Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[BL]

Considering upgrading the PC running my instance of JRiver.  Looking for opinions on an existential question. I recently saw a video on YouTube (RMAF16: Computer Audio Demystified - https://www.youtube.com/watch?v=JEztALnHhQg) where the presenter compares two systems running JR and it appears most people in the room hear a better sound when JR feeds the DAC from an high-end powerful laptop vs an old low-end weak laptop.  His point is that more powerful systems will drive better sound on the analog side of the hifi setup.  The 90-minute video/presentation goes on explaining the many ways hardware on the digital side impacts the (analog) sound coming out of the DAC.  You don't need to watch it to reply to this post.  The concepts are not new.

I have an IT background and I have a hard time believing more CPU power, more memory (RAM), etc... can impact the output of a DAC. It's not only that audio processing is not very demanding, it's also that getting 1s and 0s transferred between systems without error has been a no brainer in the IT world for a few decades now.  If the 1s and 0s are delivered to the DAC in time, how else can hardware improve the DAC's performance?  Sincerely, this whole presentation on the impact of hardware on sound comes across to me as a bunch of baloney.  In addition, I suspect the JRiver team does not believe that since the JRiver Id is not based on the most powerful NUC available.  If they thought more CPU power made their software sound really good, I suspect they would not hesitate to use a stronger NUC and make it shine.

This being said, I remain open to ideas as to why more computing power can help.  Maybe I have yet to hear the correct explanation.  This is a sincere question.  I don't want to impose my opinion on others.  Thoughts anyone?

[BryanC]

Ignore "results" that question the underlying nature of digital audio and do not include double-blind ABX testing. That's been the policy of Hydrogenaudio for many years, and I think it's a good one.

[landiepete]

At the risk of opening yet another can of worms, the 'it's only 1's and 0's' argument doesn't really hold a lot of water, since there are multiple variables in play, one of the most important ones in experiencing music is timing. When, for instance, listening to a live performance, and in particular larger ensembles playing acoustically (i.e. where the information from all the performers is not captured, fed through third rate electronics and then fed to really bad PA speakers), a lot of the information is decoded by our brains on the basis of timing, in real time (obviously). Timing information is not captured in the digital stream, but rather controlled by the 'clock' (in the dac when it is asychronous, in the source - in this case the pc - when synchronous) of the master. So when the timing is not perfectly accurate, the brain will interpret the information differently.

This is only one example where 1's are not 1's and 0's are not 0's. Well, they are, but they're in the wrong place.

Pete

[ths61]

Use Thesycon's "DPC Latency Checker" to see if your system is handling interrupts without excessive latencies when under normal music playback loads.

https://www.thesycon.de/eng/latency_check.shtml



Does your USB DAC have sufficient galvonic isolation to prevent ground loop hum ?  DACs vary in their ability to handle noise rejection.  Balanced outputs also help.

Also check how asynchronous DACs handle timing.

[ths61]

... I recently saw a video on YouTube (RMAF16: Computer Audio Demystified - https://www.youtube.com/watch?v=JEztALnHhQg) where the presenter compares two systems running JR and it appears most people in the room hear a better sound when JR feeds the DAC from an high-end powerful laptop vs an old low-end weak laptop.  ...

I have an IT background and I have a hard time believing more CPU power, more memory (RAM), etc... ...  Thoughts anyone?

I watched about half of the video in the hopes to learn something new. 

The first thing he did was told the audience what to expect to hear before playing the first sighted A/B listening test.  He basically set the expectation bias (tainting the response) seconds before pressing play.

He did not measure the outputs to make sure they were level matched but told the audience to trust him that they were the same and that said that lower noise floors give the perception of louder volumes.  If he took measurements, he could have demonstrated the relative volume levels and noise floor levels.  I think this was a [un]intentional missed presentation opportunity.

He suggested configuration optimizations for audio that are the same configuration optimizations any good IT person would do to optimize their own hardware to get the best performance for computing.  He did NOT mention tuning off non-essential resource consuming processes/features to make the most of the hardware one has which is another miss in my opinion.

He brought up EYE PATTERNS and jitter and implied it was an issue.  He did NOT ask the question, does it matter to well engineered asynchronous DACs and if so how bad does it have to be to make a difference.  He then brought up reclockers and mentioned the ReGen product as an implied remedy. 

Independent tests demonstrated the ReGen product made no improvements to well engineered asynchronous DACS (even inexpensive DACs costing less than the ReGen by itself).  During the independent testing, measurements showed the ReGen's digital switching wall wart power supply was actually introducing noise into the signal so it was actually doing more harm than good.  After much back and forth between the independent tester and the vendor of the ReGen product, the ReGen vendor finally fessed up to the PS noise and said he would change their digital switching wall wart power supply vendors. 

There was only 1 poorly engineered asynchronous DAC (of the DACs tested) that showed measured improvements with the ReGen reclocker.  It was cheaper to buy a properly engineered asynchronous DAC than the tested DAC plus the ReGen so money ahead to not use either.  All of the other cheaper and more expensive DACs showed no improvements.

He also glossed over the counter question about memory playback.  If USB drives are so error prone, there would much bigger problems with functioning computers.  Entire operation systems are run off of USB thumb drives without a glitch or crash.

He brought up some good points to computer newbies, but unfortunately I did not learn anything new.  There appears to be some room for refinements in his presentation IMO.

[danfl75]

I'm in IT too and have a EE background. I don't think it's about error rates, but have read there's 3 types of jitter (never saw them listed out tho).  RFI/EMI might also be involved.

The why it works is hard to answer IMO, but you should focus on if it makes a difference since you're skeptical. Buy an AQ Jitterbug, iFi iPurifier 3 and/or good usb cable with money back guarantee and play with it. I've had great success with usb chains. Even changing the power cords to these devices or PCs makes a noticeable difference. Ethernet cables too. There are lots of threads on audio forums with lots of opinions and options. If you do try, tell us what happened and good luck!