Level Matching for DBT

[Foxenfurter]

Can more experienced members please advise on acceptable ways to level match 2 sources prior to posting DBT test results. I am looking for the simplest/easiest/cheapest and least risk approach for a newbie to follow.

 

[solderdude]

For starters we would need to know what you want to DBT.
Files, DAC's, Amps, cables, speakers ?

 

[Soniclife]

For starters we would need to know what you want to DBT.
Files, DAC's, Amps, cables, speakers ?

Sources were asked for, other than phono is the process not the same for everything else?

 

[Foxenfurter]

Wasn't expecting a response that quickly - thanks.

I normally try and A/B using Digital sources that have been synced. I have both speakers and headphones. I am probably more likely to hear differences on headphones unless I am testing for background noise.

I assume that I can use a microphone and level meter with the speakers and then switch to headphone if I choose.
If I go that route, do I use white/pink noise or a sine wave?
How close in db do the sources need to be matched?
Or is an voltage match the only way that would be acceptable?

 

[Blumlein 88]

Measure voltage. Measuring with a sound level meter isn't close enough for DBT.

Measure voltage out DACs, and at speaker terminals. If you have a way of matching them the one that matters is at the speaker terminals.

You need both sources within .1 db of each other optimally. That would be a difference between voltages of 1.2% or less.

So if one source is 1.00 volts the other should be between 1.012 volts and .988 volts.

 

[levimax]

Can more experienced members please advise on acceptable ways to level match 2 sources prior to posting DBT test results. I am looking for the simplest/easiest/cheapest and least risk approach for a newbie to follow.

I just went through this and made some home made attenuators out of a 20 K ohm 12 turn potentiometer and some old RCA cable ends ... these are easy and accurate to adjust and they don't change once set. I think you can buy something like this but if you can solder the parts are only a few dollars. I then used REW (free software) to generate a 500 Hz signal to send to the speaker and used the attenuator on the higher gain (louder) device to match the voltage levels between the devices. Good luck and I think you will find it worth the trouble.

 

[Foxenfurter]

Yes, that was going to be next question about how to fine tune the levels to ensure they match, I had been trying to do it by ear, but was not convinced I had it right so gave up. I don't have the time or ability to add an attenuator, and even then there would be complaints from some parties that this itself is colouring the sound. So it seems like there is quite a high barrier to entry for making meaningful listening comparisons between audio sources, which also explains why we end up either having measurements or impressions and it is so difficult to reconcile the objective and subjective worlds.

 

[Speedskater]

Measure voltage. Measuring with a sound level meter isn't close enough for DBT.
Measure voltage out DACs, and at speaker terminals. If you have a way of matching them the one that matters is at the speaker terminals.
You need both sources within .1 db of each other optimally. That would be a difference between voltages of 1.2% or less.
So if one source is 1.00 volts the other should be between 1.012 volts and .988 volts.

Note that you don't need an accurate or a true RMS meter. The meter's error will be the same for both readings.

 

[Blumlein 88]

Yes, that was going to be next question about how to fine tune the levels to ensure they match, I had been trying to do it by ear, but was not convinced I had it right so gave up. I don't have the time or ability to add an attenuator, and even then there would be complaints from some parties that this itself is colouring the sound. So it seems like there is quite a high barrier to entry for making meaningful listening comparisons between audio sources, which also explains why we end up either having measurements or impressions and it is so difficult to reconcile the objective and subjective worlds.

Okay, do you run a separate preamp? You can mark or use the volume control or rather have someone do it for you.

Also how different are the devices you are trying to compare? If they differ by say 6 db or less, I'd copy the file and attenuate it digitally in software so the two devices match. When you switch devices switch files. The result would be even. This pulls in some possible confounders, but nothing too serious in most cases.

There is a barrier, but it isn't all that high. Plus, no matter how many hours you compare sighted and unmatched, it is a waste of time. If you want meaningful results you have to meet certain minimum criteria.

 

[Foxenfurter]

" If they differ by say 6 db or less, I'd copy the file and attenuate it digitally in software so the two devices match. "

Great idea. Thanks

 

[Bob-23]

Not only level matching is necessary but also instantaneous switching between the sources; small differences - which we usually find in amps or dacs, if there are any, at all - can only be discerned by using an AB-switch box. When switching seamlessly, you usually get immediatly, if there's a difference, or not, even if it's not fully 'blindly' done. (But you may let somebody else do the switching, or you try to obfuscate before yourself, which position of the switch (if it's a toggle) switches on which amp, or you find a push button at which both positions of the button are undiscernable, optically and acoustically (hard to find).

If you don't already have a multimeter, here's a cheap one with astounding capabilities (I'd not test mains voltages with it!):
The Aneng 8008, Dave Jones tested it on EEV Blog.

https://www.amazon.com/Voltage-Amme...d=1&keywords=aneng+8008&qid=1603882201&sr=8-4

 

[Soniclife]

I don't have the time or ability to add an attenuator, and even then there would be complaints from some parties that this itself is colouring the sound.

I only see that as an issue if you get an identifiable difference that is surprising, if you cannot detect a difference it does not really matter.