performance per dollar... how to measure?

[Dana reed]

Here's a picture of the ratio of SINAD/$ for all the DACs measured here with SINAD>100 (if you let the <100 ones in, then the dongles like the apple USB C dominate due to price). Obviously there are other factors in play with many of these comparisons, as some of these only have USB, some also include a HP amp while some don't, some have XLR and RCA, etc.
The top one here is the zorloo, but that doesn't have very high output level, and the prices vary around a lot for the topping products depending on where you get them (I used either msrp, or what Amir listed paying for them)

FYI, the items here are still ordered the same as in the master graph by increasing SINAD, just with each one divided by price. Thoughts on how to systematically evaluate bang for the buck?

 

[amirm]

I was thinking about creating a scatter chart with Sinad on one axis and price in the other. But as you have noticed, we probably need to put similar DACs together rather than everything in one chart.

 

[Ron Texas]

Before doing anything like that, it might be a good idea to establish a floor of audibility. There is no doubt about it, the testing here has proved select DAC's selling for $250 and lower are transparent. They are the greatest buys in audio.

 

[Dana reed]

yeah, I guess even the apple dongle would be transparent to my ears. You just need more gain than it can give in a lot of cases

 

[Soniclife]

How do you account for the different functionality offered? That has a crucial impact on value, but not in the same way for each person.

 

[Dana reed]

yeah, it's annoying to do a scatter plot without filtering by some other criteria, as there are too many items to use color coding, and labels with callouts is tedious. At least in excel I don't know a way to automatically arrange labels to not overlap. I had to put price on a log scale to make room in the 90-900 dollar range.

Certainly other features like input capability voltage out, balanced outputs, headphone amp, are all different here. The hp laptop has similar performance to the yggdrasil, but has a few features that the yggy does not.

 

[Dana reed]

would also have to put some bars around price for some of these, and also should put some error bards on SINAD, as these are mostly measurement of one unit, and wouldn't take into account variability in manufacturing, and if a few others would pony up and buy another AP analyzer and set up identical measurement rigs, we could put some confidence intervals around these numbers. Like how we sometimes send round robin samples around to multiple labs at my work and see how well all the measurements agree. Of course we're getting paid to do that, and didn't have to pay for the capital equipment budget so it's more likely to get done.