[Artemio]

I am new around here and am very inclined to measure the gear I find with my limited means. About a year ago I went up to the Home Theater Shack, and learned about their Room Equalization Wizard Software and the use of the BFD. I won't go into much detail here, but you can follow the links and read for yourself if you are so inclined (great information in there for those using speakers and subwoofers).

The thing you need to know is that the REW software is used to measure frequency response and with the BFD you can use the parametric equalizer and a decent microphone (such as the ECM8000) to compensate for the bass response of your room's acoustics and obtain great results. What interests us here is the FR graph. The process is simple, you calibrate your soundcard (the line in and line out jacks) with the help of a single cable that connects them, and trace the response. The software creates the filter to compensate for a flat response from your card. After this you can check out how other equipment affects the FQ response (for example, in case of measuring the subwoofer the response from the mixer used to power the mic).

I though... what about measuring the Corda Move to check out the FR? And so I did when it just arrived. I previously owned a Porta Corda III and did measure it's response, but didn't think of that until after I had done some burn in. But for the Move I did graph the FR before and after. Unfortunately, I made a mistake back then and erased the graph before burn in without the crossfeed filter. The Move burn in process was done with pink noise and it has over 100 hours with several more of music.

I present to you all the graphs I thought would be relevant as information so you can give your thoughts on the subject.

Out of the Box Corda Move with Crossfeed on: (linear scale for Hz)

Out of the Box Corda Move with Crossfeed on: (logarithmic scale for Hz)

Corda Move after burn in with Crossfeed on: (linear scale for Hz)

Corda Move after burn in with Crossfeed on: (log scale for Hz)

Corda Move after burn in with Crossfeed off: (linear scale for Hz)

Corda Move after burn in with Crossfeed off: (log scale for Hz)

[size=small]And the most revealing ones:[/size]

The three graphs: (linear scale for Hz)

[size=x-small]Red: New CF On
Green: Burn in CF on
Purple: Burn in CF off[/size]

The three graphs: (log scale for Hz)

[size=x-small]Red: New CF On
Green: Burn in CF on
Purple: Burn in CF off[/size]

I placed a marker on the "The three graphs (linear scale for Hz)" graph for better reference on the impact of these.

Here are the same last two, but overlapped:

The three graphs: (linear scale for Hz)

[size=x-small]Red: New CF On
Green: Burn in CF on
Purple: Burn in CF off[/size]

The three graphs: (log scale for Hz)

[size=x-small]Red: New CF On
Green: Burn in CF on
Purple: Burn in CF off[/size]


Personally, I am happy to verify with measurements that there is indeed a change in FR as a result from burn in. Of course this is just a measurement made with a sine wave.

What are your thoughts on these?

 

[RedLeader]

Cool!

 

[mdarnton]

I don't much understand it, but I'm going to try.

 

[klemchang]

Quote:

Originally Posted by Artemio /img/forum/go_quote.gif What are your thoughts on these?

Fantastic data! Thank you!

 

[jamvanman]

Very interesting to see the change from burn-in. Thanks. Might be interesting to see how they look after another couple of 100 hours - presumably it would show the Move doesn't need to be burnt in for much more than 100 hours?!

 

[Filburt]

How many samples were taken at each point in time of each configuration?

 

[biged]

Awesome! I would love to see something similar with cable and amps.

Thanks for a great thread.

Ed

 

[Artemio]

Quote:

Originally Posted by Filburt /img/forum/go_quote.gif How many samples were taken at each point in time of each configuration?

The total was 131,072 samples. It was done at 48Khz and it took 2.73 seconds. The sweep was from 2 Hz to 20 Khz.This was done for all configurations. So it means that 6.55 samples were taken for each frequency.

I'll post the Porta Corda III measurements I made tomorrow (I left them on the other computer), I am glad you guys have liked these so far.

 

[oicdn]

I'm not understanding how to read it...

I don't have time atm to read the article (on my way out the door) but I'll read it when I get back....

 

[bludo]

Is this with the DAC or amp only? Recorded with the microphone or direct line out > line in ?

Thanks

 

[Crossfeed]

Could you please post the last graph with a more zoomed Y-axis for example using the range 72..80. It's hard to see much difference as it is now.

What's your estimation of measurements faults/variations? Some error bars are always nice

 

[Artemio]

Quote:

Originally Posted by bludo /img/forum/go_quote.gif Is this with the DAC or amp only? Recorded with the microphone or direct line out > line in ? Thanks

This was with the amp only, direct line out > line in and calibrated as described above to eliminate the sound card response.


Quote:

Originally Posted by Crossfeed Could you please post the last graph with a more zoomed Y-axis for example using the range 72..80. It's hard to see much difference as it is now. What's your estimation of measurements faults/variations? Some error bars are always nice

Here are the same graphs but zoomed in to the range 72..78, you are right they show more detail. But for those that could feel that they also show bigger flaws keep in mind that the biggest variation here is 2 dbs.

The three graphs: (linear scale for Hz)

[size=x-small]Red: New CF On
Green: Burn in CF on
Purple: Burn in CF off[/size]

Corda Move, the three graphs: (log scale for Hz)

[size=x-small]Red: New CF On
Green: Burn in CF on
Purple: Burn in CF off[/size]

Burn in
Corda Move Burn in compare: (linear scale for Hz)

Red: New CF On
Green: Burn in CF on

Corda Move Burn in compare: (log scale for Hz)

Red: New CF On
Green: Burn in CF on

Crossfeed
Corda Move Crossfeed compare: (linear scale for Hz)

Green: Burn in CF on
Purple: Burn in CF off

Corda Move Crossfeed compare: (log scale for Hz)

Green: Burn in CF on
Purple: Burn in CF off


Porta Corda III
Crossfeed compare: (linear scale for Hz)

Blue: CF on
Orange: CF off

Crossfeed compare: (log scale for Hz)

Blue: CF on
Orange: CF off

Move vs Porta Coda III
Crossfeed On compare: (linear scale for Hz)

Green: Corda Move
Blue: Porta Corda III

Crossfeed On compare: (log scale for Hz)

Green: Corda Move
Blue: Porta Corda III

Move vs Porta Coda III
Crossfeed Off compare: (linear scale for Hz)

Purple: Corda Move
Orange: Porta Corda III

Crossfeed Off compare: (log scale for Hz)

Purple: Corda Move
Orange: Porta Corda III

Regarding the error, I am not sure of an exact number.. but I can tell you that I don't notice a difference every time I measure with all things kept equal (same cables, volume, etc)

 

[lerio]

First thank's for this amazing job

I like to have Fact to add with personal human taste to compare headphone or Amp and you start to do it for Amp like headphone.com made graph for headphone

That say i'am Newbie Total in this theory stuff and i'll want to ask you if you can in simple word, to explain what kind of line on the graph should be to show to say that one is better then another one ?? does it have to be the closer possible to line 75 ? the most horizontal possible without any inclination ?

Interesting to see that burn int in Fact make a change on a Amp which i was not sure

Welcome here

 

[Filburt]

Quote:

Originally Posted by Artemio /img/forum/go_quote.gif The total was 131,072 samples. It was done at 48Khz and it took 2.73 seconds. The sweep was from 2 Hz to 20 Khz.This was done for all configurations. So it means that 6.55 samples were taken for each frequency. I'll post the Porta Corda III measurements I made tomorrow (I left them on the other computer), I am glad you guys have liked these so far.

I think you missunderstand. I mean, how many trials did you do for each. E.g., how many times did you test each state before graphing.

Secondly, can you describe your methodology here? Is there a volume control involved? Was the load identical in both cases? Was your input device operating identically in both cases? I think substantially more information about this is needed here before conclusions, like those some posters are drawing, can really be made soundly.

 

[Crossfeed]

Nice! Easier to see now

It was some time since I studied signal theory. It seems like Move is responding equally for all frequencies when crossfeed is off (given that measurements is like +-0.3 dB correct). Move seem to respond in the same way when crossfeed is off except for the range 20-~924. I wonder if this is intentional to compensate for the feeling of lack/decreased amount of bass when using some crossfeed filters.

Do you know how much "load" Move experiences during these measurements? When using phones the load might perhaps be quite different.

It would be really interesting to compare Move with some other amp.