You said your measurements showed better distortion, noise and power.
This has been one heck of an inquiry indeed. From day one I felt that something wasn't right so I wanted to figure it out and stay in communication. As we were up in the mountains with no reliable internet service for a week, that wasn't easy. Talk about perfect timing and a ruined vacation... but customer commitment comes first.
Our internal measurements are not performed the way Amir performs measurements, so we needed to create apples to apples conditions. AP handbook differs in some of its recommendations, but it's not worth getting into that rabbit hole now.
Since getting back after thanksgiving, we engaged with our staff to make new measurements and run tests. All the while, we had to juggle shipping and covid distancing, so quite a dance all around.
Here is what we know so far:
- Initially I wasn't clear what the issue was (THD, noise, power). This Wednesday it finally dawned on me -- I had forgotten that the digital gain pot is the immovable object. In all fairness, we designed the product two years ago and since then designed many others, so it wasn't fresh in my mind anymore (I wear many hats, cannot remember everything).
- Realizing this caused me to sit down, retrace the design decisions and write this paper, which I published today, if anything as a memo to myself (and anyone else who would care to read it). The end result: HiFi-M8 THD will remain at 0.004%. That will not change unless we wire across the digital pot. I suspect people do want volume control, so that's not a likely mod. Otherwise, we can easily bring distortion another decimal point down.
- Regarding noise, first there is the driver, which is in progress. And then, there is also the output level, about which there is a bit of a debate around the office. Amir wants to measure the output at 2V, which assumes that the output is a line output. But HiFi-M8 is not a desktop DAC and it doesn't have a line output. It has 4 headphone outputs, all different. Which one is fair to choose as a line output? And if we have to dial down from FS in digital domain to achieve 2V, would we be artificially limiting dynamic range? We are at a crossroads on this at the moment. Hope clarity will emerge with additional discussions.
- Lastly, regarding power, HiFi-M8 behaves better than DACport HD, which was recommended on this forum. And yet HiFi-M8 was failed. Amir was frustrated with the driver, but subjectivity shouldn't enter into his judgements, so there must be something else. Did he not have the woefully inefficient Dan Clark headphones earlier? We don't have them, so we cannot check with them. We check with LCD2, etc. and by the time you hear any distortion it's so loud, your ears hurt. Should you really be listening (and evaluating) at those levels?
In summary, we are happy to continue the inquiry (time permitting) and trust that clarity will emerge. At this point we know that THD will not improve. We may or may not continue with power as the effort to match the experimental conditions is starting to become a serious resource drain vis-a-vie other standing commitments. We still believe that noise (aka dynamic range, aka resolution) will improve once the driver is ready.
EDIT: Time permitting, we have been reading up on this forum, to study Amir's process and to understand the general discourse. We hadn't been here much before and it's a rather large resource at this point (how does he even have the time?)
Knowing what we know now, we should plan to design future products to make it easier to evaluate them within this framework (hint: 2V output). Having to jump through these same hoops again is definitely not an attractive proposition.