I have to disagree with that statement, especially since it’s such a blanket statement.
There is a great deal of difference between something like a Schiit Magni (which is a very good “cheap” amp) and something like their MJ3 or even Jotunheim 3, not to mention any number of amps getting into the thousands of $. Same goes for DACs.
Of course you get into the diminishing marginal returns problem the higher you go up the cost ladder, but saying there is little difference or that it doesn’t matter is just wrong. I’ve definitely been able to hear a difference as I’ve moved up the ladder with better gear.
What you can say is that you can’t hear a difference, or that it is a negligible improvement to you. Yes transducers (headphones) matter the most, and the levels of distortion in a transducer is orders of magnitude greater than that in an amp or DAC. Which is why the ASR folks are chasing shibboleths when they fret over .005 vs .0005 levels of THD. But saying something like amps and DACs don’t matter is simply incorrect.
Definitely. Case in point: The Canever ZeroUno DAC/AMP sounds different to anything I’ve owned prior to getting this. It’s a pretty unique beast across the board, from both a hardware and software design.
Also, some DAC’s aren’t designed to handle the higher sampling frequencies. A ripped CD file upscaled via PGGB will surely sound different compared to a DAC that cannot support the added processing.
Again I have to reiterate this point, my position is not that differences don’t exist, it’s that the differences that do exist are inconsequential, particularly in comparison to the differences you get from FR adjustments. It’s a point that shouldn’t be controversial, but unfortunately seems to be. And I say this purely as a result of having done blind testing for many of these devices - even for some of the products you’ve mentioned.
I’m not going to say you’ll have the same experience I did when doing blind tests, but it’s a worthwhile experience to have to reveal the degree to which this stuff does or doesn’t matter. Even without blind testing, SPL differences are a hell of a drug.
I don’t want to gatekeep ‘the truth about amps and dacs’ or whatever behind blind testing and I can appreciate the impulse to reject kinds of statements I’m making about the degree to which DACs and amps matter - I’ve been there too. If folks don’t want to do blind testing, that’s understandable - it’s not the easiest thing to do correctly. But if you really do want to test your assumptions on this stuff, it’s rather necessary.
Ultimately I don’t care if people want to spend more money and get their joy from DACs and amps, that’s up to them. What I’m pushing back on is the idea that Amps and DACs can ‘fix’ headphones, while EQ cannot. I would caution anyone who encounters statements like that.
I understand your point, but again they are not incosequential. They may be inconsequential to you, but that’s about as far as you can go with this statement.
That is something I can definitely agree with, some headphones simply can’t be “fixed”. Now this is my subjective opinion, but the Grell OAE-1 is a case in point. There was no amount of eq that could fix that headphone for me. It had comfort and fit issues as well, but I never could get it to sound “right”, and simply gave up on it.
Depends on the equipment used to conduct blind testing. I would argue that if one compares headphones with a a lower end DAC/AMP setup that cannot support high upscaling, then run the tests again with a higher end setup that supports the upscaling, the results may indeed vary. I agree that a different DAC/AMP setup will not “fix” a given headphone, but some headphones do indeed benefit from better electronics.
Regarding frequency response graphs, they do not explain the perceived differences with low level detail playback. The higher end headphones tend to often be cited as having greater detail, even if they do not measure as well as a lower priced headphone.
Getting EQ right is a challenge. I have found that some tools, such as Sonarworks, more often than not, does provide decent EQ. The ability to add EQ in PGGB can also improve the playback.
Whether the end user likes the difference is up for debate.
Sure, though I conjecture that upon doing proper controlled level matched testing, many of the differences people report with traditional comparisons would disappear. Auditory memory alone is a very tricky thing.
But, your point is a good one, I do not personally care about this stuff after having done those tests. Others on our team like Cameron, who have also done those same blind tests, do still find there are reasons (very small minor reasons but still reasons) to care. And, so long as people aren’t suggesting “you need to spend x amount of money on your chain for this or that headphone, don’t bother with this EQ stuff”, then I’m all for it.
I actually think with the right tools and enough time, even that one could be fixed with EQ. It’s just hard to do if your only sense of the FR is the graph, and not how it’s behaving in-situ.
It really does, the problem is you’re still thinking of FR in terms of the graph alone, and that’s not all there is to FR. I have to say this time and time again, and it bears repeating here as well, what the FR is doing at your eardrum is not the graph. It is not the Harman target, it is not ‘tonality’, it is not what you see measured on squiglink and various other sources.
In-situ response is the missing link between typical measured performance and ‘technicalities’, people just struggle to interpret things like “detail” in terms of the in-situ response. I suppose that shouldn’t be surprising because they’re hearing the effects it imparts on the music they’re listening to. But anyone who really takes time to examine effects such as ‘detail’ or other subjective qualities closely will be able to eventually interpret these things in terms of FR relationships as well.
So, it sounds like what you are saying is that the “actual frequency response” the end user hears cannot be adequately measured for display for viewing. There are differences between driver designs and performance characteristics, which must account for the ability to reproduce low level detail clearly. IMHO, no amount of EQ can address this aspect of headphone listening.
There must be additional measurements developers use to design headphones that the public is not privy to.
I read Sennheiser measures their drivers rather closely, and discard any that don’t meet specs. I’ve experienced driver variations with several brands other than Sennheiser, which throws a spanner into trying to EQ the cans.
But again, how do you know this isn’t FR? Just because your experience of a headphone is better or worse for low level detail doesn’t mean there’s a “low level detail” metric only designers are privy to. ‘Detail’ is a subjective description of your experience, of the music you’re listening to with this playback equipment. It sounds more detailed to you. There is no acoustic category of ‘detail’, because that requires a human being to perceive it.
Moreover, the best engineers are designing for in-situ FR performance, and with enough resources are even measuring FR on real humans. Yes, they’re also designing for things like low harmonic distortion and ideally consistency, but I can promise you there is no secret metric out there that only engineers can see. They’ll tell you basically the same thing about FR in-situ I just did, and we know this because we talk to engineers regularly. Where there may be some disagreement is in terms of what ‘good’ ultimately is. But as far as the what of it - the metrics, almost everyone is in agreement that it’s FR in situ.
Sure, the ability to distinguish low level detail is part of the overall frequency response, but some drivers are better suited to achieve this than others. EQ can help to some extent, but it can’t fix everything.
Take the review of the Sony MDR-Z1R from Tyll Hertsens. He basically said that the 70mm dynamic driver causes anomalies that (I don’t think) can be fixed by EQ. Also remember a video from Sennheiser that talked about strange things that can occur once the dynamic driver exceeds 38mm in size.
What I take away from this is that the reviews and graphs can provide some basic idea about how a headphone sounds, but only in-situ listening can tell the end user if the headphone is worth pursuing, and how much EQ (if any) may be needed to optimize playback.
This would’ve been at a time when people didn’t realize how differently HpTF can trend between rig ears and human ears. Driver size is a parameter that can influence a lot, but the end result is what matters not the fact that it’s a specific driver size. This is a matter of necessary and sufficient conditions.
With that said, you may come across things like deep phase cancellations, and while it’s exceedingly rare, there are also some situations where headphones aren’t minimum phase. So the possibility of driver-based limitations and various other design issues can lead to anomalies.
I think this is in general a good outlook to have. Though I do think the FR graph is useful as an initial guideline. It won’t tell you how it sounds to you, but it does tell you how it could sound to any given person. That’s better than nothing.
I feel that by “shaving down a bit” some of the rougher parts of the FR from the graph, I got my OAE1 to sound pretty good while retaining their unique presentation. However I’ve never tried using Owliophile on them. Imma try that.
