JaZZ

The headphone I like best with my iFP-395T is the Etymotic ER-4X (40 ohm); second best is PX 200, third best PX 100, which has a special synergy with the iRiver to my ears, but doesn't isolate at all. You just have to turn the bass down by -6 dB at 50 Hz and -3 dB at 200 Hz as well as slightly increasing the treble (+3 dB at 14 kHz and maybe +3 dB at 3 kHz) -- I just don't know if the iFP-390 has the same EQ, but you get the point. I like the PX models. The PX 200 is somewhat critical in terms of individual anatomics due to the small contact surface (seal!) though.

No doubt, I'll buy the HD 650 as soon as it's in the stores. Anyhow I trust Sennheiser to be able to really improve the HD 600. And from the single criteria I could judge it appeared like an improvement: deeper, tighter and cleaner bass, more brilliant and clear treble... The uncertainty is the sonic balance: the bass was too strong for my taste, and the midrange lacked the intimacy and warmth of the HD 600's. But as has been said: the copy I've auditioned (with lousy music from probably lousy electronics) possibly wasn't broken in. So this is a very provisional report. However, now we (at least I) know that the HD 650 indeed sounds different from its predecessor. And I'm not sure if the less warmth isn't in fact a good thing. I preferred the HD 600 for some time without the foam pads; this made the bass leaner and the treble clearer, also at the expense of intimacy and warmth, but I preferred it to the stock sound, although it seemed to damage its coherence to a certain degree. Let's hope Sennheiser's own «mods» manage to combine the best of both worlds, and -- this my personal concern -- that the bass isn't too strong.

Rameish... ...I really hope you can wait. But till then take care to have a low-impedance source at your disposal!

Although I have no info about transformer distortion, this sounds very plausible (thinking of magnetic saturation of the iron core). And it corresponds exactly to my listening experience with the Reference Tools: Although close to the typical (resistor-based) passive attenuator, the sonic signature was leaning somewhat towards headphone amps: warmer than the pot-based attenuator, like «filling the gaps between the notes» - creating a greater coherency, which is what's missing a bit with the pot type, which in turn seems to sound downright too clean at times. Given that the amps' distinct sonic characters are based on their harmonic-distortion spectra, it's a small step to postulate an equal behavior for transformers, just in a clearly milder occurence - which is deserving from a puristic point of view. The output voltage isn't the problem, but the output impedance. My former DAC, the Theta Pro basic II, offered 6 volt (and 5 ohm), whereas my now Bel Canto DAC2 contents itself with modest 2 volt (and 20 ohm). (BTW, 2 volt are a common value; few DACs offer less than that, but some even more.) Although it has less volume headroom than the Theta, it works equally well with driving headphones directly, and I never reach the upper limit of the volume control. 2 volt is enough.

Rameish... ...it should arrive at you in four days or so. In the meantime... well, enjoy your life! And so do afterwards... Firefox... ...I don't get your post. Is it that you want to insinuate that the transformer needs a sort of «buffer» to sound loud enough? That's certainly not the case. It's designed to provide a high input and a low output impedance, so to create «normal» circumstances for both line-out stage and headphone - thus it's a buffer itself; that's its actual raison d'être. The crux is the Reference Tool's low efficiency, which logically is a matter of (the extremely low) price. John Chapman from Bent Audio, with whom I had an extensive exchange of ideas, has said to be able to build a transformer-based attenuator for headphones + line outs with the same 1:1 efficiency as a resistor-based attenuator. But let's not talk about the price...

matbon... ...sorry, I don't understand your question.

In fact line-out amps are of a similar design and not so much weaker than headphone amps - 2 volt is really enough to drive headphones, and some are even stronger than that; the only caveat is their ability to supply enough current with low impedances. I have already mentioned that some are even explicitly used as headphone amps (Sostenuto-1 portable DAC + headphone amp).

As I said: I'm not the one to do some soldering inside a CD player or DAC, so removing op-amps is not my cup of tea; and unfortunately amps are indispensable when it comes to drive sound transducers... I just don't like the idea of having a strong signal from an output stage amp, boosting it up with a second amp in series to a level way too high to be useful with any *headphone* so that it has to be attenuated by a volume pot again, notably to a level clearly below the maximum of the original line-out signal (spoken in volts). Here's an example for a quite cheap transformer-based attenuator specially designed for headphone direct-path operation. Unfortunately it's not loud enough for normal sources and headphones, at least it requires low output impedances, high-sensitivity headphones (preferably the HD 600) and loud recordings... But the sound is excellent, slightly different from resistor-based attenuators, but on an almost comparable quality level as to the signal accuracy. Explain! No, you don't need one. It wouldn't hurt the signal accuracy that much as an amplification stage though, but the unbuffered direct signal path provides the greater sonic accuracy. In the case of the transformer, this one is itself the buffer; and in the case of the resistor (= voltage divider) solution, the 500-ohm pot doesn't corrupt the signal that much. You shouldn't see headphones as speakers; they don't necessarily need a really high damping factor to work properly. Some of them are even designed to work best with serial resistors (some Beyer models, Etymotic ER-4 series), and some (OTL) tube amps have a similar output impedance as the effective serial resistance of the 500-ohm pot in this case. So what's your point? Do you think amps alter the signal less than an unbuffered 500-ohm pot or a dedicated transformer? In contrast to you I have some experience with the direct path, and I have systematically tested the impact of serial resistances and low-impedance loads. With an appropriate source, you get unbeatable accuracy and neutrality. It's a well-kept secret that amps aren't really neutral - none is, not in the least. Actually that's quite logical, because every one sounds different than the other. So if you can renounce an amplification stage, do it! And be it just out of curiosity, to experience the true nature of your amp's sound. You may like the more colorful presentation through the amp better though, but don't mix this up with accuracy. Also analog discs often sound better than CDs, even when the same digital master has been used. Rameish... ...«sonic truth» explicitly referring to the path «source line-out --> headphone». Not absolutely (of course not!) . Sometimes it's hard not to be misunderstood... BTW: the pot is on the way.

Well, I still don't see how direct-path experiments are expensive. Basically all you need is a (cheap) potentiometer. You can even use some stepped switch instead (e.g. an input selector) and equip it with resistors corresponding to a 500-ohm pot at the different settings. If you decide to make a lasting solution out of it, there's just some handicraft with some available aluminum housings to be accomplished to mount RCA and headphone jacks - and that's it. Maybe US$ 25 totally. I for one wouldn't dare to swap op-amps in a player, so if I don't have a low-impedance source, I'm no direct-path candidate... Yes, I know the sonic merits of headphone amps - in the form of a Corda HA-2 and an Earmax Pro. I may be a «direct-path advocate» , but I use the amps clearly more often than my 500-ohm pot: because of the «better» sound. Don't get me wrong: the «better» (= more euphonic) sound in fact is a fake and doesn't represent greater accuracy, but quite the opposite! The sound directly from the source (or its proprietary line-out amp, resp.) without the detour through an additional amplification stage represents the sonic truth - in the ideal case.

But beware: it's rather likely that this output has a high impedance as well.

Firefox... ...why is it a waste of money? It's an interesting experiment (not just buying gear, but also being sort of creative), and it's really cheap, just the price of a 500-ohm stereo potentiometer (though hard to find) or a switched attenuator, resp., with as a result, in the ideal case, an accuracy unbeatable even by the most expensive amps. And even if you don't like the sound: who isn't interested to know how his amp or any amp alters the original source sound? Of course the precondition of a low output impedance has to be observed. But this isn't really hard to do. Rameish... ...OK, I'll have a look in the store.

Hi Rameish! It seems that I've been lucky to find some 500-ohm pots. They're nothing special, a no-name product, with the usual wire inside as it seems, but sufficiently precise. If I find some more I could send them to you if you like... ...but it may be easier to use a switched attenuator instead; it could be a simple one with say 6 or 12 positions for test purposes. You just have to equip it with the equivalent resistors.

Rameish... ...just a little misunderstanding: the ideal pot value is 500 ohm, not one in the kiloohm range... The serial resistance of a 100 kiloohm pot is much too high in view of the say 300-ohm drivers.

Yeah, what went wrong...? First of all: rule number one has been disregarded: choose a DAC/CDP line out with a low output impedance! I think 100 ohm should be seen as an upper limit - the more so as the indispensable passive attenuator will add even more resistance. As already mentioned in an earlier thread below, it looks like the majority of common players have line-out impedances of 200 ohm and above. Don't waste your time which such players! They're not suitable for direct-path purposes. Rameish, you have tried several different op amps with the Rotel player. How is the output impedance with them? The 600 ohm you first mentioned is way too much! Imagine the drastically reduced damping factor and its impact on the frequency response, in the form of a huge hump around the resonance frequency - even more so with 32-ohm headphones! Hence your «muffled sound». You didn't mention the attenuator, but it looks like you were using a potentiometer. With which value? It's very important for the resulting sound. 500 ohm has turned out to be ideal, 1000 ohm will work too, but sounds slightly more «muffled». I even wonder if you had to attenuate the signal at all when I think of the high output impedance... A scientific, systematic approach for this «test» would have been to choose an appropriate source which fulfills the main criterion instead of a random player which turns out to be unsuitable. Furthermore, to evaluate the accuracy of the direct-path sound merely based on personal esthetic measures doesn't make sense. My goal was to detect the impact from the low and complex load represented by a headphone on the output signal (because - as you stated - it is in fact «not designed» to drive other loads than amp inputs), and with my corresponding test configuration and my equipment I came to the result that it hadn't any, so the direct path could be seen as a virtually perfect signal transfer from the source to the headphones' voice coils with no perceivable sound alteration - as would happen with additional electronics in the signal path. It wasn't my main intention to achieve a better sound; I just wanted to know how the source signal really sounds, in the awareness of the unignorable sound alterations caused by all amps. In your case it seems logical that what you get has little in common with the original source sound and at the same time lacks a lot with esthetics. My experience is that in terms of esthetics the direct path has no advantage over the classic amplifier-based operation. I have described this in the earlier thread. The reason isn't clear, but what's clear is that amps add some convenient euphonic colors to the sound - and thus may mask some deficits with the source signal (keyword «digital artifacts»). You may be tempted to jump on the conservative bandwagon and simply state: «Line-out amps are not designed to drive headphones.» This doesn't prove anything. But you can test the impact of a headphone as a load on its signal accuracy (like I've already described earlier): Listen to some music through a headphone amp. Then switch an additional headphone in parallel to the amp's input (thus to the source) and try to detect if this makes a difference to the sound. I bet it doesn't. So the only remaining criterion is: to what extent does the serial resistance (output impedance and attenuator) between source and headphone affect the signal? If you use a source with very low output impedance and a 500-ohm potentiometer, the average effective serial resistance will be in the same range as OTL tube amps - thus absolutely practicable. Just look at the thousands of people who have their headphones plugged into their soundcards' line outs! This works really fine if your soundcard doesn't have a dedicated headphone jack, and in most cases it even sounds better - and even louder! - than the headphone output. Obviously line outs of soundcards usually have low output impedances. The great advantage here: you don't need a passive attenuator, the computer software will do it. Or there's a guy on Head-Fi called aos. He builds portable DACs with a built-in headphone amp - which in fact is a line-out stage! Who says line-out amps can't drive low impedances! This may apply to some of them because of current supply limitations, but the main criterion seems to be their output impedance. Well, strictly speaking the term «direct path» isn't really accurate, I agree. There can be no direct path at all in sound reproduction; the whole signal chain consists of dozens of processes, from the microphone over the mixing console to the A/D converter... etc. I use it in the sense of «as direct as possible». And there are indeed some superfluous stages: the preamp is by nature completely senseless for normal line-level sources, except for some critical cases (long cables, capacitance issues...). The headphone amp can be renounced if the source's line-out amp has a low enough output impedance and is capable to supply the relatively high currents needed for headphones. An additional amplification stage would inevitably make the sound less accurate. This is the theoretical basis. Nevertheless your ears may tell you that the additional amplification stage makes the sound more beautiful and musical anyway. After all that's what counts and the reason why I mostly use an amp, too. BTW: The usual 2 volt of a CDP is more than enough to drive both power amps and headphones.

Mackie... ...you may just as well put the blame on my foreign mother tongue. English sometimes isn't easy to understand.

Hi Well, I don't see headphone amps as a gimmick. Many of them can increase the listening pleasure and the lifelikeness of the music reproduction compared to headphone jacks in CD players and integrated amps. E.g., my NAD 3220, which I use in my computer system, has a really mediocre sounding headphone out. But I've read of some classic integrated amps with very good ones. However: if any amplification - be it superfluous or not - is done in the signal path, if you want good sound, good components and a sophisticated design have to be used. That's definitely not the case in normal consumer devices, where the integrated headphone amps have more or less an alibi function; they're built as cheap as possible. The integration of headphone amps is not comparable to active speakers. While you may complain about a less than desirable quality of the built-in amps there, after all the integration provides a clear advantage in the form of active filters as well as individual and direct drive of the transducers. Of course, to have an amp on the desk which you can touch and look at, which reflects its value, the money you've paid for it, can make you feel good - so that's a perspective worth considering (and another possible reason why «handicrafted» passive attenuators for direct-path purposes don't find many followers at Head-Fi). But believe me, if you have a great source and a great headphone, a good sounding dedicated headamp makes the difference. If not a passive attenuator.

And BTW: this is a really beautifully designed forum!

That's a fascinating thread. Thanks, Adrian, for inviting me to this forum! Sorry for my lengthy debut posting - but what can I do but lay all my cards on the table... I would agree that amps are overrated, but not so much in this forum, as it seems, in clear contrast to Head-Fi, where I use to post. It's hard to find the understanding for the direct-path idea there, where it obviously is perceived as a red rag. The reason could be that the headphone amp often is the most expensive part of the individual setup, so no one likes to read that it's actually superfluous or a pure effect device. I have to confess that occasionally I was tempted to exaggerate it that way, but actually I just wanted to demystify it to a certain degree, make its real role clearer. Which is primarily to amplify a signal. Is this really necessary in the case of a high-level source device such as an FM tuner or CD player, precisely: is a preamp indispensable, to drive a power amp? Definitely no! At least in most cases with not too extreme output and input impedances (...sensitivities, capacitances, cable lengths...). A reasonably designed passive attenuator based on resistors, potentiometers (both acting as voltage dividers) or transformers do the job equally well, with even clearly less alteration of the source signal. In fact in ideal cases the signal is virtually unaltered. No amp can reach such a high fidelity. Meaning: signal accuracy in the concerned path (!), not in terms of the subjectively perceived realism. My direct-path career actually has begun with momentarily renouncing the preamp and using the only source with variable volume I had at my disposal then - an old Revox Tuner - to feed the power amp, with the goal to evaluate the fidelity of different preamps. Not a really great source, but the only thing that mattered was the deviation caused by the additional electronics in the signal path. I finally found the most neutral preamp in the form of the Conrad Johnson PV2. And also that the differences among the various preamps were quite big, as still was the one between the PV2 and the source sound. It was obvious that the amps smoothed and colored it in a euphonic way. The original sound was kind of dry in comparison. I can very well reproduce Mackie's experiences with the «direct path» in this context: improved transparency and clarity at the expense of smoothness and «musicality». Later, with digital sources, I managed to compensate for the thus exposed «digital» artifacts by fine-tuning my self-constructed speakers' crossover networks, adapting them precisely to this specific sound. Interestingly the phono playback barely suffered from that measure, as it would have been in the reversed case. So the conclusion that the digital technics are to blame is not too unlikely in my opinion. On the other hand I remember the dry sound from the Revox tuner which then still was completely analog (as to the broadcasted music). So it's the weaknesses of the sources, generally speaking, which have to be masked by the smoothing amp electronics to make the sound «musical». Strange, but obviously true. However, in the case of my speaker rig I have sticked to the direct-path philosophy, taking benefit from the maximum transparency. This excursion into preamps hasn't happened accidentally: Both direct-path versions have very similar effects. Renouncing the headphone amp and just plugging headphones into the line-out RCA jacks or using a passive attenuator, resp., offers the same advantages, such as increased clarity, transparency and transient speed, at the expense of a certain coherence and musicality. This may lead to the suspicion that those line-out amps aren't really capable of adequately driving headphones, which isn't their design purpose anyway. But I have seriously tested this. With my gear the low and relatively complex load represented by a dynamic headphone has turned out to have no influence on the signal - which is fed into a headphone amp and monitored through this one - when it's connected in parallel to its input (so that the line-out amp drives both headamp and load headphone at once). But one thing I've learnt since my militant Head-Fi episode is that the majority of CD player's and even DAC's line-out stages are not really appropriate for driving headphones directly - mainly because of their high output impedance. My own equipment with its 5 ohm (Theta Pro basic II) and 20 ohm (Bel Canto DAC2) seems to be the exception in this context, whereas my DVD 963SA with its 200 ohm rather represents the normal case. The main issue with the high output impedance is the missing volume, the accompanying influence on the bass response is less striking. The common maximum output voltage of 2 volt (or more) is enough to drive even inefficient headphones, provided the output impedance isn't too high. High-impedance headphones are preferable in view of the output impedance factor as well as the actual purpose of line-out amps: to drive high-impedance loads. The HD 600 is one of the most suitable headphones around with its 300 ohm and its quite high efficiency. There hasn't been reported any damaged line-out stage because of such «misuse» and «overload»... but you never know. The ideal value of a potentiometer for direct-path experiments appears to be 500 ohm. Probably not easy to find (so you may use 1000-ohm types instead for experiment purposes). Combined with a 250-300 ohm headphone, the resulting effective impedance (as to the electrical frequency response depending on the headphone's impedance curve) is around 100 ohm on the average, thus in the range of OTL tube amps. You can also use switched attenuators instead and dimension the resistors accordingly. And there's a cute little passive «amp» available: the ASL Reference Tools, which attenuates the signal in six steps using a transformer. Unfortunately its efficiency is very low, so even with maximum volume setting you need a HD 600 and loud recordings to get passably sufficient listening levels. (According to John Chapman from Bent Audio it should be no problem to build a much more efficient unit.) But one thing it can serve for is to prove the nature of the source signal and to unmask the real role of amplifiers without leaving the door open for any refutation such as to suppose that line outs generally aren't able to drive headphones because they're not designed to drive low-impedance loads. The load impedance created by the transformer is high enough to be uncritical and passably comparable to amplifier inputs, whereas the output impedance in turn is low enough to be uncritical in terms of possible colorations in cooperation with the connected headphone. As to its sound: it's somewhat smoother, slightly rounded, with slightly reduced transparency and dynamics compared to resistor-based attenuators (leaning towards tube amps), but subjectively doesn't suffer from any loss of focus and detail, just presenting the music in a slightly smoothed manner, with somewhat greater coherence. Indeed it's a tiny step towards amplifier sound, but a really tiny one. The basic signature is the same dry and uncolored, highly resolving and detailed sound as from a typical resistor-based passive attenuator. Every amp I've compared to it seems like adding a lot of colors and warmth and smoothing the music signal in an absolutely convenient way, but sacrificing a lot of transient sharpness. Imagine the details in a recording to be colored points, randomly distributed on a white surface. That's how the direct-path sound looks like. With an amp switched into the signal path, the points become larger and fuzzier, reducing the gaps and diminishing the white between them. That's how smoothness, coherence and colorfulness are increased and the reason why sometimes even details can be perceived as more striking (since they are «bigger» now). In comparison the source sound can be perceived as downright «empty». But no doubt: it's more transparent, more focussed and more dynamic. No lack of bass impact or extension at all. As to dynamics: It's well known that compressed music can sound more dynamic than uncompressed. The beat of a bass drum, e.g., is a more important event in a compressed recording because of its prolonged duration. The real dynamic events are very short and not really flashy to the ears. That's why even in terms of perceived «dynamics» amps seem to have no shortcomings, even rather the opposite. To be honest: for most of my music listening time I use one of my two headphone amps: Earmax Pro and Corda HA-2. I fully agree with tee and its excellent posting: headphones provide an artificial sound anyway, so it doesn't make much sense to categorically maintain a purist standpoint. What counts is credibility, musicality, to be emotionally touched by the music reproduction. So I push all theoretical considerations aside and just enjoy the music. And even have to agree that through the amps it somehow sounds more realistic than through the passive attenuator. That doesn't mean the direct path sounds bad at all - but when I have the choice... Nevertheless, I'd really like to have John Chapman build me a perfect transformer-based attenuator once he has the time. And I think I would like it a lot. But it would cost even more than a really good (sounding) headphone amp. So, finally, what's the merit of the direct path? It's the cheapest way to achieve excellent headphone reproduction with even the purest signal possible. The only precondition is a minimal technical skill to build the attenuator, either a potentiometer or a resistor-based switching type (which is more expensive). The main purpose in my perspective is a «scientific» one: to realize the role of amplifiers, their distinct sonic colorations. If one thing still is questioned: the general capability of line outs to drive headphones, one other thing can't be ignored: the considerable differences in terms of sonic signatures among amps. So the conclusion is justified that all ot them color the source sound more or less. Just as well to be understood as «more or less euphonically». (Tube amps are more on the euphonic side, but according to my experience they don't alter the original sound more than their solid-state counterparts.) One concrete benefit from this awareness is being able to question the sense of really expensive amps, with way oversized power supplies. (A guy on Head-Fi drives his R10 with a $15,000 amp.) You can be quite sure to have your money invested in a gigantic effect device. I'm not sure if the extraordinary low load can damage some line outs. There are mixed opinions around, but (fortunately) no concrete damage event has been reported so far. There are millions of computer users which have their headphones plugged into their soundcards' line outs, myself included, and so far I haven't heard any bad news on this subject.