Conceptual Framework for Identifying System Bottlenecks

I saw a post from last year where a poster asked for advice about his digital front end, whether source or pre would be better to change out, and he was besieged with comments about fuses and power outlets and other nonsense, running him off the site. Although it's too late to help that poor soul, I hope this will help some of you if you're trying to identify system bottlenecks.

Definition of term: "bottleneck" in this case means a component that is significantly holding back the other components, and which can be identified with one's own hearing, using instant switching between the components being tested, and taking careful notes of specifics. This is *not* about "slight perceived differences in timbre." YMMV.

To that end, I recently compared:

6 Sources (2 transports, 2 network streamers, 1 NAS, 1 rPi: price range $90-$3500)
6 DACs (3 ESS, 2 TI, 1 CS: price range $125-$2500)
5 Preamps (3 SS, 1 tube, 1 passive: price range: $50-$3500)
3 Power amps (Class AB, Class D, integrated AB: price range: $500-$3500)

My goal was to identify bottlenecks *between those components only*. (Meaning, I used the same interconnects and power cables for all tests for maximum consistency, and which, although they are not cured in Yak piss for 3 months prior to "break-in", actually include measurements that I can trust. While there might be slight differences in sound between cables, there is zero evidence of any kind that they cause a "bottleneck" except in the most demanding professional environments or terrible power situations, neither of which apply here.)

I used my reference speakers throughout. I will not name any names, as cognitive bias is the hobgoblin of little minds and there is so much unconscious brand identity in this hobby. I did all tests using balanced connections wherever possible, to minimize level-matching, and unbalanced to unbalanced when necessary.

It is not possible to test every single permutation in depth, but some things became apparent early on, and consistently throughout several weeks of testing. In other words, these were real-world tests, not speculative quasi-religious rituals of consumer self-medication. Pfffft.

Here are my observations:

For my system, the biggest bottleneck was the Class D amp (Hypex nCore). It is the same module used in much more expensive amps, and the amp was built by a very good company, but it made DACs and sources sound virtually identical, whether $2500 or $250. This was the biggest surprise to me, as I've been an enthusiastic disciple of Class D for weight/energy/efficiency reasons. My Class AB integrated of a similar price class just blew it out of the water. This was a real eye-opener to me that measurements (like ASR does) in no way determines sound quality.

Put it this way: the Hypex allowed me to hear all the cymbals on a tambourine; the Class AB allowed me to hear where they were in relation to each other. They both have tremendous detail, but the AB was vastly more 3-dimensional. 

The difference was even bigger when comparing pre-amps--which i think in my system is the most important piece at this point. Maybe that will change, but the fully balanced differential circuit end to end, Class A > AB, is just amazing. 

Other thoughts: 

Differences between DAC chips themselves were negligible. The smallest difference was in chips from same manufacturer, close in model (ie, ESS 9006, 9016, 9028Pro); there was audible difference between chip manufacturers, though none sounded "bad". My personal preference in my own system was for ESS chips, followed by TI. I found the CS to be slightly oversharpened and grainy, but it was still very good and I only had one version on hand.

Sources: Output Vrms might be skewing things here, but I preferred the disc transports, followed by LAN files, followed by high-res streaming, though streaming remains my main source. This is one test area where I'm open to testing different routers and ethernet cables. I compared MP3, redbook, SACD, ISO, Flac, etc.) Btw, stock rPi USB out is amazing, and easily fits in a $10K+ system as a reference source.

I don't know if this helps anybody, but I learned a lot and had a ton of fun doing these tests. I've sold my class D amp and high-end dedicated DAC and am focusing on fully balanced differential Class A/AB signal chains for now. 
It is neigh impossible to evaluate your posting unless you use specific brand and models.

If you wanted to be more helpful you could be specific in talking about what you heard in amps as well as DACs.

The idea of selecting a component because we know something about it’s construction (Tweeter fabric, class of amp, DAC chip, etc.) is not something I have found can be relied on.

OP, I’m that poor soul you speak of. I ended up ignoring the loons here and upgraded both my preamp AND my digital source lol.
"Put it this way: the Hypex allowed me to hear all the cymbals on a tambourine; the Class AB allowed me to hear where they were in relation to each other"

I find this hard to believe the AB amp allowed you to  hear where the cymbals on a tambourine were in relation to each other? Aren't they about a few inches apart? 
Most systems are smitten with compounding errors some of them negating each other such as a dull cartridge making bright speakers sound better. So, each system is an individual, room included. What is good for one system may not be good for another. The vast majority of people can not even apply a method as they have no idea where they are starting from. You have to measure the in-room frequency response. If you think it is anything close to flat guess again. 
Having said all this the weakest link is by far the loudspeakers. Next is the cartridge followed by the amplifier. The differences between good electronics is relatively minor except for advanced processors such as the Anthem STR preamp and the Trinnov Amethyst. The Anthem is a fabulous value at less than 1/2 the price of the Trinnov. Both units will dramatically increase your understanding of audio in general.