Help me understand audio graph's?

[Psittac]

Can someone post graph's to any of the headphones dac's or amps in my sig and explain to me how that is going to manifest to my ear holes?  I'm more of a hands on person but I think that if someone put pictures to what I'm experiencing maybe I'll get a greater respect for graph's.  My philosophy has been to hear what a reviewer says about them and experience myself, I think now that I have enough of a handle on terms used I can possibly transition partially to graph's.

[191x7]

Well, you can, like everyone else, google "headphone_name frequency response graph" and see for yourself.

Reading the graphs is easy, they pretty-much show how loud (pronounced) every measured frequency is. And for the frequencies you have low (bass), mid and high (trebble).

[Psittac]

1 hour ago, 191x7 said:

Well, you can, like everyone else, google "headphone_name frequency response graph" and see for yourself.

Reading the graphs is easy, they pretty-much show how loud (pronounced) every measured frequency is. And for the frequencies you have low (bass), mid and high (trebble).

Yeah I probably should, but it's just something that doesn't translate to me, I get high mid and low but that's not much of a nuanced look into a headphone.  I'm just not a visual to audio learner, back in elementary school I quit playing the saxophone because I couldn't read the notes, just didn't make sense to me so I just mimicked the people next to me.  I was hoping that someone could elaborate on something that would bridge the gap for me but it looks like it's something I'm just going to have to learn on my own if at all.

[CTR640]

25 minutes ago, Psittac said:

Yeah I probably should, but it's just something that doesn't translate to me, I get high mid and low but that's not much of a nuanced look into a headphone.  I'm just not a visual to audio learner, back in elementary school I quit playing the saxophone because I couldn't read the notes, just didn't make sense to me so I just mimicked the people next to me.  I was hoping that someone could elaborate on something that would bridge the gap for me but it looks like it's something I'm just going to have to learn on my own if at all.

That's so me too! The notes are like mathematical chinese gibberish to me, I never could understand them. Now fast forward to present, I just can't understand the audio visual graphs and if I'm correct, they are also called FR. So a super highly evolved intelligent human can read how the highs and lows are lol. Anyway, audio is personally meant to enjoy and not to be a scientist. 

[Spuriae]

https://diyaudioheaven.wordpress.com/tutorials/how-to-interpret-graphs/frequency-response/ is a good guide for mapping frequency response graphs to subjective impressions. Keep in mind when viewing graphs that different test rigs can produce significantly different measurements. Similarly, especially in the low bass and high treble, measurements may not tell the whole story because the shape of the listener's head and ears can have a significant effect on the sound they hear.

 

If you to be able to intuitively make these connections yourself, or if you want to be able to produce rough frequency response graphs by ear, check out Harman's listener training software: http://harmanhowtolisten.blogspot.com/2011/01/welcome-to-how-to-listen.html. It has you listen to music with certain frequency ranges boosted or attenuated, and practice matching the sound you hear to the correct graph.

 

AutoEQ has a nice index of graphs here: https://github.com/jaakkopasanen/AutoEq/blob/master/results/INDEX.md. For pure frequency response, oratory1990 and Crinacle are the measurements that are most applicable to most listeners since they use the test rigs most standardized to the average human; Reference Audio Analyzer is the least applicable. The red line in the graphs represents the difference between the measurement and "flat", and is the one you would compare to the frequency response guide.

 

Depending on where you get your measurements, they may have a different idea of what neutral ought to be. Whenever I'm dealing with an unfamiliar source, I usually compare the measurements of the headphones I'm interested in to the same tester's measurements of the HD650 (which, at least to me, are a good neutral reference in the mids and treble). Even then, I often have to mentally adjust one of the graphs up or down because the tests were not properly volume-matched.

 

Spoiler

Example: comparing the Harmonicdyne Zeus to the HD650, using Crinacle's measurements

 

The green line is the difference between the frequency response for the Zeus and the HD650, and I've made it partially transparent to insert on top of the frequency response guide.

 

 

Based only on these measurements (I haven't heard the Zeus):

• the Zeus has more upper bass/lower mids than the HD650. The low end is warm- syrupy even- but doesn't quite reach the level of being obnoxiously muddy or boomy (the general ~3dB boost from around 70Hz to 300Hz). Extension in the low bass is reasonably good for an open back but nothing particularly special (the graph is close to flat below 60Hz).
• the Zeus has some small dips in the upper mids compared to the HD650. The timbre might occasionally be a touch too distant (the dip around 1.5kHz) or relaxed (the dip around 3kHz) to sound completely natural, but is well within the range of reasonable personal preference. It does this without giving up too much in the way of clarity – though it is definitely not the most detailed headphone in absolute terms (the entire range from 1kHz to 4kHz isn't too recessed, and the level at 2kHz is close to the HD650).
• the Zeus has better treble accuracy for the average ear than the HD650 (note the green line usually being closer to the dotted line, or the neutral target curve, than the HD650, which is acting as the 0dB reference in this graph, above 6kHz); its treble is relatively neutral and reasonably textured without being offensively grainy (the treble doesn't have any massive spikes, while roughly following the contours of the neutral target curve). The Zeus isn't exceptionally energetic (the average treble level isn't much higher than the HD650), but it does mostly capture the edge and energy which the HD650 can sometimes lack (particularly good target curve tracking compared to the more recessed HD650 from 6-8kHz, which is where the initial edge of percussion mostly resides). It shouldn't be piercing or sibilant for most users (response never significantly exceeds the HD650 or target curve).
  • High treble measurements tend to be inaccurate, but the loss of tracking around 9kHz implies that very fast percussion impacts might sound slightly rounded-off.

 

The above works for headphones, but there isn't really anything similar for DACs or Amps because the measured differences between those usually aren't nearly as significant, especially between well-designed products.

 

There are a few other types of graphs that can correspond with things we hear (for instance, soundstage), but they aren't as significant as frequency response, and are also much harder to explain.

[spartaman64]

this is a very contentious topic. the main measurement that the debate is centered around is sinad which is signal to noise + distortion. some say that sinad is pretty much the end all be all measurement and as long as an dac/amp has reached a certain sinad it sounds exactly the same as any other dac/amp that hit that number.

other people say sinad only tells you if a product is broken and there are differences between dacs/amps even with the same sinad. 

 

im in sort of a middle ground because ive heard differences between dacs/amps with > 90 sinad that group A says should have no difference but ive also heard minimal / no difference between dac/amps that group B says should be a huge improvement. also in general i havent heard any huge improvements at least compared with changing headphones