The No.1 Website for Pro Audio
 Search This Thread  Search This Forum  Search Reviews  Search Gear Database  Search Gear for sale  Search Gearslutz Go Advanced
Mastering with a music server? Digital Converters
Old 1 week ago
  #1
Here for the gear
 

Thread Starter
Mastering with a music server?

I am curious, has anyone here experimented with using a music server to send a final mix to your DAC for mastering OTB. In short, they are supposedly cleaner than a computer (electrical noise etc) and less jitter prone than say computer USB protocol. The only disadvantage I can think of (besides convenience) is not being able to clock via BNC from your AD or an external clock (In brief research, I've only seen BNC outs on high end units with clocks built in). Also, if it's a server+DAC unit, no cable (pre analog domain) seems like it would be purer than any other cable protocol senario. I haven't seen a lot of info on the web about this, specifically for mastering. Any thoughts?
Old 1 week ago
  #2
Lives for gear
 
Greg Reierson's Avatar
 

Verified Member
It's a non-issue. There is no jitter and no electrical noise that affects audio on a USB connection unless it's broken.
Old 1 week ago
  #3
Gear Maniac
Quote:
Originally Posted by Greg Reierson View Post
It's a non-issue. There is no jitter and no electrical noise that affects audio on a USB connection unless it's broken.
There are lots of threats in diyaudio that mention usb noise. Galvanic or optical isolation does a lot to be honest :-)

But a music server might be unneccesary surely. The problem lies often with the noise tgat affects the clock.
Old 1 week ago
  #4
Lives for gear
 
Greg Reierson's Avatar
 

Verified Member
Those would be some seriously dodgy systems.
Old 1 week ago
  #5
Here for the gear
 

Thread Starter
Hmm, well, something as simple as an iPurifier 3 has made a noticeable improvement in my setup (I can identify it in a blind A/B test FWIW). There's a lot of info/published tests out there about computer noise/jitter effecting USB 2.0 protocol. Of course, there's a lot of misinformation on the web as well. I research, but come to my own conclusions by ear. Here's an example of some measurements done: The Well-Tempered Computer
Old 6 days ago
  #6
I sometimes wonder how an audiophile manages to install an operating system or doing online banking without a nervous break-down.

Imagine the miles and miles these signals have to travel through, how often do you see a zip break? a character flip? Or a 25GB game installation fail? Maybe these informatics ppl are less stupid than they use to look. Ignore what they do and what their universities teach (e.g. information theory*), just don't complain about giggles in the background.





* Good overview here: http://web.mit.edu/6.933/www/Fall2001/Shannon2.pdf

Last edited by FabienTDR; 6 days ago at 03:55 PM..
Old 6 days ago
  #7
Transfer Types from the USB 2.0 specification:

- Control Transfers: Bursty, non-periodic, host software-initiated request/response communication, typically used for command/status operations.
- Isochronous Transfers: Periodic, continuous communication between host and device, typically used for time-relevant information. This transfer type also preserves the concept of time encapsulated in the data. This does not imply, however, that the delivery needs of such data is always time-critical.
- Interrupt Transfers: Low-frequency, bounded-latency communication.
- Bulk Transfers: Non-periodic, large-packet bursty communication, typically used for data that can use any available bandwidth and can also be delayed until bandwidth is available


Audio is transferred in Isochronous mode (of course).
Streaming generally assumes that the transfer will continue even if there are uncorrectable errors.

You MUST have a memory buffer to stream anything from USB (with audio, it's simply the audio buffer, but USB also has its own buffer below it). It has nothing to do with error correction. USB sends small packets at high speed, with pauses in between, and the DAC wants one sample per sample clock tick.

The technically possible range of error only allows the production of very obvious sudden drop outs. i.e. a complete black-out, not an increase in distortion or noise! A change in color or timbre requires a regular error, literally impossible when operating a digital system within specs.

Of course, a too slow processor or too small buffer will inevitably lead to issues, but clearly an "out of spec" misuse of the technology.

Last edited by FabienTDR; 6 days ago at 06:08 PM..
Old 4 days ago
  #8
Lives for gear
Sorry, I have no time to enter this discussion ro debate but if you simply just experience to listen music through a computer via USB to a DAC. Changing the USB Cable, you will never hear the same thing. You can hear big changements in tonal balance, transient etc....
I used to be a ME and know I'm reselling gear to studio and hifi for love of sound.
I have sold for instance more than 150 Wyred4sound Recovery to make the USB more immune to EMI/RFI and 100% (studio guys and hifi) heard all a significative improvement.
So, it's possible to demonstrate that USB is immune but it is not.
Ihmo, I would absolutely not master with USB digital distribution.
Old 4 days ago
  #9
DAH
Lives for gear
 
DAH's Avatar
Quote:
Originally Posted by priko View Post
Sorry, I have no time to enter this discussion ro debate but if you simply just experience to listen music through a computer via USB to a DAC. Changing the USB Cable, you will never hear the same thing. You can hear big changements in tonal balance, transient etc....
I used to be a ME and know I'm reselling gear to studio and hifi for love of sound.
I have sold for instance more than 150 Wyred4sound Recovery to make the USB more immune to EMI/RFI and 100% (studio guys and hifi) heard all a significative improvement.
So, it's possible to demonstrate that USB is immune but it is not.
Ihmo, I would absolutely not master with USB digital distribution.
enhanced 0s and 1s?
Old 4 days ago
  #10
Lives for gear
 
evosilica's Avatar
 

Quote:
Originally Posted by DAH View Post
enhanced 0s and 1s?
The transitions from 0 to 1 and vice versa become much sharper, therefore better transients.
Old 4 days ago
  #11
Lives for gear
 
Apostolos Siopis's Avatar
 

Verified Member
Quote:
Originally Posted by evosilica View Post
The transitions from 0 to 1 and vice versa become much sharper, therefore better transients.
???
Old 4 days ago
  #12
Gear Head
 
moostapha's Avatar
 

Quote:
Originally Posted by Apostolos Siopis View Post
???
I'm almost certain it was a joke.

Almost.
Old 4 days ago
  #13
DAH
Lives for gear
 
DAH's Avatar
The better cable statement can work if only USB protocol used to transfer audio does some sort of correction like the Redbook error correction system when a CD is played back. However, I highly doubt it.
Old 4 days ago
  #14
Motown legend
 
Bob Olhsson's Avatar
 

Verified Member
The data can be perfect but a slightly different clock speed, which can be caused by all manner of things including RFI, will sound different. People forget that the clock is good ol' analog including its can of worms.
Old 4 days ago
  #15
Quote:
Originally Posted by priko View Post
Sorry, I have no time to enter this discussion ro debate but if you simply just experience to listen music through a computer via USB to a DAC. Changing the USB Cable, you will never hear the same thing. You can hear big changements in tonal balance, transient etc....
If anything, drop-outs appear. You can expect most audio USB devices to be reasonably buffered. Most ME's will likely use a solid computer with a generously buffered, asio compatible audio interface.

Of course it's easy to "hear" changes. I clearly hear music sounding different in dependence of my mood, with exactly the same cables. Eating, drinking, smoking, flirting totally changes the music experience. Listening to the same loop for minutes also impressively shows this effect. Turing lights off does the same.

"tonal balance, transient etc" are either long term or level dependent distortions, the appearance of these is technically excluded by design (welcome to digital)! They would be easy to measure and demonstrate. If audiophiles would only ever find time for that.



What's maybe less well known though is USB's rather limited tolerance for cable length (from wikipedia):

The USB 1.1 standard specifies that a standard cable can have a maximum length of 5 meters (16 ft 5 in) with devices operating at full speed (12 Mbit/s), and a maximum length of 3 meters (9 ft 10 in) with devices operating at low speed (1.5 Mbit/s).

USB 2.0 provides for a maximum cable length of 5 meters (16 ft 5 in) for devices running at high speed (480 Mbit/s).

The USB 3.0 standard does not directly specify a maximum cable length, requiring only that all cables meet an electrical specification: for copper cabling with AWG 26 wires the maximum practical length is 3 meters (9 ft 10 in).

In that sense, using a 6 or 9m cable could indeed produce some trouble. Clicks, dropout or complete silence. But that's clearly "out of specs".
Old 4 days ago
  #16
DAH
Lives for gear
 
DAH's Avatar
Quote:
Originally Posted by Bob Olhsson View Post
The data can be perfect but a slightly different clock speed, which can be caused by all manner of things including RFI, will sound different. People forget that the clock is good ol' analog including its can of worms.
How can a USB cable or powering voltage/current fluctuations affect the clock on the DAC?
Old 4 days ago
  #17
Here for the gear
 

Thread Starter
Should have known this would turn into another joke/theory/hifi debate forum. Haha . Just wondering if anybody has actually tried it. Thanks for all the input though, appreciate it. I sit somewhere between both worlds I suppose. Though, as a ME, I have done extensive tests with a few cables, and there are audible differences when running mixdowns (AB-ing after, and they do not null). I don't care enough to prove it though (to people who are obviously smarter than me in this area), I guess it'll be a never-ending debate. On a similar note, how about Dante, MADI, Ethernet, Thunderbolt, USB 3 for audio? What is your preference, and why? Insert 1's and 0's joke again here haha.
Old 4 days ago
  #18
Motown legend
 
Bob Olhsson's Avatar
 

Verified Member
Quote:
Originally Posted by DAH View Post
How can a USB cable or powering voltage/current fluctuations affect the clock on the DAC?
It can't ONLY if the DAC and power supply are properly designed. We simply can't assume that.
Old 4 days ago
  #19
Gear Maniac
 

Quote:
Originally Posted by Bodhifile View Post
Should have known this would turn into another joke/theory/hifi debate forum. Haha . Just wondering if anybody has actually tried it. Thanks for all the input though, appreciate it. I sit somewhere between both worlds I suppose. Though, as a ME, I have done extensive tests with a few cables, and there are audible differences when running mixdowns (AB-ing after, and they do not null). I don't care enough to prove it though (to people who are obviously smarter than me in this area), I guess it'll be a never-ending debate. On a similar note, how about Dante, MADI, Ethernet, Thunderbolt, USB 3 for audio? What is your preference, and why? Insert 1's and 0's joke again here haha.

Wait wait wait. So you're saying you run the same mix through the same outboard gear and DAC, and you record it, and then do nothing but change the USB cable, record it again with same setup, and it sounds noticeably different and doesn't null (it wouldn't necessarily null any way running the same outboard chain twice, esp. if tube gear is involved)? And you're not willing to prove that?? All you have to do is post a couple 30 second clips of each mix recording, and we can all judge for ourselves...
please?
Old 4 days ago
  #20
Here for the gear
 

Thread Starter
Quote:
Originally Posted by dogoftears View Post
Wait wait wait. So you're saying you run the same mix through the same outboard gear and DAC, and you record it, and then do nothing but change the USB cable, record it again with same setup, and it sounds noticeably different and doesn't null (it wouldn't necessarily null any way running the same outboard chain twice, esp. if tube gear is involved)? And you're not willing to prove that?? All you have to do is post a couple 30 second clips of each mix recording, and we can all judge for ourselves...
please?
Sure, send me a clip. I'll post, yes, I understand that analog gear contributes to not nulling too.
Old 4 days ago
  #21
Simply pass any audio through you DA/AD loop, change USB cables, do it again. Upload the results.
Old 3 days ago
  #22
Lives for gear
 
Greg Reierson's Avatar
 

Verified Member
What? Your text nulled...
Old 3 days ago
  #23
Lives for gear
 
Greg Reierson's Avatar
 

Verified Member
Do it again without changing the USB cable just to make sure your system is stable.
Old 3 days ago
  #24
Lives for gear
 
evosilica's Avatar
 

Quote:
Originally Posted by DAH View Post
How can a USB cable or powering voltage/current fluctuations affect the clock on the DAC?
Gravitational waves can modulate the clock via doppler effect. Sounds very bad.

But seriousy, DACs do not derive their clock signal from some analog signal off the USB lines.
There is a buffer right in front of the dac that spits out the samples with a stable frequency, so it doesn't matter how much jitter there is on USB, as long as the data is not corrupted and/or not too slow. If that happens it'll start to crackle ...
There's no way any kind of slight degradation can happen due to bad cables. It's either a valid transmission or horrible noises.
Old 3 days ago
  #25
Here for the gear
 

Thread Starter
Hmmm, well... Let me start off by saying that I am not interested in being right as much as finding out what's true. Especially on the web, there is SO much conflicting info and opinions, it can be quite a maze navigating it sometimes. I do believe there are strong psychological factors that contribute to the way we perceive music. I do find myself second guessing and wondering how much placebo plays into impressions often. BTW, when I said "people who are smarter than me in this area", I was being honest, not condescending. I am just curious and obsessed about this kind of stuff, and somewhat new to what is considered the "Hifi" world (which I definitely think can get wonky). Just a humble recording/mixing/mastering engineer (around 15 years experience, which is a bit, but nothing compared to some of you guys). Not here to create conflict. Though if I can give you a chuckle, that's cool too.

This may be a little long winded, so bare with me.
The results... they did null! Well... let me explain:

This was the test (Test 1): I took an old mixdown and looped it through (for simplicity) my Prism Lyra 2 AD/DA on it's internal clock with 1M Vovox Sonorus Direct TRS cables. One with a random USB cable I had laying around, one with an AudioQuest Diamond USB cable (which I got pretty cheap on eBay), then one with the Diamond and the iFi iPurifier 3. Then one with a chakra infused incense stick, but my chi must have been off cause it wouldn't stay in the USB hole, oh well*.

So... yes, they did actually null with each other (not with the actual inbox mixdown, I know the AD>cables>DA changes it). Which did surprise me some, because last time I did this test, they did not... after some thought, I came to the conclusion that when I did this specific test before (with this cable), it was a direct inbox mixdown, not a mixdown of a previously bounced stereo mix. There are certain plug-ins, like reverb that have fluctuating algorithms, or something of that sort which contributes to them not nulling from actual realtime plug-in differences.

However, I did hear some very small differences in detail, the imaging seems a little better, and the highs more refined on the Diamond. Also, the iFi creates a wider 45 degree image that is apparent, but it also nulls. Overall, very subtle, yes subtle enough that it could be placebo (some sections I couldn't tell when looped), they do null after all! This definitely had me questioning my ears.

So for Test 2: I recorded the AD/DA loop directly from the DAW mix playing back in real time. These did not null, you can hear reverb discrepancies etc. Though, I think the results of the cables are a little more apparent here. As the source is more direct (another 1's and 0's debate with digital summing), I can hear the differences clearer (especially on a more resolving headphone amp, I used an Antelope Audio Zodiac Gold for this). The random cable seemed a bit dull, lacking some dimension and clarity. All this is very subtle in the grand scheme of things obviously, but mastering is mostly the art of the accumulation of subtitles.

Will a USB cable make better music? No.

I am curious to hear your impressions, here's the link:

Uploadfiles.io - Archive.zip

Do you guys hear no difference at all when toggling through these on your system?

Anyways, I really need to get back to work. I don't really have the time to be doing this at the moment.
Old 3 days ago
  #26
Lives for gear
 
evosilica's Avatar
 

Quote:
Originally Posted by Bodhifile View Post
I do believe there are strong psychological factors that contribute to the way we perceive music.

Quote:
Originally Posted by Bodhifile View Post
However, I did hear some very small differences in detail, the imaging seems a little better, and the highs more refined on the Diamond. Also, the iFi creates a wider 45 degree image that is apparent, but it also nulls. Overall, very subtle, yes subtle enough that it could be placebo (some sections I couldn't tell when looped), they do null after all! This definitely had me questioning my ears.
Your answer is in the first quote. Some people find it hard to believe, some even reject it, but still it's the way it is.

Quote:
Originally Posted by Bodhifile View Post
So for Test 2: I recorded the AD/DA loop directly from the DAW mix playing back in real time. These did not null, you can hear reverb discrepancies etc.
Reverb algorithms (among other effects) have some randomness. LFOs starting at different phase etc. You cannot do valid null tests with them, because they will produce a different output each time. So you have to print them and the use the audio file for further tests.
It has nothing to do with the cable!




Big for doing these tests. Usually it's just claims followed by "I don't have time for this, if you don't hear it, that's your problem".
Old 3 days ago
  #27
Gear Maniac
Quote:
Originally Posted by evosilica View Post
Gravitational waves can modulate the clock via doppler effect. Sounds very bad.

But seriousy, DACs do not derive their clock signal from some analog signal off the USB lines.
There is a buffer right in front of the dac that spits out the samples with a stable frequency, so it doesn't matter how much jitter there is on USB, as long as the data is not corrupted and/or not too slow. If that happens it'll start to crackle ...
There's no way any kind of slight degradation can happen due to bad cables. It's either a valid transmission or horrible noises.
No, indeed not. However, the clock is analog. And it is quite sensitive to fluctuations in the power. So in a good ADC/DAC design, the clock power is isolated form the digital domain and only transfers the actual clock to the DAC/ADC.

then there is the bit-perfect thing. When you have a buffer with a FIFO, the stream of bits just come in, regardless if the square is slightly less square.

So if the buffer is working with the clock and a fifo, there is a theoretical perfect transmission of the bits.

In cheaper dac designs, the clock is often powered by the USB or whatever computer connection. There is much noise on the line and actually does influence the power to the clock. Without a buffer, this would definitely affect the sound quality.

The downpart of a buffer, is that it introduces latency. With mixing/mastering, that slight delay might not be the issue in a reproduction status. But when you record, then you have a much bigger issue. But now the focus on the DAC.

If you have a properly isolated analog part in the DAC from the digital domain, the results can be great. It is however not a cheap solution, and that's why good dacs cost a bit more. It's also very dependant on the voltage regulators. Even better than the voltage regulator is using a battery. Now this is a very puristic solution, but it actually -does- work. And the data is very available.

So, as many DAC builders say; its not just the DAC chip, but it is the implementation what makes the difference.

For more information, you should check out the DIYAudio forums and read up there (with many measurements). Its super interesting.

When I was building my streamer, I had a jitter problem, and I included a buffer, an isolated clock board, and powered the clock separately. When I powered the clock with a voltage regulator (good one), it sounded great (bigger soundstage, more defined instruments), but when I used the battery. It was such a big difference. They are really not subtle.

The better dacs like Cranesong, Lyra, etc are good DACS. They have some way or another solved most of these issues.
Old 2 days ago
  #28
Here for the gear
 

Thread Starter
Quote:
Originally Posted by evosilica View Post
Reverb algorithms (among other effects) have some randomness. LFOs starting at different phase etc. You cannot do valid null tests with them, because they will produce a different output each time. So you have to print them and the use the audio file for further tests.
Yes, I came to those conclusions. Did you download the files and listen for yourself?
Old 2 days ago
  #29
Here for the gear
 

Thread Starter
I still hear differences on resolving systems, and can pick them out blindly, mostly on Test 2. Subtle, yes. Not really on stock sound cards etc though. My hunch is that nulling doesn't tell the whole story, I could be completely wrong though. I remember reading similar debates in the BLA MK III XB vs Antelope Audio 10M clock comparison thread a while ago.

Please, someone else try it out since I took the time and it was requested. Even just to confirm that it's still just placebo.
Old 2 days ago
  #30
The developers behind whatever signal processing gear are likely well educated and experienced. Otherwise, they simply wouldn't have that position. While doubt is a healthy attitude, this sector is particularly well documented (with every step mathematically proven!). Design patterns are very well developed, as is the design of highly critical software system.

It's somewhat arrogant to expect such a developers to oversee such banalities. In this sector, nobody gives you a million budget without solid former references. Even in China, India or wherever.

A developer specialized in audio gear will of course know very well about naive design mistakes, and already have elegant solutions at hand to avoid them altogether, "by design". It's not that difficult anyway, and the solutions are well known (those guys literally wouldn't have passed their exams without understanding these solutions in full depth).

All this is just too easy to test and exclude automatically, even before any line of code has been written.


If you see a problem, for the sake of fairness, document it properly, explain it in all depth. That's interesting!

But it's just too easy to hallucinate. It's too easy to underestimate the depth and mountain of cleverness hidden behind essentially any field from bakery and agriculture to electronics, IT or whatever.
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