solderdude
Administrator
measureutternutter
Posts: 4,886
|
Post by solderdude on May 12, 2014 4:53:28 GMT
It would have been easy if they had recorded some white noise.
You can 'guess' it, if they also uploaded the original signal and AB'ed that file.
I have not seen measurements of the 50X yet, I don't know if they differ from the non-x versions.
Sibilance may be possible to reduce as it is easier to reduce passively than to boost.
|
|
|
|
Post by c61746961 on May 12, 2014 14:09:13 GMT
The original file is available: soundcloud.com/sonic-sense-pro-audio/sets/headphones-testIn Tyll's graphs the M50X has better distortion and reduced 10 KHz peak. Both sound somehow 'grayer' than some other phones in the comparison, and the M40X has more midbass presence, but clearer highs. It's very nice to be able to make relative comparisons like this. Edit: I've done further comparisons to other closed cans. I did this via the fb2k ABX tool with ReplayGain track gain adjustment on (which should get the tracks reasonably close in loudness via the EBU R128 profile), which makes the differences pop out very clearly due to the null switching time. Among the lot I think I still prefer the M40X, the upper midrange and highs are of a higher quality compared to most other closed headphones in the < $200 range, the dynamics are good as well, downside is the elevated midbass, which by no means sounds bloated to my ears, just very present. Not a bad thing in a noisy environment, I think. I know it's not closed, but the Superlux 681 EVO fared very decently in this test, too, I can only guess how good Frans' fixed version sounds like  |
|
solderdude
Administrator
measureutternutter
Posts: 4,886
|
Post by solderdude on May 12, 2014 19:13:14 GMT
I examined Tylls plots and think the 50X is even darker sounding than the M50. I like my headphones a bit brighter and would probably prefer the M40X myself. For a closed headphone you could also consider the Roland RH300. It looks a bit like a Sony but sounds way better. It measures flat and sounds really good (tried one myself) but may not be easy to buy.  They also have an open version which also sounds very good... the RH-A30. It is slightly clearer/brighter sounding but otherwise very comparable. Not for bassheads !  I think these 2 particular Roland headphones get too little coverage. The EVO and modified EVO sound very, very, different ! |
|
solderdude
Administrator
measureutternutter
Posts: 4,886
|
Post by solderdude on May 12, 2014 19:39:35 GMT
The M40fs and M40X at soundcloud sound VERY VERY different indeed.
They also have the HD681 and EVO in those comparative files.
The DT1350 is clearly NOT having a seal and sounds like a laptop speaker !
In reality (with good seal) it sounds completely different.
|
|
|
|
Post by c61746961 on May 12, 2014 20:18:04 GMT
The Rolands look very dandy as well, sadly they don't ship here. I think I will go for the M40X  Yeah, the DT1350 sounds wrong, so does the TH02, both feature a finicky fit I presume. I liked the DT250 a lot, too, the bass was more in line with the original files but the mids sounded slightly cupped and the treble a little low-fi, I think they are starting to show their age. |
|
|
|
Post by c61746961 on May 18, 2014 1:08:44 GMT
The guys at Sonic Sense are upgrading their library with recordings from a new setup. I think the DT1350 is properly coupled in this one.
|
|
solderdude
Administrator
measureutternutter
Posts: 4,886
|
Post by solderdude on May 18, 2014 5:52:08 GMT
The guys at Sonic Sense are upgrading their library with recordings from a new setup. I think the DT1350 is properly coupled in this one. I might upload music recorded through my headphones using the same file as used with the Kameleon amp. My rig isn't very suited as it has a very poor s/n ratio which doesn't matter for FR but is impossible to use for THD measurements and recordings. Will give it a try though and have done this before in the early stages of the filter board to 'verify' by ear if the filter is bringing the signal closer to the original. To really appreciate it they should be played back on a REALLY neutral system (which few own) otherwise you hear the coloration of 2 headphones combined ! It could give you a comparative idea though as the 'original' file also get's coloured by the used headphone. So it could say something about the tonal balance. For instance when you would listen to a recording that contains great LF extension or treble extension with a speaker or headphone that doesn't have this it will not sound as good as it sounds in reality, nor will the file sound great on systems that have less resolving reproduction than the 'tested' headphone. I can only upload a few headphones and all modified (and or 'corrected') |
|
|
|
Post by c61746961 on May 18, 2014 15:03:27 GMT
I gave a listen to the T50RP track and it did help my modded pair a fair bit, now I want a Kameleon  Also, the new recordings put the M40X amazingly close to the source file, my pair is on its way and I will surely report back if they really do sound that neutral. |
|
solderdude
Administrator
measureutternutter
Posts: 4,886
|
Post by solderdude on May 18, 2014 17:46:33 GMT
Well you can get an idea by analysing the files. below is the spectrum of the source file which must be considered 'flat' If anything it almost looks like pink noise (spectrum wise) when you click on the picture it enlarges and you can use those for your analysis.  Then for several (the more the better) points in the FR you must plot the differences from the file below which is the FR of the M40X  You must observe the scale as it differs in the dB range. So when 50Hz in the source = -19dB and in the M40X file it is -37 the 40X will be -18dB At 500Hz the source = -33dB (so -33 dB = 0dB) and the M40X is -30dB difference = -3dB At 8kHz the source = -60dB and the M40X = -54dB so +6dB at 8kHz. You can make a detailed plot this way. Needless to say from the plots (and sound) the M40X is very good in the mids, rolls off the lows below 300Hz and has a significant increase in treble. This last part can be passively compensated. The loss in lows cannot. The lack of bass kills it for me, the treble can be reduced easily. |
|
solderdude
Administrator
measureutternutter
Posts: 4,886
|
Post by solderdude on May 18, 2014 18:43:35 GMT
I 'recorded' some of my headphones some modified, some filtered. you can DOWNLOAD the ZIP file which contains a few files. 01 = original 02 = original but with 'room correction' it would sound like this on a very good speaker in a 'normal' room. 03 = original HD650 04 = HD650 via Kameleon 05 = modified T50RP on a 'normal' amplifier 06 = Modified T50RP played via the filter circuit as described in my T50RP article. 07 = Modified T50RP played via Kameleon When you compare the sounds it is quite obvious the 06 file (filtered T50RP) is closest to the original sound (01 file) almost a 'copy' The 04 (HD650 via Kameleon) is quite close to the 'speaker in room' sound and depending on what my mood is I grab one of these. The 'unfiltered' T50RP is substantially different sounding. The T50RP via Kameleon is somewhere between 01 and 02 file but also quite pleasant and better than without filter. 08 = modified K551 via Kameleon with Bass-boost engaged. Will also design an Olive-Welti signature filter. 09 = modified HD681-EVO (half felt mod + inductor + Wang Yifei pads) it does actually sound pretty good and prefer this over the HD681 10 = Modified HD681 which is still V shaped so it retains the bass and treble and actually sounds better than the recording at lower listening levels. 11 = DT1350 with HD25 pads and a filter circuit. To record it properly I had to fit 'adapters' but have only one so didn't use them this makes the recording sound weirder than in real life. 12 = TH-02 in DT330 housing and with filter inside, added it just for fun and do not use this one myself. Download link for the recorded headphones is below (available till may 25th) www.filemail.com/d/lrphuivpcdwenrnthe sound samples used are MP3 320kbs (to keep it smaller) and for comparative purposes. The rig the headphones are placed on has a very poor S/N ratio and is not really suited for this task. It does show the tonal balance. Weird part is that the brain adapts itself easier to the actual headphone sound than the recorded sound. In other words the headphones sound 'better' in real life and less 'coloured' than how they appear on the recording. I am still amazed how accurate the filtered T50RP is and how 'speaker like' the HD650 is. Also the HD681-EVO with mods is the budget champion ! |
|
|
|
Post by c61746961 on May 19, 2014 3:00:10 GMT
Thanks for the recordings! The plots you made look like the old recordings, in the new recordings the source file is 32-bit, 44.1 KHz and the headphone recordings are at 32-bit, 96 KHz. 89.2 MB and 194 MB respectively. I resampled the source file to 96 KHz using SoX' High-Quality preset to match and produce the following plots: Original  M40x  Overlay (original is purple, M40x is green)  Too good to be true? |
|
solderdude
Administrator
measureutternutter
Posts: 4,886
|
Post by solderdude on May 19, 2014 4:54:35 GMT
Yes..
the plots I posted are actually M40X and the original but they seem to be 128kb MP3.
My guess would be the differences shown are from the same files and caused by resampling ?
|
|
|
|
Post by c61746961 on May 19, 2014 5:52:33 GMT
The files I used I got directly from SoundCloud, via the 'Download this track' button that appears when you hover any item. The files were distinctly named upon download, so I don't think I screwed up there. The dt1350 shows a response akin to the one measured by GoldenEars when compared to the source file, so does the M50x and the V-Moda M80 and XS. Edit: Yep, I just double checked, here's the waveform display:  As you can see, the pop sections of the original recording are heavily compressed, and the headphone recording is not. The FR of the second recording is within 4 dB of the original, broadband, which admittedly is a bit hard to believe. |
|
solderdude
Administrator
measureutternutter
Posts: 4,886
|
Post by solderdude on May 19, 2014 6:09:50 GMT
I downloaded the files as well and they are indeed quite different.
The original seems highly loudness war'red while the M40X recording appears quite dynamic.
The M40X seems to have the same peak around 8kHz (+6dB) and peaks around 10Kz and 13kHz.
Bass seems to be flat to 50Hz at least.
All other headphones appear to show their distinct properties but the M40X seems to be very 'truthfull' in general (aside from the erratic HF behaviour)
There is only one solution to find out... buy it and try it.
Maybe from a mail order company with return-right ?
The few reviews I read say it sounds very good and sounds like the M50 without the bloated bass.
|
|
|
|
Post by c61746961 on May 19, 2014 6:25:53 GMT
Yes, I have a pair on my way already, I should be able to tell if it's a dud, so far it doesn't seem like it.
|
|
