Quote:
"Nice" is in inverted commas for a reason.
And yes, I know you were saying the Woo and the M^3 with OPA627s sounded the same. I was speculating based on potential crosstalk problems as to why the GS-1 sounded different. As nick_charles says, it still seems unlikely that it would be audible, but as I say, wild speculation abounds in the absence of measurements.
I didn't notice the inverted commas, but I knew exactly what you meant when you used the word "nice".
Never the less, I'm not sure you should classify a device with a sound signature as "broken". Many times, but not all the time, a sound signature can be an improvement and sound very good.
Take the 650s for example. They have a bassy laid back sound signature, yet they are able to reproduce the sound of my piano more accurately than any of my other headphones, including my new T-1s.
I don't disagree with you and xnor and bigshot in the science of it, but if you've gone to meets with your favorite headphones and CDs, or a music containing thumb drive, you've noticed that as you plug into the various rigs, and play your music, they don't all sound the same. At one meet one of the guys brought a portable DVD player with a rock concert DVD... we went around attaching the analog outputs (with permission of course) to many of the amps at the meet and listened with our own headphones. I was surprised at the differences I heard and we were always able to agree after listening, that we had heard a different sound signature.
So what I'm saying is that although the science end of it sounds correct, the reality is different.
I also have a problem with restricting your (not you personally) sample to things that measure similarly and sound similarly, and saying that everything that doesn't is either "broken" or a "bad design".
Also, we have avoided answering this question: Does the new 32 bit Sabre chip have more resolution than older chips, and if it does, is it audible?
Quote:
This statement only makes sense in situations where there's no established set of measurements.
But we have several measurements (frequency and phase response, THD, IMD, SNR, and so on) that can be used to assess the performance of audio gear like amps and DACs and even components like operational amplifiers.
Another way to reply to this is: I've measured the right thing but you fail to understand the results and therefore misinterpret them.
What I mean with this is that you have to know what levels of distortion, crosstalk, frequency response deviation from flat etc. are audible. If you don't, how do you define "measures bad"? Is -70 dB crosstalk bad, what about -40 dB? Just because the result can look bad on a graph doesn't mean the measurement is bad or that you can hear differences.
Regarding your M^3, what gain does it have? Ti aka amb advises against the OPA637 depending on the configuration.
And how did you proceed in matching the volume between amps?
gain of 11 on the 637 M^3.
See above. I don't disagree with you guys about the science, but I've found that the reality doesn't necessarily echo the science.
Quote:
If equalising sounds like crap, you're either doing it wrong or have a godawful EQ. Try a decent 24-bit one - Electri-Q is the only name I can remember offhand. With something like the ipod's equaliser, there are tonnes of problems, which shouldn't exist with decent software EQs. If you immediately notice a degradation in sound with no attenuation or boosting at all, something has gone horribly, horribly wrong somewhere.
EDIT: Xnor was rather more to the point I see.
EDIT2: Regarding the M^3 and potential problems with opamps, I think we can safely assume the 637 was not used in a low-gain amp, as I think upstateguy's M^3s are professionally built (rockhopper audio).
+1 on
Electri-Q
and regarding the M^3, you can't use a 637 in the ground channel at any gain.... you have to use the 627.
Quote:
Several mastering grade, super expensive, parametric EQs can be nulled with freeware EQs. What this means is that you can, in fact, achieve the same thing. This is not suprising if you know some digital signal processing basics. I've developed graphic as well as parametric EQs and there is no magic going on behind the scenes. It's all based on the same maths.
To not turn this into "all EQs sound the same": Yes, there are different transforms but the simplest one is used in most parametric EQs, some even use iterative methods. Yes, there are different types of filters like IIR and FIR. Yes, some EQs have a feature to oversample in order to make IIRs behave like their analogue counterparts near nyquist. And so on..
But all of this boils down to transfer functions for each filter which mathematically define how the input signal is related to the output signal. Y = H . X, where Y is the output, H the transfer function and X the input.
You can analyze any parametric or graphic EQ quite easily. You can record what the EQ is doing and look at the frequency and phase response of each filter. All of this is completely transparent if you know DSP. If you want to know more attend a course or read a book on the topic.
@bigshot: Yeah, but room correction is just one of many areas of application. You can color the sound to whatever you like with an EQ. In the wrong hands you can cause clipping or veil and ruin sound quality.
EQs are powerful tools. If you know what you're doing I'd argue that you can improve almost every playback system, no matter if you have speakers or headphones.
The deeper we delve into this, the more things don't sound the same.....
The science says they should, but going to meets and listening to rigs says they don't. Sometimes its sound stage, other times is frequency response differences. like the difference between 880s and 650s and still other times it's resolution, where one Dac, Amp or entire Rig is "cleaner" and "clearer" than another.
One would think that if all gear sounded "relatively" the same, all the rig combinations would also sound "relatively" the same.