solderdude
Administrator
measureutternutter
Posts: 4,881
|
Post by solderdude on Jun 25, 2014 16:44:11 GMT
The Philips SHP-9500 is the successor of the SHP-9000. It has similar weight (very light weight) and similar type of cloth pads. The headphone is very comfortable and nice looking. It is more neutral sounding than the SHP-9000 and has better highs extension but lacks bass extension. Bass itself sounds good and tight but bass heads should look elsewhere. Bass wise it is similar to AKG K500 / ATH AD700. The treble is somewhat peaky and elevated and thus it sounds overly sparkly. The dip around 3kHz removes some ‘clarity’ but the elevated treble ‘compensates’ more or less. Below the CSD of the SHP-9500 which isn’t as ‘clean’ as some other headphones in this price range (€ 100.- = $ 150.-) but resonances in the 3-5kHz range aren’t as bad as in higher and lower parts of the frequency range. Left = blue, Right = red. Fortunately this headphone reacts very well to EQ and when the missing bass is ‘added’ and the treble peak is lowered it sounds very good ! When 2mm thick felt is mounted in front of the drivers the peaky treble is lowered to more normal levels and this headphone sounds really well with this simple mod. The effect is shown below in the overlay of the CSD. Original SHP-9500, SHP-9500 with felt in front of the driver. Resonances are shorter and the amplitude of the treble is lowered to ‘normal’ levels. Aside from the bass extension (bass itself sounds fine and tight) this headphone sounds surprisingly good. Blue is stock SHP-9500, green = 2mm felt in front of the drivers. Below the SHP-9500 with felt (green trace) compared to the much more expensive Philips X1 (red trace). The SHP-9500 with felt is a little brighter (which I thought was lacking in the X1) but doesn’t have the bass extension. Fortunately with EQ this bass issue is easily solved and in my opinion the SHP-9500 (with EQ/felt) sounds better than the X1. All in all a very light weight and comfortable headphone that sounds very good (certainly with the felt and some EQ) and can recommend it.
|
|
solderdude
Administrator
measureutternutter
Posts: 4,881
|
Post by solderdude on Jun 25, 2014 18:59:51 GMT
Stock SHP-9500 measured via Kameleon (no felt in front of the drivers): I may 'up' the bass just a little more and lower the treble just a tad more. The 'hole' at 3kHz is filled in just a few dB and sounds 'right' now but still measures as a dip. The slight 'peak' around 6kHz seems to kind-of compensate the dip around 3kHz making it sound VERY balanced and flat. Extremely analytic and truthful. The cable is low in microphonics (better than X1 in this regard as well). My only gripe is the pads... they cannot be replaced as the cloth is glued onto the plastic and will tear when you remove it. The pad material is cool feeling cloth (AKG style) and probably aren't available as spare parts anyway.
|
|
|
Post by Brett on Jul 21, 2016 3:48:43 GMT
All in all a very light weight and comfortable headphone that sounds very good (certainly with the felt and some EQ) and can recommend it. Could you please post your EQ settings for this set of headphones mate?
|
|
Rabbit
Administrator
Posts: 7,087
|
Post by Rabbit on Jul 21, 2016 7:44:44 GMT
I would imagine that you'd need to look at the level of the peaks and reduce them by around the amount shown as boosted......
from 80 down needs boosting. around 2k needs a boost. 4-15k needs lowering slightly.
Then you might have as flat as you might need and then tweak for your own preference perhaps. For me, I often like a bass boost of around 3db.
|
|
|
Post by slowpickr on Jul 21, 2016 12:18:55 GMT
I have a set of these and they can be modded and they sound great afterward. Here is the Magick Man mod from Head-Fi: Originally Posted by Magick Man View Post
Okay, I've finished my tinkering, and even though I feel that I could extract more from them in one way or another, other areas start to suffer when I do. So, without more ado, the budget "flagship killer" is...
the Philips SHP9500.
They're that good? Yeah, they're that good. After listening for quite a while trying to figure out what these sound like, I figured it out, they're very similar to the Koss ESP/950s, only they have ~3dB more mid bass. The mids and highs are practically ripped directly from the 950s and are just beautifully smooth and silky without any harshness, and the soundstage, though not the largest I've heard, is substantial and all instruments are well placed. Resolution is fantastic too, THD numbers must be near rock bottom because I detect almost no coloration or grain and what warmth they do give off is natural and unobtrusive. Overall they're really great, and when you factor in their price (~$85US on Amazon) they're miraculous. What makes them so special is their drivers, they actually have 50mm neodymium ring drivers (no joke). I don't know of any other inexpensive full-size cans packing such hardware. (Also the cable is swappable, it's a standard 3.5mm M-M cable, for those who want to use something different from Audioquest or whatever. rolleyes.gif )
But, they aren't quite that sonically gifted right out of the box, they need a little work to pull out more of their potential. Without mods they're a little harsh ~6kHz, mids are a bit honky, and bass, though tight, is simply too recessed. Here's where the simple mods come in.
Other downsides? -the fabric used for the earpads and headband feels a little like burlap, it's not the softest out there -the plastic is pretty cheap feeling -the size adjustment mechanism is kind of crappy -head clamp is very, very light (might be a plus for many), I prefer a little more
What you'll need:
A sheet of 2mm thick craft felt, it's pliable but fairly dense A sheet of 4mm thickness EPDM butyl, it feels pretty rubbery Glue stick Scissors X-acto knife Ruler
1. Feel around the inside of the ear pads, underneath them, and you'll notice that the padding is glued down a bit. What you need to do is firmly but gently pull that glue apart. Don't worry, the fabric will hold the pad to the baffle, we just need to get some space underneath the earpad itself. 2. Using the pad itself as a kind of template, you'll want to draw an oval in the craft foam. It'll be 10cm wide by 12cm tall. Then inside that you'll draw another oval that's 6cm wide by 8cm tall. That will give you an oval ring with a 2cm "wall". Now cut out that ring; I used scissors on the outside and an X-acto for the inside, you want to keep it in one piece. It'll look like this:
3. Cut a circle of felt 8cm wide then apply a light layer of glue from the glue stick to one side, then apply that to the exact middle of the ear side of the baffle (apply it to the fabric back there), tucking the edges underneath the earpad. Let the glue dry, takes ~an hour. Here's a pic of the felt in place:
4. After the glue has dried, take your rubber oval that you cut and slowly work it underneath the earpad, it'll act like it wants to bunch a little but just keep working it under the pad until it's completely under the pad. You can see the edge of it barely visible under the pad here:
5. That's it! Seriously. I stripped one set all the way down to the frame and monkeyed with them like crazy (completely ruining the earpads in the process), but this is by far the most balanced mod for these. In this instance, less really is more.
Get them while you can, though, Philips didn't distribute many in the USA and I think they may even be discontinued now. Sad.
The pictures didn't copy so here is a link for the above post which has pictures: www.head-fi.org/t/600088/an-audiophile-and-petrolheads-journal-buckle-up/7770#post_11121790Edit: I would like to note that I didn't use EPDM mentioned above to mod my set. Mainly because I couldn't find any in town. I simply rolled a paper towel into thin strips (apprx. 1/4" in diameter) and tucked them around the perimeter of the ear pads on the inside. This resulted in added space between the drivers and ears. Enhanced the sound slightly and also gave my ears more room inside the cups so they weren't touching the felt.
|
|
solderdude
Administrator
measureutternutter
Posts: 4,881
|
Post by solderdude on Jul 23, 2016 8:51:45 GMT
All in all a very light weight and comfortable headphone that sounds very good (certainly with the felt and some EQ) and can recommend it. Could you please post your EQ settings for this set of headphones mate? I don't have any settings as I don't use digital EQ. The EQ used was created by a Kameleon filter module which boosted the 3kHz by about +7dB narrow band and subbass (20Hz +10dB) 50Hz +3dB 100Hz 0dB. Treble was reduced as well. Didn't like the fact that the pads are glued on and appeared to be non replaceable. Sound was good though.
|
|
|
Post by dodi on Aug 3, 2016 13:34:44 GMT
solderdudeWhy all headphones measurements are on 90dB? Can u draw frequency response curve that u are aiming for? Looks like most headphones have some relation to Fletcher-Munson equal loudness curves. alturl.com/7o5ii. Some of your mods end up like that: alturl.com/w3xtr. Do they sound more natural than shp9500? How shp9500 will measure on 40dB because all of these headphones are measured on 90dB? tnx
|
|
solderdude
Administrator
measureutternutter
Posts: 4,881
|
Post by solderdude on Aug 3, 2016 16:54:12 GMT
They are measured at 90dB because these levels represent peak levels at which most people listen to when playing at 'realistic' levels during the day(means 75dB to 80dB average SPL). Also my measurement rig (based on WM61A) is not very linear above 100dB and at 90dB there are sometimes peaks well over 100dB. At 100dB (@ 1kHz) there can be peaks well above 110dB which will skew distortion measurements. Aside from that it is a level most other measurers measure as well so kind of a standard which most likely is sprung from: 'audiophile listening levels', linearity of most headphones, limits of some amplifiers/sources, trade off between distortion and background noise, surrounding noise.
At 40dB average SPL you would not hear much music any more and noise levels of the mic and pre-amp would dominate the plots. Sounds of 20dB SPL are practically inaudible in a normal room so having 20dB dynamic range would be way too small.
So 90dB SPL (give or take 1 or 2 dB in absolute sense) is a trade off against distortion and the always present background noise.
The plot you linked to is of a stock T50RP by the way and is not a target plot for mods nor one where I tend to EQ towards. A perfectly 'horizontal' line in my plots is where I aim for. More on this further down this reply.
Fletcher-Munson curves merely show that at softer listening levels the preceived amount of bass is smaller than at higher listening levels. They have nothing to do with any compensations, room curves nor target curves.
You do raise an interesting point about EQ though. Should one always listen to average levels of say 60dB, which already is really really soft and everyone would turn the volume up at these levels except when listening to highly compressed rock perhaps. Then, to perceive the bass notes (60Hz) equally loud you need 86dB level at 60Hz. But when listening to 80dB average (which is a normal level which you can listen a whole evening to) you need 99dB at 60 Hz to perceive bass equally loud. At these levels peaks are there reaching 90-95dB (hence why we measure there). Bass levels (60Hz) thus are +19dB opposite 1kHz at 80 dB average. At 60dB average bass levels of +26dB are needed to perceive the exact same tonal balance.
So at -20dB level difference (at 1kHz) you need +7dB more bass to have the same 'feel of fullness to the sound'.
This is why super bassy headphones (those which measure +10dB or so opposite 1kHz) can sound 'full and rich in tonal balance' when playing quite softly yet become overblown when turned up a lot louder. It is also why 'flat' headphones, which sound 'full and rich in tonal balance' at higher levels (you know the volume where you turn the volume down after 1 song or so) sound 'lifeless, boring and midrangy' at softer listening levels.
This is what Fletcher-Munson plots are showing by the way.
This means that if you ONLY listen to headphones at VERY low listening levels and like a 'realistic, rich sound' then a bassy headphone will sound better than a flat one IF the goal is to reproduce the sound of instruments and voices faithfully. A headphone that measures +10dB (around 60Hz) opposite 60Hz will sound 'realistic and full' When you usually listen at 'just below average' levels and want the same tonal experience a headphone a headphone with a mild boost of say +6dB will sound realistic. When you like to listen to music at 'just above average levels' a boost of +4dB opposite 1kHz would be 'realistic'.
The thing with the Fletcher-Munson curve is that this 'curve' is already embedded in the recordings. When music is mastered in the studio or recorded live they listen at 'well above' average levels and the bass player has an objectively higher level than say a 400Hz flute. This is why in analysed recordings we see more presence in low frequencies than in the mids. Because our ears work according to the Fletcher-Munson curves we 'perceive' music as real when lower frequencies are higher in an absolute level than mids.
Where does this leave my frequency plots.
Well if I had no compensation in my measurement rig and I would measure a 'flat' line. Meaning bass levels, mids and treble would be equal in amplitude then a recording played back through a 'flat' headphone would sound utterly real at 'live' listening levels (way above normal listening levels). So for this reason I compensate for low frequencies about the same way as Golden-Ears and Innerfidelity does.
A 'flat' headphone thus would measure flat from 300Hz to 200khz and would drop off below 300Hz down to about -5dB at 20Hz.
When a headphone measures 'flat' on my plots (so a horizontal line between 20Hz and 20kHz) then in actual SPL the headphone would show a +5dB rise for the lowest frequencies. This 'compensates' live levels to 'normal = can listen to this levele for hours' levels while sounding every bit as realistic as 'live levels' would on a 'flat' headphone.
A perfectly horizontal line (at 90dB) thus represents 'realistic = audibly neutral = 'flat'' sound. A raised bass or treble above the 90dB line thus means it sounds bassy and trebly. A roll-off in the lows and treble thus sounds that way.
My EQ (applied in Kameleon) thus makes the headphone sound realsitic at normal levels. Bassheads and those listening at lower levels may prefer a slightly raised bass on my plots.
The SHP9500 is not 'neutral' to me. It is lacking (sub)bass 'foundation/body'. Bass is there and tight and realistic, maybe a tad 'thin' on the bottom side. This is something only only misses in this headphone when one KNOWS that it should have been there. The mids are very realistic indeed and is one of the strong points and why most consider it 'flat' sounding. The dip around 3kHz takes away the 'edge' of lesser recordings and is what a LOT of people prefer over a 'realsitic presence' in voices and instruments. The elevated treble also creates an extra illusion of 'very detailed' sound which, again, is what a lot of people like unless they have really crappy recordings.
Flat... it is not. With EQ you can get it closer to 'my ideal' where it has a better 'foundation' in the lows, a better presence and less 'slightly over the top' treble/detail.
|
|
solderdude
Administrator
measureutternutter
Posts: 4,881
|
Post by solderdude on Aug 6, 2016 19:48:08 GMT
|
|
|
Post by dodi on Aug 12, 2016 9:09:35 GMT
Thank you very much for your detailed explanation.
|
|