[fotios]

Hi

I know that there are another threads for this issue. In this (i mean the link at the end unfortunatelly

) except pictures of modification you can see as well a documentation with FFT analysis. The software used is the Virtins M.I. Pro 3.1 and the interface are two Essence STX cards themselves. The one is the modified, and the second is the original as is shipped from the factory. Few words about MI Pro 3.1 (just yesterday released the new v3.2 from Virtins but i had not time to re-calibrate it, to uninstall-reinstall cards etc. so i used the old v3.1) it is one of the most trusty softwares for FFT analysis. It uses the ASIO buffer of sound card, so it is independent from the patch or the windows mix. If it is calibrated correctly in conjunction with the sound card used, then everything can be adjusted from its very sophisticated signal generator. The results are not "cooked" by no way, if we do the right setup before the execution of a measurement. The secret is found on the page 116 of its manual were is a table refereed at the correct frequency of signal generated in relation with the FFT size and the sampling rate. There is no reason of using smoothing windows after this. In the resulted plots that i quote, you can see that these are very close to those offered from ASUS and which are obtained with an expensive Audio Precision analyzer. You can download the trial version (full operational for 1 month) as well the manual. For users of Xonar Essence STX, i can offer ready the calibration factors. Open the calibration window and simply write in the "Range (V)" box +/-2.91646 and the same in the "Others/ASIO" box. Then click OK. Open the patch mix of Essence STX, and place the master volume knob at full, the playback sliders at 76 and the record slider of line in at 100. Mute everything else. Open now the signal generator of MI Pro 3.1, and for a typical FFT analysis at 24bit/48K write on the frequency boxes 1007.8125. Place the sliders of generator at -5.0dBFS which corresponds in 1.6Vp, which you may see in the "Output Amplitude V" boxes if you have did the calibration that i said and which is a safe value to not overloaded the ADC converter of Essence STX. The rest calibration of the FFT analyzer you can see in my plots. If you want to select a different SR and bit depth you may see at page 116 of pdf manual the related table. For further assistance ask me please.
Back to the Essence STX now. The upgraded card it is not mine. It is of a friend which has not the possibility to make the upgrades by alone. My Essence STX, remains as is shipped from the factory and under the continuous danger to fried during an external loopback test. Because i use this card mainly for measurements, if i knew that has very delicate Line In then i had bought the PCI EMU1212. I had the EMU-0404 USB, which was very rugged in overloadings but with a lot of pots, switches etc on it. You can't get trustable results of a measurement with a such card.
Because my friend has a PC with desktop case (low height) i placed the Bursons vertically on the Essence STX. See pictures and plots at: welcome to* The Electronic Audio Labs* web magazine
I ask your comprehension for the link, but the size of plots is big and if is resized, then the numbers will be no visible. Unfortunatelly, i can upload images up to 19.5KB only.

Instead, in diyAudio Forums i can upload files up to 195.3KB.
I leave in you the conclusions.

Fotios

 

[leeperry]

you mean http://www.eal.gr/P.C.%20AUDIO.htm

I don't see a crosstalk graph...and don't you still own your stock STX for comparisons?

 

[fotios]

Quote:

Originally Posted by leeperry /img/forum/go_quote.gif you mean ÍÝá óåëßäá 0 I don't see a crosstalk graph...and don't you still own your stock STX for comparisons?

Hi leep

Yes, this is the correct link. Do you know how many times i tried to insert this link without success? I don't know the reason. So, i was forced to insert the link of my web page which is translated from the system in welcome to* The Electronic Audio Labs* web magazine instead "http://www.eal.gr". I will try now to insert the direct link again: ÍÝá óåëßäá 0. Can you see in what is translated?

Fotios

 

[fotios]

Quote:

Originally Posted by leeperry /img/forum/go_quote.gif I don't see a crosstalk graph...and don't you still own your stock STX for comparisons?

Leep, sincerely i had plots of crosstalk for comparison. Simply, it was my decision to not include them because i don't like to worry people. Anyway...
the FFT analysis shows that the original Essence STX has a crosstalk of 100dB, instead the upgraded has... [size=xx-small] 44dB[/size].
As for my STX: I fried the analog input and the ADC of the first in which i had replaced (you remember) the DIL sockets. Fortunately, the seller refunded me with a new. In the second card i did not any change. It is as shipped from the factory.

Fotios

 

[fotios]

Now, i am looking in this: http://www.minidsp.com/onlinestore/details/46/7/minidsp-kits/minidsp-kit-revb
I will contact with the seller to ask him for some details. From a first view, this card seems to be fine for my work, it is unexpensive, upgradable, and above all can accepts up to 24Vdc in its input without risk. It is not a card for games. Tell me your option.

Fotios

 

[fotios]

Quote:

Originally Posted by fotios /img/forum/go_quote.gif Now, i am looking in this: http://www.minidsp.com/onlinestore/details/46/7/minidsp-kits/minidsp-kit-revb I will contact with the seller to ask him for some details. From a first view, this card seems to be fine for my work, it is unexpensive, upgradable, and above all can accepts up to 24Vdc in its input without risk. It is not a card for games. Tell me your option. Fotios

This again

Can you see that the link does not works?
Please go in this site: MiniDSP

 

[ROBSCIX]

Quote:

Originally Posted by fotios /img/forum/go_quote.gif Now, i am looking in this: http://www.minidsp.com/onlinestore/details/46/7/minidsp-kits/minidsp-kit-revb I will contact with the seller to ask him for some details. From a first view, this card seems to be fine for my work, it is unexpensive, upgradable, and above all can accepts up to 24Vdc in its input without risk. It is not a card for games. Tell me your option. Fotios

That is an interesting item. Might have to check into one of them myself.
You might want to check your link again.

Here is the fixed link: LINK
Anyway, are you considering writing your own source code/firmware for the board or just using the ready made plugins for your application?

 

[leeperry]

Quote:

Originally Posted by fotios /img/forum/go_quote.gif Yes, this is the correct link. Do you know how many times i tried to insert this link without success? I don't know the reason. So, i was forced to insert the link of my web page which is translated from the system

the forum will automatically parse url's, you can either disable it by unchecking "automatically parse links in text" at the bottom of the edit window, or do it yourself by only keeping the http url between <url> and </url>(of course I meant [] instead of <>)
Quote:

Originally Posted by fotios /img/forum/go_quote.gif Leep, sincerely i had plots of crosstalk for comparison. Simply, it was my decision to not include them because i don't like to worry people. Anyway... the FFT analysis shows that the original Essence STX has a crosstalk of 100dB, instead the upgraded has... [size=xx-small] 44dB[/size]. As for my STX: I fried the analog input and the ADC of the first in which i had replaced (you remember) the DIL sockets. Fortunately, the seller refunded me with a new. In the second card i did not any change. It is as shipped from the factory.

ah ok! I didn't realize that the STX on which you switched the sockets was the first one that had its inputs out of order after a while.

another guy here killed his STX line input twice in a row, hopefully the Xonar PM you have sent a corrective design to will have it worked upon some day(don't count on it, this said)

still, you could run the same tests on your stock STX...and compare to the pimped card, or? I did RMAA my ST because it was sounding terrible, it didn't kill the card.

it's strange that the bursons give such poor crosstalk results in RMAA...but RMAA goes through KMixer so it only worsens things up.

anyway, ppl keep whining about specs on these audiophile-wannabe forums....a burson measures BAD, and sounds great huh

Quote:

Originally Posted by fotios /img/forum/go_quote.gif This again Can you see that the link does not works? Please go in this site: MiniDSP

looks nice! now to see about the drivers? because you want automatically bit-matched ASIO/KS/WASAPI to avoid resampling, and this is pretty much impossible on the Xonar cards.

you can get an Audiotrak Firewire external soundcard for less than 100 EUR shipped from thomann.de, maybe a more worthy option? the drivers are great! the Xonar cards are expensive toys, that fail at everything except for the shiny box and the metal shield to keep the noise voodoos out of teh computer

 

[fotios]

Quote:

Originally Posted by ROBSCIX /img/forum/go_quote.gif That is an interesting item. Might have to check into one of them myself. You might want to check your link again. Here is the fixed link: LINK Anyway, are you considering writing your own source code/firmware for the board or just using the ready made plugins for your application?

Yes, it seems very good item. Except the 24bit/192Ks possibility of its ADC/DAC, there is a meaningful detail: IN / OUT dynamic ranges are referred greater than 98dB UNWEIGHTED . This corresponds to 115dB A-Weighted, at least.
As for the rest - in the case of purchase - i think that i will use the ready plugins for the first time. As i can perceive, this company it is not like ASUS or Creative, so i can suppose that will be helpful in any demand.

Fotios

 

[ROBSCIX]

Quote:

Originally Posted by fotios /img/forum/go_quote.gif Yes, it seems very good item. Except the 24bit/192Ks possibility of its ADC/DAC, there is a meaningful detail: IN / OUT dynamic ranges are referred greater than 98dB UNWEIGHTED . This corresponds to 115dB A-Weighted, at least. As for the rest - in the case of purchase - i think that i will use the ready plugins for the first time. As i can perceive, this company it is not like ASUS or Creative, so i can suppose that will be helpful in any demand. Fotios

It seems like an interesting item for doing some experimenting.
Are you going to try your hand at programming it? That is one of the more interesting aspects for me.

 

[oohms]

I think if people are using discrete op-amps in the xonar, they should upgrade the power supply (at least the -12v psu)

The bursons don't look shielded either, so each half can induce noise into the other half.

If i had a xonar, i would get the ST with the already improved clock, upgrade the filtering for +12v and upgrade the -12v psu, and add more decoupling caps (low ESR and tantalum). Then i would try regular op-amps first, before thinking about discrete ones (too expensive, big), and maybe alternate I/V solutions.

 

[leeperry]

except that the ST measures much worse than the STX...look at the stereophile review

and you obviously haven't heard the bursons.

 

[fotios]

Good morning from Greece

@Leeperry: thanks for the instructions. I disabled the corresponding box, and now the links are OK.

@ROBSCIX: I have restricted knowledge in programming. My knowledge is restricted in writing flowcharts, in some basic language and in assembly language for ST microcontrolers. I don't know if this knowledge is enough for writing firmware etc. As i said, i will contact with MiniDSP to ask him for my demands. I need a hardware addressed in FFT analysis to use it with Virtins MI Pro 3.1. Thus any of the toys offered from sound cards it is useless and injurious for this work. The ASIO buffer it is good, i described in the first post the reason. I tried the PCM2902 of TI with success, but it is restricted at 16bit/48KHz.
Practically speaking: look at this plot, which has been obtained with a precision RMS voltmeter:
http://www.eal.gr/CHARISMA%200K6%20F...SE%201Wrms.jpg
My amplifier - i don't claim that it is a clear Hi-End device because it is a BJT implementation at whole - can touch the 150KHz (-3dB point at 1Wrms/4Ω) which is the worst case. The better sound card and the MI Pro 3.1 of Virtins with their restricted response at 192Ks:2=96KHz can't obtain this measurement BY NO WAY. As well, a 16bit depth is translated in: 16 X 6 = 96dB dynamic range. Look at this plot:
http://www.eal.gr/0K6_THD_1Wrms_8Ohms_ASUS.jpg
The noise level of my amplifier it is 93,36dBr. If i change the position and the direction of the monster power transformer of 500W, then i can improve the noise level by 4dB. Then, a 16bit sound card can't measure this level. I need a depth of 24bit for such measurements. For frequency response tests, i use only trusty bench-top instruments. The crappy SAR type ADC converters of sound cards are for measuring only THD+Noise etc. OTOH the flash type ADC converter included in all DSOs, although it has terrible bandwidth in the range of GHz its depth of 8bits X 6 = 48dB is very poor for such type measurements.
Hope the above description was enough to you understand my place.

Fotios

 

[oohms]

Quote:

Originally Posted by leeperry /img/forum/go_quote.gif except that the ST measures much worse than the STX...look at the stereophile review

I read the online version but it only showed measurements of the STX, with a small footnote at the end about the STX measurements being slightly better than the ST.

They tested the ST in a 1.4 ghz pentium 4 pc which probably has a HORRIBLY BAD switch mode psu - i repair these things for a living, versus testing the STX in a shuttle pc with a much more modern active PFC model. This alone would cause me to discount their comparison.

They also incorrectly point out the AV200 as an LSI chip (when it is a re-branded c-media chip) and infer that it has a hardware DSP

 

[fotios]

Quote:

Originally Posted by oohms /img/forum/go_quote.gif I think if people are using discrete op-amps in the xonar, they should upgrade the power supply (at least the -12v psu) The bursons don't look shielded either, so each half can induce noise into the other half. If i had a xonar, i would get the ST with the already improved clock, upgrade the filtering for +12v and upgrade the -12v psu, and add more decoupling caps (low ESR and tantalum). Then i would try regular op-amps first, before thinking about discrete ones (too expensive, big), and maybe alternate I/V solutions.

I don't think but i am sure for this: a discrete implemented opa to show its benefits must be supplied with +/-18Vdc AT LEAST. I don't exclude from this rule neither the ICs. For this reason LM4560 has possibility of +/-22Vdc supply. As for the Burson, i am sure that it can obtain its maximum performance with +/-24Vdc supply at least. Another one thing of big importance it is that, discrete implementations have very large open loop gain compared to ICs and thus are most sensitive in any type of interference like EMI, RF etc.