Budget DAC Review: behringer UMC204HD

[Blumlein 88]

The Audient ID 24 and older ID 22 bypass the mic preamps if you use the return trs jacks.

The SPL Crimson units have line inputs separate from and bypassing the mic preamps.

 

[gino1961]

The Audient ID 24 and older ID 22 bypass the mic preamps if you use the return trs jacks.
The SPL Crimson units have line inputs separate from and bypassing the mic preamps.

Thanks a lot for the interesting information When i hear by-pass i imagine nothing more than a switch/selector
The level could be regulated at the mic preamp level to avoid overloading of the AD section
If i had the skills i would try to trace the mic preamp section and try to send the signal there and place two panel jack sockets on the back of the unit
But it is all smd these days
Maybe i am wrong but the mic preamps on board are quite noisy I think i will look for a another sound card without mic preamps I understand that great mic preamps can be built from kits

 

[Blumlein 88]

Thanks a lot for the interesting information When i hear by-pass i imagine nothing more than a switch/selector
The level could be regulated at the mic preamp level to avoid overloading of the AD section
If i had the skills i would try to trace the mic preamp section and try to send the signal there and place two panel jack sockets on the back of the unit
But it is all smd these days
Maybe i am wrong but the mic preamps on board are quite noisy I think i will look for a another sound card without mic preamps I understand that great mic preamps can be built from kits

What makes you think mic preamps are noisy? The EIN values of good ones are a handful of db more than noise from the resistances involved. The gain is what makes that thermal noise noticeable for the most part. Also, self noise in microphones and ambient noise from venues overwhelm any electronic noise. It is misplaced enthusiasm to worry about it for most gear. When audio interfaces that run thru mic pres can achieve noise levels of -105 or -110 db I would not call that noisy. There are plenty that can manage this.

Study this until it makes sense to you. Especially the part about EIN (Equivalent Input Noise).

https://www.ranecommercial.com/legacy/pdf/ranenotes/Selecting_Mic_Preamps.pdf

It seems you are imagining all kinds of issues with mic preamps which are not really a problem.

Revisiting the thing about noise, a circuit which added none of its own, has a lower limit of noise based upon the impedance of the input source. At 150 ohms that will be a little lower than -130 dbV. If you had such a perfect circuit and amplify it by 60 db, that noise will be at -70 dbV. You cannot get it any lower than this. If you don't need much gain or no gain because the source is high enough in level then your noise floor can be lower. Some of the best like RME will get you down to -120 dbV with a line level input. Many will get you down to -105 dbV or better even running thru the mic preamp. The best external mic preamps are almost sure to need 30 db of gain and often more. So you wont be harmed feeding that to an interface with noise at -100 dbV.

 

[gino1961]

What makes you think mic preamps are noisy?

Hi i am attaching the noise floor i see with Arta SW with open input/selector on line/gain at max
If i select inst the noise jumps up in a shocking way
i dont know if it is generated noise or it comes from the power supply PC power supply is not the cleanest around for sure

The EIN values of good ones are a handful of db more than noise from the resistances involved.

that is why i bought it My idea was to use it as a cheap audio analyzer Clearly it is not up to the task

The gain is what makes that thermal noise noticeable for the most part. Also, self noise in microphones and ambient noise from venues overwhelm any electronic noise. It is misplaced enthusiasm to worry about it for most gear. When audio interfaces that run thru mic pres can achieve noise levels of -105 or -110 db I would not call that noisy. There are plenty that can manage this.
Study this until it makes sense to you. Especially the part about EIN (Equivalent Input Noise).

https://www.ranecommercial.com/legacy/pdf/ranenotes/Selecting_Mic_Preamps.pdf

It seems you are imagining all kinds of issues with mic preamps which are not really a problem.

I see Then why that aweful spectrum ? i did not expect that on the basis of the specs and some comments i read here and there

Revisiting the thing about noise, a circuit which added none of its own, has a lower limit of noise based upon the impedance of the input source. At 150 ohms that will be a little lower than -130 dbV. If you had such a perfect circuit and amplify it by 60 db, that noise will be at -70 dbV. You cannot get it any lower than this.

well i see now and this could be the case here Actually if i turn down the gain the noise lowers a lot I have seen audio analyzers with very low noise Like -140dB
Like the QA403

https://www.bestcaraudio.com/wp-content/uploads/2023/03/Noise-Floor.png

What Audio Component Harmonic Distortion Numbers Mean

A look at the actual facts behind audio component harmonic distortion numbers and how the values are calculated.

www.bestcaraudio.com

If you don't need much gain or no gain because the source is high enough in level then your noise floor can be lower. Some of the best like RME will get you down to -120 dbV with a line level input. Many will get you down to -105 dbV or better even running thru the mic preamp. The best external mic preamps are almost sure to need 30 db of gain and often more. So you wont be harmed feeding that to an interface with noise at -100 dbV.

Not at all My idea was to check THD+noise of low gain preamps at 2Vrms output (the level needed to drive most power amp to full power)
Actually there is a very nice E1DA Cosmos ADC that looks impressive indeed I am thinking to buy a separate signal generator but the cheap ones have also high distortion
Clearly precision instruments cannot come cheap I am puzzled It is just a hobby not a profession

 

[Blumlein 88]

Hi i am attaching the noise floor i see with Arta SW with open input/selector on line/gain at max
If i select inst the noise jumps up in a shocking way
i dont know if it is generated noise or it comes from the power supply PC power supply is not the cleanest around for sure

that is why i bought it My idea was to use it as a cheap audio analyzer Clearly it is not up to the task

I see Then why that aweful spectrum ? i did not expect that on the basis of the specs and some comments i read here and there

well i see now and this could be the case here Actually if i turn down the gain the noise lowers a lot I have seen audio analyzers with very low noise Like -140dB
Like the QA403

https://www.bestcaraudio.com/wp-content/uploads/2023/03/Noise-Floor.png

What Audio Component Harmonic Distortion Numbers Mean

A look at the actual facts behind audio component harmonic distortion numbers and how the values are calculated.

www.bestcaraudio.com

Not at all My idea was to check THD+noise of low gain preamps at 2Vrms output (the level needed to drive most power amp to full power)
Actually there is a very nice E1DA Cosmos ADC that looks impressive indeed I am thinking to buy a separate signal generator but the cheap ones have also high distortion
Clearly precision instruments cannot come cheap I am puzzled It is just a hobby not a profession

It is because you tested an open input. That has effectively an extremely high impedance (many megaohms) as a source and maximizes thermal noise. Selecting instrument is the same situation. Typically those are padded by a 1 megaohm resistor. So the thermal noise of a megaohm resistor is very high. If you connected a 150 ohm resistor across the input the noise would drop dramatically. 150 ohms is something like the common microphone output impedance. A cheat on EIN specs is to measure with a shorted input. That lowers noise more, but is not realistic.

Also you when you say audio analyzers have like -140 db noise are you referring to the level shown in an FFT? That is not the level of total noise due to FFT gain. I don't know your level of knowledge so forgive me if you already know this, but do you know why the FFT will graph at lower noise levels that what the noise spec is?

 

[gino1961]

It is because you tested an open input. That has effectively an extremely high impedance (many megaohms) as a source and maximizes thermal noise. Selecting instrument is the same situation. Typically those are padded by a 1 megaohm resistor. So the thermal noise of a megaohm resistor is very high. If you connected a 150 ohm resistor across the input the noise would drop dramatically. 150 ohms is something like the common microphone output impedance. A cheat on EIN specs is to measure with a shorted input. That lowers noise more, but is not realistic.

Thanks a lot I will try asap I like this unit expecially considering its price It seems quite well built I bought it exactly for its very low ein declared in the specs
But i should have read this first

The standard approach for measuring EIN is to mimic a source resistance with a 150 Ω resistor, apply a gain on the pre-amp, measure the output noise, then subtract the gain to arrive at the EIN

Also you when you say audio analyzers have like -140 db noise are you referring to the level shown in an FFT? That is not the level of total noise due to FFT gain. I don't know your level of knowledge

very low I dont even understand what a FFT is ... i like the graphs when they are clean

so forgive me if you already know this, but do you know why the FFT will graph at lower noise levels that what the noise spec is?

i have no clue May i ask why ?

 

[Blumlein 88]

very low I dont even understand what a FFT is ... i like the graphs when they are clean

i have no clue May i ask why ?

If took and measured all the noise with a voltmeter over the 0 hz to 20 khz band let us say you get -90 db V. Now imagine you take and make a high pass filter for 10,000 hz to 20,000 hz, and a low pass filter for 0 hz to 10,000 hz. Measure the voltage for each portion. If it were white noise, each half of the bandwidth would measure -93 db V. Next imagine you filter it so you have 0-5 khz, 5-10 khz, 10-15 khz, and 15-20 khz. With your voltmeter each of those would now measure -96 db. As you can see as you slice the bandwidth up to 20 khz into more and more small sections the noise in each section is lower and lower.

So when you use a 16k FFT it splits the band into 8192 slices. Yes each one is a bit over 2.44 hz wide. So you are in effect measuring the noise in each little segment, and it will read much lower. So if something has that noise floor in an FFT graph running along at -140 db V, the total noise over the 20 khz bandwidth is far above this. Like around -100 db. So the FFT has let you reach into the noise floor another 39 db. That is the FFT gain I mentioned.

Now in terms of doing good measurements of other devices, your measuring instrument needs to be 15 db better than what you are measuring optimally. Otherwise you are limited in finding the ultimate limits of the Device Under Test. The Cosmos AD is pretty close to as good as it gets. A more off the shelf unit is the RME ADI Pro unit. Most other devices will somewhat limit you for measurements. You could still check noise floors for noisier items or frequency response or even for units with too high distortion with what you have. But many devices may have results near or even below what your device can do.

 

[gino1961]

If took and measured all the noise with a voltmeter over the 0 hz to 20 khz band let us say you get -90 db V. Now imagine you take and make a high pass filter for 10,000 hz to 20,000 hz, and a low pass filter for 0 hz to 10,000 hz. Measure the voltage for each portion. If it were white noise, each half of the bandwidth would measure -93 db V. Next imagine you filter it so you have 0-5 khz, 5-10 khz, 10-15 khz, and 15-20 khz. With your voltmeter each of those would now measure -96 db. As you can see as you slice the bandwidth up to 20 khz into more and more small sections the noise in each section is lower and lower.

So when you use a 16k FFT it splits the band into 8192 slices. Yes each one is a bit over 2.44 hz wide. So you are in effect measuring the noise in each little segment, and it will read much lower. So if something has that noise floor in an FFT graph running along at -140 db V, the total noise over the 20 khz bandwidth is far above this. Like around -100 db. So the FFT has let you reach into the noise floor another 39 db. That is the FFT gain I mentioned.

Now in terms of doing good measurements of other devices, your measuring instrument needs to be 15 db better than what you are measuring optimally. Otherwise you are limited in finding the ultimate limits of the Device Under Test. The Cosmos AD is pretty close to as good as it gets. A more off the shelf unit is the RME ADI Pro unit. Most other devices will somewhat limit you for measurements. You could still check noise floors for noisier items or frequency response or even for units with too high distortion with what you have. But many devices may have results near or even below what your device can do.

Hi thank you very much again for your very kind and extremely valuable explanation. There is so much more than what meets the eye in this field
I understand that the measurement of THD+noise is a very common practice. If i understand well THD is linked to the circuit and noise to power supply and circuit PSRR ?
The idea was starting from analyzing the noise. When i was listening to integrated with phono input i remember that setting the volume to max without input signal was producing noise from the speakers Noise is always a very bad thing.
I saw a video where a representative of a well-known hifi brand gave a nice representation of the background noise also called noise carpet
He indeed compared the noise just to a carpet.
In the case of a noise of a certain consistency he represented it as a shag carpet. If very small objects are dropped on that type of carpet, it becomes difficult to spot them.
If, however, the carpet has short pile then even tiny objects can be seen very well. Tiny objects are like low-level musical signals that still contribute, for example, to the sensation of the soundstage and softer noises. An explanation that I liked because even I understood it. Having said it all.
Thanks again and have a good day Gino

 

[SIY]

If i understand well THD is linked to the circuit and noise to power supply and circuit PSRR ?

The circuit is important regarding noise as well.

Tiny objects are like low-level musical signals that still contribute, for example, to the sensation of the soundstage and softer noises.

That's rather fanciful, a salesman pitch.

Drop a Lego into a shag carpet and than walk on it....

 

[gino1961]

The circuit is important regarding noise as well.

That's rather fanciful, a salesman pitch.

Drop a Lego into a shag carpet and than walk on it....

I guess his point was that it's much easier to spot the Lego on a short pile carpet than on a shag one
Sounds and space perception are linked
Bats locate objects by hearing the sound reflected by them
Like radars do with rf

 

[SIY]

Sounds and space perception are linked
Bats locate objects by hearing the sound reflected by them
Like radars do with rf

Yes, but that's irrelevant to s/n and the effects of that on soundstage. Noise is noise- if you can't hear if from the listening position, making it lower won't change anything.

 

[gino1961]

Yes, but that's irrelevant to s/n and the effects of that on soundstage. Noise is noise- if you can't hear if from the listening position, making it lower won't change anything.

I see However Bose has often proposed the solution Maybe they believe in the concept
I have no idea about how the sub is merged with the satellites

TrueSpace Digital Processing Circuitry
Bose technology that delivers a spacious, multi-channel sound experience with strong center imaging from just two visible speakers and a hideaway Acoustimass module.

maybe it is just a matter of scaling up the system in performance and quality
I think a dsp could be very handy to realize this ...

 

[ferrellms]

Do you know how many audiophiles have chided and laughed at me when people asked for a budget DAC and I recommended the Behringer or an equivalent mastering interface?

It's why I pushed to have this unit benched.

Now here is the kicker:

The 404HD can take external power so you could get a 1 amp, 5 volt LPS or even balanced power supply and see if you can get the 60Hz mains noise to go away.

It's $99 shipped from Amazon.

I have been happy with my 204HD for years and have no plans to change

 

[wineaholic]

Is 204HD still a good option? I read somewhere that they have changed components over the years and now it's worse?

 

[Blumlein 88]

Is 204HD still a good option? I read somewhere that they have changed components over the years and now it's worse?

Depends upon what you want to use it for, which features you need. It has an easily overloaded input, and a lower than usual output level. I'd think an Audient ID4 is a better choice. ID4 is about 50% more money and not quite as many outputs. Audient EVO 4 is maybe $130 and I'd try it over the 204HD.

I don't know of any change in components over the years. Not sure they could do much that way, but not sure on that. You also could look for 2nd hand ID4s or Focusrite 2i2 for similar money and a better device in my opinion. Even a 2nd or 3rd generation 2i2 is still a good device.

 

[wineaholic]

Depends upon what you want to use it for, which features you need. It has an easily overloaded input, and a lower than usual output level. I'd think an Audient ID4 is a better choice. ID4 is about 50% more money and not quite as many outputs. Audient EVO 4 is maybe $130 and I'd try it over the 204HD.

As the title of the thread implies use it as a DAC to connect it to an amplifier. Maybe connect a 48V microphone. No RCA outputs on the EVO are the jack outputs equivalent?

 

[Blumlein 88]

As the title of the thread implies use it as a DAC to connect it to an amplifier. Maybe connect a 48V microphone. No RCA outputs on the EVO are the jack outputs equivalent?

You can use an adapter for XLR or TRS to RCA just fine. The 204HD will do what it does. The main issue is the output level is only about 1 volt max. That may be iffy on driving some power amps adequately. Depends upon the power amp. You would like to see about 2 volts output on RCA. If your power amp is driven to its capability with only a volt or near it then it might be fine. What power amp do you have?

As an example if your power amp needs 2 volts for max output, only feeding it 1 volt will result in 25% of its maximum power. So you only get 25 watts out of a 100 watt amp in such a case. This varies with different amps however so yours may not need more than 1 volt.

 

[wineaholic]

The main issue is the output level is only about 1 volt max. That may be iffy on driving some power amps adequately.

Thanks, I was not aware of this, such behavior don't sound desirable. I will look for another DAC

 

[Willem]

such behavior don't sound desirable

It depends. I use one with a Tivoli Audio Model 2 radio and that has a quite high input sensitivity. So it is much happier with the Behringer than with a 2V DAC. In fact, there is quite a lot of gear that is not happy with the Red Book 2V input level.

 

[Jimster480]

It depends. I use one with a Tivoli Audio Model 2 radio and that has a quite high input sensitivity. So it is much happier with the Behringer than with a 2V DAC. In fact, there is quite a lot of gear that is not happy with the Red Book 2V input level.

Yes, it really depends on the gear overall.