The Audio 101 Thread

Micro detail firstly depends on the recording. With good recordings you do not need headphones with big treble peaks.

Macro detail equals appearance of clarity? That is confusing. When the music through a headphone or speaker (for that matter) sounds clear it indicates that the audio system gives you the audible impressions that you can envision the sounds in the recording as if a looking at a map of the world and low end is the South Pole and the highs end at the North Pole.

I cannot use absolute terms here, hence my metaphor. Listening to music through the years taught me that when you can imagine the sounds and pin point them in your mind as the music is playing, the music sounds clear. When that is not the case, the audio system is not clear and sounds lacking and wrong to my ears.

I strongly detest headphones and audio systems that lack that kind of clarity for me. Years of AKG headphones have spoiled me. Good thing the HD600 is pretty good.

 

Interactions between the harmonic distortion of amplifiers/tubes and transducers is really complicated* stuff along with other things like impedance matching, I think there's where whole system synergy and listening preferences comes into play. All these high performing THD%, DNR numbers doesn't really that much matter in the grand scale of things if you don't like the glare of ESS Sabre stuff for example or like how some would prefer the more mellower/forgiving sound signature of Gungnir Multibit over the higher performing Yggdrasil or etc. IMHO, It's good to know the tech/specs but actual listening, correct equipment matching and enjoyment is still king.

*THD related articles:
http://www.nutshellhifi.com/library/FindingCG.html
http://www.nutshellhifi.com/library/tinyamps.html

 

It's really nice you pointed out that topic. I am still trying to understand how exactly it works. So far I am there :

I talked last year with a sound engineer and he admitted that he performs distortion measurements in order to fulfill his requirements list because, although he may be able to interpret on a personal level the results, he also told me that he saw people reacting or behaving way differently than him. Since then I see distortion, to some extent, as something very personal in the interpretation. Best example : tube amps bring some distortion (2nd, 3rd, 4th and so on...) by "rounding" corners of square waves, but a lot of people like it :
http://www.innerfidelity.com/conten...tortion-plus-noise-part-1#M7coMWshEQBzfrsq.97

I also found some interesting work which helped me understanding the "human interpretation" level of distortion :
http://www.gedlee.com/downloads/THD_.pdf
Some people might not like Keith Howard from what I understood, but what he points out is interesting I think. To sum up : human beings tend liking some distortions and disliking others. I think the best is to try yourself in order to have an idea about what you personally like.

Hope that helped.

 

I suppose we have different points of origin for clarity. I don't think HD650 actually sucks at clarity, it isn't as good as HD800 (modded or not), but is a ton better than a lot of other wannabe totl hp-s.
HD650 has some veil even with good SETs, to my ears it manifests as some loss in dynamics in parts of lower treble and low bass. HD800 is ultimately lacking some dynamics (and clarity) in sub bass and at some regions in high treble vs good speakers.
I also think a sample of LCD-2 that had roughly similar fr response had less overall clarity than HD650. From memory I think I felt like Beyers were not as clear as HD650 either, despite being a lot brighter. T1 was more clear in bass. DT880 was worse across the fr band.

Another example of loss in clarity that is directly related to perceived dynamics is crossover between drivers. Even with good speakers I hear murk there with loss in perceived dynamics. Strangely enough passive low order filters sound less crappy than super sharp digital or amp based ones, like one in my Genelecs.
Shitty acoustics is another example, even after correcting for phase and fr response the clarity still lacks vs the same setup in good acoustics with relatively similar fr balance.

I am probably weird even by SBAF standards and fixate on dynamics too much.

 

May we talk about the factors that play into timbre?

I've thought of timbre as the realistic reproduction of the subtle nuances of voices and instruments. The realistic part of that is the key, of course, because I find it hard to define timbre except that things sound real and you can tell when a cymbal, violin, voice or bass sounds wonky.

I think there may be a number of factors, including distortion (maybe this could play into why n-order distortion sounds better to one person versus another), attack/sustain-the time element, and tonality.

However, regarding tonality, I've heard a number of dark sounding, for example, iems that have excellent timbre (dynamic driver ones, usually, but not always). Flat FR may not be the last word on timbre and tonality and the connection between overall FR and timbre is not 1:1.

Maybe it's not overall tonality, but the local topography in terms of a relative spike or gully compared to the frequencies directly above or below which can throw off things to our ears.

I'm just putting out some thoughts to start the conversation, looking forward to hearing your take.

 

timbre
ˈtambə/
noun
noun: timbre; plural noun: timbres

1. the character or quality of a musical sound or voice as distinct from its pitch and intensity.
   "trumpet mutes with different timbres"
   synonyms: tone, sound, sound quality, voice, voice quality, colour, tone colour, tonality, resonance,ring
   "the Czech orchestra have just the right timbre for Smetana"

Per wikipedia. Raspiness of a trumpet vs smoothness of a clarinet?

 

I think timbre is composition of fundamental tone + it's harmonics in a correct ratio and with realistic dynamical rendition (characteristic to a particular sound/instrument/voice).

 

Yes! The part in parenthesis is important because it covers that special something that makes an instrument sound just right like it should.

Fundamental tone and harmonics is a good start to what goes into making timbre what it is.

Whoa, that is not how I pronounce timbre at all ('tim-bur, like what you yell before a tree falls, all American English). But yes, that special raspiness of a trumpet or nasaliness of an oboe that is characteristic.

 

Okay, now that we've cut-and-pasted definitions, which answers the unasked question "What is timbre?", I'd love to hear more about my original question, which I'll adjust slightly:

May we talk about the factors that play into what makes an earphone excellent at reproducing timbre?




I'd also like to explain why I think these cut-and-pasted definitions of timbre aren't so helpful: They lack any real insight. What is Timbre? It is what gives an instrument it's sonic signature. Okay, but how? What is that signature made out of?

This doesn't give us insight into what makes an earphone better or worse at timbre, which is the real question. A better definition of timbre would give us a lot of clues toward that, just as the discussions on what is plankton did that.

 

I'll give it a try :

As mentioned previously :

This, for me, sum up very well the whole thing and we can go forward with the analysis :

• The dynamical rendition is important. It means for me that everything which helps pulling apart a frequency from the other is important. Therefore, a "neutral" FR in order to avoid masking effects looks promising.
• Any kind of distortion is going to have a bad effect on timbre in order to truly reproduce it if you have a good FR because you are going to pollute your harmonics. It's a hard criteria and one may take into consideration the instrument playing in order to narrow down what distortion is really affecting its timbre. For instance (and this is most likely untrue) I could imagine that a Guitar would excites only even harmonics whereas a trumpet would have a very strong 2nd, 3rd harmonics and then nothing more. Knowing that you may get something interesting.
• HOWEVER, I think a headphones which has a bad FR for one instrument may still do the job if it brings extra energy through distortion. But my feeling on that point is, even if you do so, distortion has always a "nasty" character and you wouldn't get something really clean. You will, somehow, loose in resolution.

That's it for my point, I wanted to do a correlation with structure such as open-back, closed and so on, but I think it's pointless. Or at least I don't have the knowledge for it. Feel free to rocket my message, that's only my point of view.

 

I think you're missing an essential aspect - timbre is not constant, and so one cannot address how a headphone can render timbre better or worse across the board. Rendering the timbre of a trumpet will be radically different from rendering the timbre of a flute, or a cello, or drums So you would have to constrain your observation on rendering timbre to the timbre of a specific instrument in ideal terms, or a type of instrument in less ideal terms (as in this specific stradivarius or violins in general). A headphone cannot be 'good at timbre' in a general sense.

 

That suggests that whatever it is, it is a valid thing, we may just be overloading a term in using 'timbre' to describe it. Can you try to describe what timbre reproduction means to you? Is it more meaningful with certain types of music?

 

That said, there are aspects of timbre that are...consistent: harmonic content, attac, decay and vibrato: http://hyperphysics.phy-astr.gsu.edu/hbase/Sound/timbre.html.

I still feel these will differ across different instruments and will tax a headphone in different ways for different instruments. Also, an instrument can have better or worse timbre - what happens when you reproduce bad timbre accurately?

BTW, my wife is a music major, I've got a bit of a lesson in this myself after asking her about it .

 

@GoodEnoughGear you are right saying that timbre depends on the instrument. There is a timbre for Trumpet, for Guitar and for any other instrument because of their mechanical construction/properties. Really, there is everything in what @Priidik said :

The definition of timbre is not the problem to me. What you suggest and which I think is true is that it is hard to get a headphones which reproduces well timbre of all instruments. For me, this is true simply because it is hard to find a headphones which is flat on the whole bandwidth. What follows is that you can chose a headphones because it renders better the timbre of X instrument or Y instrument. For me, it is necessary to get to this point in order to proceed with our talk.
On the other hand, the question of @Kunlun was different, at least to my ears. He is asking what parameters should we optimize for X timbre or Y timbre, thus leading to the discussion regarding distortion, FR, why not CSD and so on...

 

The electronic music reference is interesting - a smooth generated sine wave would have no (or neutral?) timbre. Music created this way without timbre would be...muzak? For electronic music to elevate itself it would need to introduce electronic timbre: harmonics, vibrato, modified attack and decay and so on. Or be notable for its absence of timbre, as opposed to just devoid.

So I can't escape feeling that timbre reproduction must be a composite of other performance aspects of reproduction in a headphone. You would need accurate tonal reproduction across the FR, good macro and microdetail, resolving ability etc, as timbre would be embedded anywhere in the music. In a way it IS the (character of the) music?

 

@MoatsArt

 

I agree with this, though if you've heard the output of an honest-to-goodness analogue synth being piped directly into a nice stereo then you will have an audio-image of what it's supposed to sound like that will be indelibly imprinted in your brain

I suspect that the reason that we are having trouble with timbre is that it's a kind of third-order descriptor. At the low level we have things like frequency response, impulse response, slew rate, THD and so on that we can measure and assign to gear. Above that we have things like microdynamics, macrodynamics, tonality, imaging, clarity, speed and so on which we can sort of correlate with the technical measurements but without a really clear mapping between them. I feel like timbre is on a third level above even these - it's something which we can sort of correlate to microdynamics, macrodynamics, tonality and so on but again without a really clear mapping. So all we have are hints and guesses.

EDIT: i.e., wot @GoodEnoughGear said

That being said, my best guess as to what is required for good timbral reproduction:

• an EVEN frequency response (overall tilt doesn't matter so much, but peaks and troughs on the way do)
• no phase shifts
• no ringing
• great microdynamics
• accurate transient response

The Schiit multibit w/ burrito upsampling filter seems to be particularly excellent at timbre, so maybe there are some hints there.