DAC Comparisons using AI


After a couple of years trying different DACs in my system, I ended up with the Aries Cerat Helene (R2R) and the SMc Audio DAC-2 (delta-sigma) in my main system.  

I have been considering other options, and decided to use AI to help me imagine the possibilities.  I have found it actually works pretty well if you are able to specifically address what you are looking for. Anyone else here believe you are getting helpful answers by using AI when considering a purchase?

I have been asking specific questions like:

  • What sonic difference would be achieved by upgrading the Aries Cerat Helene to the Kassandra Reference II?
  • Might someone who enjoys the sound of the Aries Cerat Helene find the EMM Labs DV2i to sound fatiguing?
  • Compare the sonic signatures between the Aries Cerat Helene, MSB Technology Premier, and Totaldac D-1 Triunity.

I have not yet encountered answers I would consider total BS, and using AI has sort of bridged the gap between different industry reviews, like when you finish reading a review and wish, if only the reviewer had compared X to Y.

 

mitch2

I use AI daily. At work and at home. M365 Copilot, github copilot, Gemini, chatgpt. 
Knowing how to properly prompt it is key. Learning new things and educating yourself is amazing. 
Trusting AI to compare DACs for you and take that information as a bible is nutty. 
 

AI can be fun, and can be useful in assembling information but beyond that I don't trust it much for subjective evaluation because AI can't hear and can't account for a paramount consideration-our own subjective preferences. What I regard as bright and fatiguing, you might hear as natural but detailed-not sure how AI can account for that. I also think the limited AI tools available to us are. pretty good at telling us what we want to hear. Fox News and MSNBC have proved that to be a fertile market.

Nevertheless, it a fun tool for preliminary information gathering.

WRT evaluating audio equipment, I find AI to be a database management tool (sort of like the box of Stereophile magazines many audiophiles used to keep), that also provides a level of cognitive and predictive analysis (sort of like an audio publication reviewer or a sales person at your local audio shop). The result is information that a user can choose to consider to whatever degree, or not at all.

Some may use AI for tactical information, i.e., does DAC X or Y provide a better S/N ratio? Others use it for more strategic input, i.e., if my preferences are for hard hitting bass, extended high frequencies, a maximum level of detail, and front-row, in-the-room soundstaging, which of the following DACs should sound best to me, and why?

The result is still simply information, not much different from a manufactuer’s marketing verbiage, a salesperson’s spiel, a reviewer’s conclusion, or a forum poster’s exclamations.  Like any information, from whatever source, the end user will need to decide how much weight to give to it.  Like most folks, I would never blindly accept information I receive from AI, without first comparing it to my own experiences, and whatever other information or data I can accumulate on the issue.  As several here have mentioned, hearing a component or speakers for yourself (preferrably in your own system) remains the gold standard.

Just as an aside, ask AI about these Tai Chi ads we are seeing on Audiogon and elsewhere. It is an interesting story.

@sholladay - I think that’s kind of obvious, no? Listening to music is a sensory experience… We may research wine based on ChatGPT based on reviews, but the experience of tasting the wine is always subjective and personal. And it’s that experience which is most important. 

 

As Mitch has described, AI can be used as a data consolidation and synthesis tool. I have been rather surprised by how apt the descriptions have been on audio components I currently own. I’m sure it’s based on aggregated comments, reviews, user comments, and of course marketing from the manufacturers.