Emotiva ERC-4 CD Player (DAC) Review

[amirm]

Something really must have gone wrong, as an SPDIF-Receiver into AD1955 into LM4562 I/Vs and outputs should generate a nice and clean better than -110dB spectrum at 1kHz. I doubt Emotiva R&D and QC engineers would give the shown performance a pass.
LM4562 have a tendency to produce "popcorn" noise, maybe that's what we're seeing in one channel.

I can hardly accept the excuse of not having a test signal CD at hand as it takes no more than a few minutes to generate the files and burn them on CD (the player even takes data-CDs).

Assuming company spec follows your argument that it is from CD, this is what they spec:

And this is what I got:

0.00195 in good channel versus 0.00150. Definitely in the same neighborhood.

The original request from member was to test the DAC and that is what I signed up to do. At the end, when I saw these results, I did think about spinning my test CD but found it broken as I reported. I only have one old computer left with CD drive. Not only do I have to create the tracks and burn them there, then I have to mess with AP software to measure in asych mode which is a pain to do. So I ran with what the member requested.

From past experience, these results only get worse anyway when you switch to 24 bit to 16 bit as the latter limits you to 93 dB SINAD.

 

[restorer-john]

On professional gear, genuine balanced output was either fully floating transformer-type or active servo-balanced which then gives the same output voltage regardless of how it is connected.

1:1 transformers with 600R output impedance. Output levels the same for single ended or balanced. That was the professional standard until silly home HiFi audiophools decided they 'must have' XLRs on their gear to make them feel like 'pros' and the cheapest way to do that, was to strap on an inverting unity gain opamp and call the source 'balanced'. Consequently, the output level doubled for no good reason.

And we all know, lifting the voltage up another 2V will buy some 'free' S/N.

 

[AndreaT]

1:1 transformers with 600R output impedance. Output levels the same for single ended or balanced. That was the professional standard until silly home HiFi audiophools decided they 'must have' XLRs on their gear to make them feel like 'pros' and the cheapest way to do that, was to strap on an inverting unity gain opamp and call the source 'balanced'. Consequently, the output level doubled for no good reason.

And we all know, lifting the voltage up another 2V will buy some 'free' S/N.

Hi restorer-john, remind me the best way to have a balanced output…is it perhaps a fully balanced circuit for the preamp? Why do you say the output “doubled for no good reason”? It seems a common trend in preamps to have an increased output voltage to 4 V and higher, especially in professional hardware.

 

[JaccoW]

I wanted to test it as a CD player as well but my test CD had broken in half.

What disc are you looking for? I picked up one of these Denon Audio Technical CDs yesterday for next to nothing and I believe the shop had plenty more for little money.
Or I could just send you a copy of course.

First disc I ever found with 99 tracks on it.

 

[GPJ]

So what's your favorite track?

 

[KSTR]

Assuming company spec follows your argument that it is from CD, this is what they spec:
View attachment 271046

And this is what I got:

0.00195 in good channel versus 0.00150. Definitely in the same neighborhood.

The original request from member was to test the DAC and that is what I signed up to do. At the end, when I saw these results, I did think about spinning my test CD but found it broken as I reported. I only have one old computer left with CD drive. Not only do I have to create the tracks and burn them there, then I have to mess with AP software to measure in asych mode which is a pain to do. So I ran with what the member requested.

From past experience, these results only get worse anyway when you switch to 24 bit to 16 bit as the latter limits you to 93 dB SINAD.

I'm not doubting your results... it just would have been interesting have a bit more info, notably idle (dithered zero) spectrum etc, different test frequencies etc to get a grip on the 600Hz and multiples needles and the rising LF noise. For example, if the 600Hz-related stuff is constantly present, this would be very bad (and even noticable during very soft passages of music) plus it is natural that AP's linearity measurement thengets influenced by this at very low levels as the bandpass is not infinitely sharp.

I'm not requesting that you make additional findings public with like the other graphs (for good reasons) but would encourage you do these tests -- really takes only a few minutes more and post sort of a sidebar comment, or is it really that you have little curiosity and zero time to dig deeper when you find something interesting? If so, that would be sad...

 

[amirm]

What disc are you looking for? I picked up one of these Denon Audio Technical CDs yesterday for next to nothing and I believe the shop had plenty more for little money.
Or I could just send you a copy of course.

It is a custom disc created with signals generated by Audio Precision. That way the results are consistent with when AP generates the signal in real-time.

 

[KSTR]

So there is a standard for outputting 2.1 volt??? I have tested probably 100+ balanced DACs that output 4 volt. I call that a standard. De-facto standard. And input version of those devices expect 4 volts on XLR input, not 2.1. So I am going to advocate 4 volts nominal.

There is no standard actually. The 2Vrms (+6dBV) voltage we see comes from the -10dBV "consumer" operating level from old analog days to allow for 16dB of headroom with digital sources (CD, and that level was agreed on by Sony and Philips).

I agree that doubled voltage of XLR vs RCA became a de-facto standard of sorts but I wouldn't call it wrong when the voltage is the same.
As for input voltage sensitivities, when I was in the active speaker business a key requirement there was to have XLR/TRS and RCA inputs to have the same sensitivity even for the consumer market devices.

 

[amirm]

or is it really that you have little curiosity and zero time to dig deeper when you find something interesting?

Both. I don't see much value in testing CD players in this day and age. And have negative amount of time to spend on one beyond what I promised the owner. As it is, I had to fix the darn thing just to test it.

 

[Scytales]

I have several Sony CDP-101s (the worlds first commercially released player). I measured THD much lower than the linked PDF above by Louis Challis. At first I figured it was my mistake until I looked at his THD calculations- see if you can spot the error/s

He has missed a "0" in the calculation of the 1 kHz THD, which is actually 0.00059 %, not 0.0059.

What did I won ?

 

[Scytales]

There is no standard actually. The 2Vrms (+6dBV) voltage we see comes from the -10dBV "consumer" operating level from old analog days to allow for 16dB of headroom with digital sources (CD, and that level was agreed on by Sony and Philips).

I agree that doubled voltage of XLR vs RCA became a de-facto standard of sorts but I wouldn't call it wrong when the voltage is the same.
As for input voltage sensitivities, when I was in the active speaker business a key requirement there was to have XLR/TRS and RCA inputs to have the same sensitivity even for the consumer market devices.

To the best of my knowledge, there actually is a standard - IEC 61938 (Audio, video and audiovisual systems - Interconnections and matching values - Preferred matching values of analogue signals), which imposes a maximum output level of 2 V RMS for consumer appliances (other than preamplifiers to be used with separate power amplifiers), irrespective of output topology. Here is the 1996 issue of the standard. IEC 61938 was amended in 2018. I cannot say for sure if the levels have changed from the 1996 version, but I don't think so, even if some adaptations has probably been done concerning digital sources due to the advent of 24 bits digital formats.

Of course, the existence of a standard does not mean that everyone adhere to it or that this standard is actually enforced by some mechanism.

 

[KSTR]

To the best of my knowledge, there actually is a standard - IEC 61938 (Audio, video and audiovisual systems - Interconnections and matching values - Preferred matching values of analogue signals), which mandates a maximum output level of 2 V RMS for consumer appliances (other than preamplifiers to be use with separate power amplifiers), irrespective of output topology. Here is the 1996 issue of the standard. IEC 61938 was amended in 2018. I cannot say for sure if the levels have changed from the 1996 version, but I don't think so, even if some adaptations has probably been done concerning digital sources due to the advent of 24 bits digital formats.

View attachment 271165
View attachment 271168

Of course, the existence of a standard does not mean that everyone adhere to it or that this standard is actually enforced by some mechanism.

Thanks for this.
The interesting part is that it states digital device outputs as 2Vrms +-3dB, quite a bit of tolerance and 2.8Vrms worst-case.

 

[Scytales]

It is quite possible that the +/- 3 dB tolerance actually copes with the possibility of introducing inter-samples overs in the case where the max level sample is not in phase with the peak level of the digitized analogue sine signal.

 

[restorer-john]

From past experience, these results only get worse anyway when you switch to 24 bit to 16 bit as the latter limits you to 93 dB SINAD.

Emotiva claiming 0.0015% at 0dBFS 1kHz on 16 bit content is right on the edge of what the very best CD players ever achieved. Your 0.0029% result on the worse channel puts its (24 bit?) performance worse than most of the 1st generation CD players form 1983.

He has missed a "0" in the calculation of the 1 kHz THD, which is actually 0.00059 %, not 0.0059.

What did I won ?

And we have a winner! Well done.

 

[Scytales]

And we have a winner! Well done.

Actually, I have some training in manual calculation of THD. Fastidious, but enlightening.

«Tests de distorsion de 720 sources numériques de 1992 à 2018» - 30043956 - sur le forum «Source dématérialisée et DAC» - 1356 - du site Homecinema-fr.com

Bonsoir Quel boulot ! Je crois avoir compris qu'il s'agit, par extrapolation, d'apprécier la qualité de conception des circuits, sur la capacité à ne

www.homecinema-fr.com

 

[KSTR]

It is quite possible that the +/- 3 dB tolerance actually copes with the possibility of introducing inter-samples overs in the case where the max level sample is not in phase with the peak level of the digitized analogue sine signal.

I was thinking about this, too, but it doesn't make sense to me, and I would think they haven't been aware of IS-overs at the time.
Nominal output is 0.5Vrms with a sine that hits -12dBFS on its maxima, so 2Vrms at 0dBFS sine wave. Let's say we add 3dB of headroom for IS-overs then everthing just shifts upwards by 3dB. Allowable max at an input would then extend to 4Vrms, etc.

 

[Scytales]

I think you are mistaken : 2 V RMS +3 dB is 2.8 V RMS, unless I myself misread or misinterpreted you.

 

[MediumRare]

Is there anything remotely audible in these measurements? How can it be broken? It is not even in the Poor category when ranking DACs. I understand there are far better options but what is the standard for Broken Panther?

 

[pseudoid]

Both. I don't see much value in testing CD players in this day and age. ...

As always, thank you for the review,
Recently, it was reported that there were more LPs sold than CDs, last year:
Do you feel any responsibility to start testing some turntables?

 

[amirm]

Do you feel any responsibility to start testing some turntables?

If it were feasible, I would. I even bought some test records but there just isn't a reliable method for measuring them.