[MLudovic]

Hi there, first post to share what I had fun doing these last days :

I'm the owner of a SR-3 and a SR-404 Signature, both powered by a 580V modded SRD-6 box. I'm very happy with the second hand SR-3 (especially for its low price) but I'm not exactly in love with the acid and harsh sounding 404, especially when listening to heavy/metal bands with loud guitars filling the whole medium/treble spectrum (Strapping Young Lad someone?)

I recently read interesting stuff about the ED-1 and ED-5 equalizers, and I decided to try to emulate their behaviour with foobar and its convolution plugin.

I eventually managed to find the schematics (thx to Mr. Gilmore, and to the filetype pdf search!) and first entered them in Switchercad III to observe their behaviour :

Then I used RePhase, a complex but terrific program, and a lot of patience to generate my impulses :

As you can see I decided to let the filters affect the phase like the hardware version does, but the EQs can be set on "linear phase" to avoid that behaviour. Another method would be to feed ScadIII with a 0.02s dirac pulse on the input of the circuit and to save the output as a wavefile, but you lose the flexibility and phase control allowed by rephase.

I was eventually able to test them with the foobar's convolution plugin. Here are the links to the .wav impulses :

http://www.petit-fichier.fr/2013/09/02/stax-ed-5/
http://www.petit-fichier.fr/2013/09/02/stax-ed-1-monitor/
http://www.petit-fichier.fr/2013/09/02/stax-ed-1-signature-1/

The custom curve is the one I'm currently tweaking to get the maximum out of my SR-404. Here you can see the green original ED-1 vs the red custom curve :

Custom SR-404 curve (EQ + Diffuse Field) : http://www.petit-fichier.fr/2013/09/02/stax-sr-404-df/


Well, that's all. I hope you guys will find that useful and you'll give me some feedback and comments. Also, I'm desperately looking for the schematic of the ED-1 Signature model...

Edit : All 3 Stax boxes + SR-404 custom correction.

 

[davidsh]

Wow, nice have to try this! It must be decent with the signature, right?

 

[autoteleology]

Very, very neat. I will have to try this myself. Great job!

 

[MLudovic]

Quote:

Wow, nice have to try this! It must be decent with the signature, right?

 
I guess so. I know that the ED-1 Signature is designed for the Lambda Sig but I've also read that the difference between the ED-1 Monitor and ED-1 Signature is rather subtle. Concerning my SR-404 Sig, it really shines now with the custom curve.
 
Quote:

Very, very neat. I will have to try this myself. Great job!

 
Thx! Looking forward to reading your impressions.

 

[davidsh]

Soo, this is changing both phase and EQ'ing, right? Have you tried searching for similar images to what you already have found on google? Or maybe take some screenshots of the manuals you found?

 

[MLudovic]

Phase is affected like in the real thing (dotted line on the third screenshot), but with RePhase you can also decide to generate an EQ impulse that won't change phase at all (linear phase). It won't sound the same even if the difference is very subtle.
 
I did a lot of searches to gather infos on the Stax boxes, and I found some interesting articles on the diffuse field theory which lead to their design. But the graal were the schematics that grant a faithfull emulation of what the hardware actually does, phase shifting included. Sadly I didn't find any user/service manual scan yet.

 

[arnaud]

Mludovic, very cool! I am quite interested in rephase as we discussed on HCFR...

I don't know if you read the interview we had with stax last year but, when we discussed about the smyth realiser, stax mentioned they would not go back to the ED days because it is mich easier to realise in the digital domain.

You've proved the point .

 

[davidsh]

I am positive they must have made a lot of compromises with the ED boxes, that could potentially be bettered in the digital domain. It doesn't work for me btw. No difference in sound with the foobar plugin convolver with the .wav files chosen.

 

[davidsh]

 

[DefQon]

Interested in this also, I also have the SR-5 and 404 Sig's and would like to see how well you go with this replication of the original ED-1/5 equalisers.

 

[MLudovic]

Davidsh your settings are correct and you should hear a sensible difference between the raw and the processed sound. I've attached a short clip with different settings where the effect of the EQ is cleary audible.
 
http://www.petit-fichier.fr/2013/08/14/stax-test/
 
Arnaud : I've read it, very interesting material ! Now I also have to dig the HRTF and Smyth realiser concepts...

 

[davidsh]

it does work for me now.

 

[davidsh]

I have thought a bit about it and tested with a violin concerto. Don't really like those custom settings. The tonality is fairly nice, but the EQ'ing of the upper overtones of the violin makes it sound boxed in and slightly smeared, but on the other hand it makes for a more weighty sound than the original ED-1.
 
The original ED-1 sounds a bit more thin with 2 much upper mids in my opinion, but it has a nice sense of being at a venue with the instruments reverberating a bit and 'standing tall', and it does add some sense of spatiality. The problem with the music I heard can be that the sound is suddenly like being in a very big hall, but the instruments still retain the same characteristics as when they were recorded. Violin recorded up close sounding like it is very far away is not good and makes for a strange sound... it might work well with big orchestras and such, will maybe try that later.
I just think the instruments are too slim/thin, and the sound seems slightly grating and on the verge of being harsh, which is not nice.
 
I like the original ED-1 the best. I am using the lambda sig's. Overall I think I'd prefer the sound without the ED-1 emulation, as I think it can be a bit fatigueing.
 
Just some thoughts based on a listening session. The music used was 'La Stravaganza' performed by Rachel Podger.

 

[arnaud]

Is the emulation for the lambda signature? If so, I guess you got used to the tonality without the correction then.

There's one sure thing: these filters are really only modyfing the tonality (and phase but that is not part of the original target, just an inevitable result from filtering be it performed in the analog or digital domain). With the intented headphone, they enable to switch from one target to another, e.g. Free vs. Diffuse field compensation (a headphone can't possibly reproduce both by design since these are different targets).

This helps to reproduce the same amplitude response as a speaker in either listening conditions (keeping in mind that the compensation is truly only valid for the particular headphone and dummy head pair used to measure the magnitude response).

But the amplitude response is only part of the story. It does affect the subjective impression of soundstage width to some extent (a reason why perceived width drastically varies between headphone of similar design), but it does not resolve the lack of crossfeed in traditional headphone reproduction (which is mendatory if you really want to reproduce a speaker sound at angle in free field conditons for example).

In that sense, nowadays, something like the Smyth realiser type (convolution something like 750ms long impulde responses using FIR filters) seems more promising, even it is a different level of investment and another bag of worms (clean acquisition of HRTFs, find a satisfactory environment to get the calibrations done...).

That does not diminish the excellent work that Mludovic shared with us, actually it is a good basis to start from and explore the strange world of hrtf compensation .

 

[davidsh]

It is not an emulation of the ED-1 signature, it is an emulation of the ED-1 which is made for the lambda pro.