Current Control vs Voltage Control: A Large Scale Study

[siarez]

Hi @Bwaslo, We are here to have these type of discussions. So your comments are very welcome
If I understood your comment correctly, you are describing the mechanism by which CD reduces distortion.
Your description is the consensus view, and I'm in full agreement with it.

 

[mtg90]

I applaud the work you have done here.

Have you compared the two methods of amplification at differing drive levels? At what is effectively 76dB / 1m the output power should be well under 1 watt for most of the drivers tested. I'm interested to see if the reduction in harmonics hold steady as output levels increase.

 

[siarez]

Thank you @mtg90. You ask a good question.

One clarification though: the ~76 dB @ 1 m estimate in my study is an anechoic measurement from a small (~3 inch) driver, so it’s not directly comparable to the SPL number someone gets by standing in a room with an SPL meter (which includes room reinforcement and possibly a larger driver). In normal listening, the measured in-room SPL can be 3-10 dB higher than the free-field value for the same driver output.

Also, for actual music playback you typically have two speakers, which increases SPL even further.

So the test level in the study may seem lower on paper than it really is in a typical stereo listening setup. Anecdotally, I can tell you that the test were pretty loud, and I could see the cone visibly move

That said, I agree that it'd be good to repeat the comparison at higher drive levels and see how the distortion delta scales. I expect the gains diminish as SPL goes up due to hitting the mechanical limits etc. For instance, CD doesn't help, if you hit the corners of the BL curve (i.e. run out of xmax.)

 

[siarez]

If you are interested to hear the difference current drive makes, I've found that two-tone tests are quite revealing.
So I recorded samples from a number of drivers and uploaded them here.
For most drivers the difference is very audible in person, but it is also audible in the recorded samples with a good pair of headphones.
I can personally tell the difference (on Airpods Pro) for all samples in the webpage, except for the Purifi.

 

[illusonic]

The FSAF measurement provides separated noise and distortion signal that one can listen to. Would be interesting for multitone and also music signals.

FSAF Measurement

 

[voodooless]

If you are interested to hear the difference current drive makes, I've found that two-tone tests are quite revealing.
So I recorded samples from a number of drivers and uploaded them here.
For most drivers the difference is very audible in person, but it is also audible in the recorded samples with a good pair of headphones.
I can personally tell the difference (on Airpods Pro) for all samples in the webpage, except for the Purifi.

I assume the drivers were EQ'ed to the same frequency response prior to making the recording?

 

[siarez]

I assume the drivers were EQ'ed to the same frequency response prior to making the recording?

Yes the tones were at a fixed dBSPL for all drivers across both VD and CD.

The FSAF measurement provides separated noise and distortion signal that one can listen to. Would be interesting for multitone and also music signals.
FSAF Measurement

Yes, FSAF provides an alternative way of auralizing distortion. I believe Michael the creator of FSAF (who is also into current drive) has done something similar to what you describe here and posted the results on the DIYAudio forum.
A difficulty with using it here is that I need the frequency response to match between VD and CD. That means I need to produce two "pre-compensated" signals, save them as wave files, and then use them with FASF. While this is possible, it is a lot of manual work with my current setup.

 

[siarez]

If you are interested to hear the difference current drive makes, I've found that two-tone tests are quite revealing.
So I recorded samples from a number of drivers and uploaded them here.
For most drivers the difference is very audible in person, but it is also audible in the recorded samples with a good pair of headphones.
I can personally tell the difference (on Airpods Pro) for all samples in the webpage, except for the Purifi.

I go my heads on a Audeze MM-100 and a Sony MDR-7506. To my surprise, the distortions are most audible on the Sony, with Airpods Pro a close second. The MM-100 is last.
The SB Acoustics SB12NRX and SB20PFCR30-8 have a fainter audible distortion than say Peerless or Faital. This was clearly audible on both Sony and Airpods, but harder to pick out on the MM-100. I expected MM-100 to be the most revealing ... maybe that's due to frequency response differences ... not sure.

Also the volume at which you play the tones matters a lot. For all three headphones the distortions became less audible at higher volumes. This could be due to headphones own distortion or some kind of psychoacoustic masking effect ...

 

[rarakijp]

I'm convinced this is a very useful and ambitious project. For the past quarter-century or so, I have enjoyed driving my own DIY 2-way and 3-way loudspeakers with current-output amplifiers. I had given up on the idea that current-driven loudspeaker systems would ever be widely and affordably available on the market, but I have hope that if more people like you undertake projects like this, there might still be a possibility.

 

[siarez]

I had given up on the idea that current-driven loudspeaker systems would ever be widely and affordably available on the market,

I don't think there is any reason why band-selective current drive amplifiers are not the default for all active speakers, other than awareness and availability of off-the-shelf solutions.
A current drive amplifier itself doesn't have to be any more expensive than a conventional voltage amplifier. Not only that, it actually allows designers to use lower cost drivers without losing sound quality. So it is a great value proposition from a business standpoint too.

but I have hope that if more people like you undertake projects like this, there might still be a possibility.

Yes, after I did this study I have been working a project to just do that for the DIY speakers community. It is a side project and I don't have big business ambitions for it. But if you (or anyone reading) is interested in giving band-selective current drive a try, I'm making small batch that should be ready in sometime in May. You can find the module specs here and a more detailed measurements/discussion of the module here. DM me if interested.

 

[rarakijp]

Yes, after I did this study I have been working a project to just do that for the DIY speakers community. It is a side project and I don't have big business ambitions for it. But if you (or anyone reading) is interested in giving band-selective current drive a try, I'm making small batch that should be ready in sometime in May. You can find the module specs here and a more detailed measurements/discussion of the module here. DM me if interested.

Thanks for your comment. I've been using two types of DIY multichannel current drive amplifiers, and I'm also very interested in your band-selective current drive amplifier.

I'll join the launch list on your website, so please let me know when it's ready.