Community!

Registration benefits include:
  • The ability to reply to and create new discussions
  • Access to members-only giveaways & competitions
  • Interact with VIP industry experts in our guest Q&As
  • Access to members-only sub forum discussions
  • Access to members-only Chat Room
  • Get INSTANT ACCESS to the world's best private pro audio Classifieds for only USD $20/year
  • Promote your eBay auctions and Reverb.com listings for free
  • Remove this message!
You need an account to post a reply. Create a username and password below and an account will be created and your post entered.


 
 
Slide to join now Processing…
Thread Tools
Search this Thread
Search this Thread:

Advanced Search
Forum Jump
Forum Jump