Monoprice HTP-1 Home Theater Processor Review

[peng]

Seems to me that all the marketing arms of these amp companies (at least those that care about THX) are trying to get that 28.8V into 8ohm THX sensitivity number into the published specs rather than for full rated power (it may coincide with it for some but not necessarily) and having trouble migrating from legacy specs.

Agreed, I said something similar before. That has actually can be confusing for some people who didn't read the fine prints as the input V for rated power has become more of a norm, than the 28.8 V THX way so people assume it is for rated output even when the spec is for 28.8 V.

If an amp isn't a fully balanced design, then effectively it is taking half the input from balanced and applying the same gain (number) as unbalanced. So, it makes sense to just state one number for both input sensitivity and gain... unless they had a different treatment for both paths without "normalizing" as they call it. The input sensitivity for XLR will be interpreted as the number per leg (Parasound states as such explicitly).
Is there some math not adding up here?

May be.., the thing is, for fully, truly, or purely balanced design, yes you will typically if not always see the 1:2 ratio consistent with the "math", but consider the "not" fully balanced design examples, and tell me what you think:

- Many not fully balanced design such as Marantz and Yamaha MCH amps, Bryston amps (except their monoblocks), Parasound Halo amps have their unbalance/Balanced input sensitivity specified as 1 V/2 V, 1.2 V/2.4 V, or 2 V/4 V etc., that is RCA: XLR in a 1:2 ratio.

- ATI's so called fully balanced or pure balanced as they sometimes called them, used to maintain the same 1:2 ratio before the "normalized" treatment, but now one same number for both RCA and XLR inputs.

It seems to me you may be right, but to me it seems to me ATI's "normalized.." is actually "abnormalized" because more people may expect the more typical 1:2 ratio, whether the amp design is "fully balanced" or just "balanced".

And, by the way, imo, the so called not fully balanced simply means not end to end, that is from the first input stage to the output stage, but the output should still be 2X the unbalanced output because of the "per leg" thing, such as the way Parasound specified for their Halo amps so I think it should be always 1:2 RCA/XLR unless it gets "normalized". Agree?

One last point, amps like ATI's with the same gain/input sensitivity for both inputs have now made them more friendly to AVPs that have limited output voltage range because it makes its XLR input's gain effectively 6 dB higher, though somehow I think the normalized treatment might have potentially degraded noise performance somewhat.

 

[Vasr]

It seems to me you may be right, but to me it seems to me ATI's "normalized.." is actually "abnormalized" because more people may expect the more typical 1:2 ratio, whether the amp design is "fully balanced" or just "balanced".

Only old relics like us who have seen the past practices. The newer trend (NAD has also started doing the same thing) is to get rid of those two separate specs and just have one. The 2:1 ratio becomes implicit.

And, by the way, imo, the so called not fully balanced simply means not end to end, that is from the first input stage to the output stage, but the output should still be 2X the unbalanced output because of the "per leg" thing, such as the way Parasound specified for their Halo amps so I think it should be always 1:2 RCA/XLR unless it gets "normalized". Agree?

One last point, amps like ATI's with the same gain/input sensitivity for both inputs have now made them more friendly to AVPs that have limited output voltage range because it makes its XLR input's gain effectively 6 dB higher, though somehow I think the normalized treatment might have potentially degraded noise performance somewhat.

I am not sure I understand what you are saying. As far as I know, ATI still expects 2x unbalanced (same as unbalanced level in each leg) in its balanced input to get the same level of output (just as NAD or Parasound for example). So, all it means is that it isn't outputting 6db more with balanced outputs. So the gain number would be the same per leg but lower by 6db if you consider the 2x as the input value rather then the "normalized" 1x as input. This can create confusion but it isn't sloppy.

People really don't buy amps based on Gain numbers except for the pedantic nitpickers like us in these forums.

 

[peng]

I much prefer something like the following, it is simple and clear.

Mcintosh:

Marantz (Yamaha's similar)

 

[krizvi786]

Seems to me that all the marketing arms of these amp companies (at least those that care about THX) are trying to get that 28.8V into 8ohm THX sensitivity number into the published specs rather than for full rated power (it may coincide with it for some but not necessarily) and having trouble migrating from legacy specs.

If an amp isn't a fully balanced design, then effectively it is taking half the input from balanced and applying the same gain (number) as unbalanced. So, it makes sense to just state one number for both input sensitivity and gain... unless they had a different treatment for both paths without "normalizing" as they call it. The input sensitivity for XLR will be interpreted as the number per leg (Parasound states as such explicitly).

Is there some math not adding up here?

Thank you that makes total sense to me now. I guess I was hoping amp was fully balanced and therefore XLR would have a real advantage.

 

[krizvi786]

I much prefer something like the following, it is simple and clear.

Mcintosh:
View attachment 84311

Marantz (Yamaha's similar)
View attachment 84312

yes much clearer

 

[peng]

I am not sure I understand what you are saying. As far as I know, ATI still expects 2x unbalanced (same as unbalanced level in each leg) in its balanced input to get the same level of output (just as NAD or Parasound for example). So, all it means is that it isn't outputting 6db more with balanced outputs. So the gain number would be the same per leg but lower by 6db if you consider the 2x as the input value rather then the "normalized" 1x as input. This can create confusion but it isn't sloppy.

It may be easier to make my point clearer using ATI's specs as follow:

AT2000 and AT3000 (AT2000 is rated 200 WPC)

Input sensitivity: 1.6 volts Note: no mentioning of RCA/unbalanced or XLR/balanced, nor "per leg", or line to line if XLR/balanced
Gain: 34 dB RCA, 28 dB XLR inputs

So if the 1.6 volts is for RCA, and same for XLR based on "per leg", the 200 W rated AT2000, with gain = 34 dB
will output a whopping 800 W !

That means:

a) if the 34 dB gain for RCA is correct then the 1.6 V input sensitivity would be wrong, it would have to be 0.8 V to output the rated 200 W.
b) if the 28 dB gain for XLR is correct, then the 1.6 V input sensitivity would be correct but would have to be line to line full voltage, not per leg.

So either way, those figures on the spec sheet is either incorrect, or to give it benefits of doubt, confusing.

Now consider the AT4000 and AT6000 (AT4000 rated 200 WPC)

Input sensitivity: 1.6 volts Note: no mentioning of RCA/unbalanced or XLR/balanced, nor "per leg", or line to line if XLR/balanced
Gain: 28 dB for both RCA and XLR inputs

In this case, the math would work, as long as the 1.6 volts is based on line to line and not "per leg". If it is meant for "per leg", as I thought initially, and you alluded to now as well, then with 28 dB gain, the output would again be over 800 W !!

That means:

a) XLR: if the 1.6 volts input sensitivity is based on "per leg" XLR input voltage, the full voltage would be 3.2 V, and 28 dB gain would allow the amp to output over 800 W, not the rated 200 W. So that's my point, either the AT4000's gain is 6 dB lower than the specified 28 dB, or the input sensitivity is 0.8 V per leg or 1.6 V line to line. That is, the ATI spec sheet is incorrect, or if there is no to it than what they show, confusing..

b) RCA: No issue, 1.6 V input sensitivity, and 28 dB will work fine, to output the rated 200 W. The 1.6 V is obviously single ended, there is only always "one leg" for RCA.

Source of info: ATI spec sheets for the AT2000/3000 and AT4000/6000 attached.

People really don't buy amps based on Gain numbers except for the pedantic nitpickers like us in these forums.

I guess my non interrupted life long career in the world of electrical engineering might have something to do with that..

 

[Vasr]

Input sensitivity: 1.6 volts Note: no mentioning of RCA/unbalanced or XLR/balanced, nor "per leg", or line to line if XLR/balanced
Gain: 34 dB RCA, 28 dB XLR inputs

So either way, those figures on the spec sheet is either incorrect, or to give it benefits of doubt, confusing.

Looks like they calculate their gain from their 4ohm measurements

20 * log((300/4)/1.6) = 33.4 rounded up to 34. XLR at 2x would have a gain of 34-6 = 28db

If there is a single number for input sensitivity, always look at it as x for RCA 2x for XLR (or x for each leg). No point in creating hypothetical otherwise.

Now consider the AT4000 and AT6000 (AT4000 rated 200 WPC)

Input sensitivity: 1.6 volts Note: no mentioning of RCA/unbalanced or XLR/balanced, nor "per leg", or line to line if XLR/balanced
Gain: 28 dB for both RCA and XLR inputs

Same calculation here for XLR using 4 ohm numbers. The rounding up/down may not be totally accurate as they are likely generated by the analyzer using greater precision for input sensitivity and power, not calculated after rounding them. It should be within +/- 1dB.

Not sure why they copied the same gain number to RCA. Most likely a mis-communication to the product marketing on the "normalizing" asking for a single number. But your point is taken here.

 

[Vasr]

Thank you that makes total sense to me now. I guess I was hoping amp was fully balanced and therefore XLR would have a real advantage.

XLR would still have advantage in avoiding any picked up noise in the cables or any ground loop noises inherent in the source signal which gets subtracted out. There would be no power output advantage.

 

[peng]

Looks like they calculate their gain from their 4ohm measurements

20 * log((300/4)/1.6) = 33.4 rounded up to 34. XLR at 2x would have a gain of 34-6 = 28db

I think you picked up the wrong formula. The 28 dB or 34 dB in their specs are voltage gain and I am sure you agree with that.

Voltage gain = Vout/Vin, or in dB form, That is Voltage gain = 20 X log(Vout/Vin) dB

If fact it is not dependent of the output power. Even if you have nothing connected to the amp, that is, 0 power output, the amp will still amplify the input voltage, the output voltage will still be calculated by the gain and input voltage, that is:

Gain = Vout/Vin, so Vout = Gain X Vin, if gain is given in multiples

Or if gain is given in dB, the Vout = Vin X antilog (Gain in dB), or Vin X 10^(Gain in dB)/20)

Given ATI's specs of 28 dB, at 1.6 V input, Vout = 1.6 X (10^(28/20)) = 40.19 V, and Power output = (40.19^2)/8 = 201.9 W, into 8 ohm, or (40.19^2)/4 = 403.81 W into 4 ohm.

It is the power output that is dependent on the load resistance (resistor load), when gain and input voltage is given.

All the formula you need are here:

http://www.sengpielaudio.com/calculatorVoltagePower.htm

Thanks for keep going, appreciated. This time though, I know I am right (typo not counted..)

 

[krizvi786]

XLR would still have advantage in avoiding any picked up noise in the cables or any ground loop noises inherent in the source signal which gets subtracted out. There would be no power output advantage.

thats too bad. so therefore over short runs no benefit

i spent all this time looking for a processor with XLR outs, for essentially no benefit! lol! is that correct?

 

[peng]

thats too bad. so therefore over short runs no benefit

i spent all this time looking for a processor with XLR outs, for essentially no benefit! lol! is that correct?

That's just another one of those hifi myth, that XLR is always better. It may be, but I have seen measurements that show either way, for short runs of cables. Just search S&V for devices measured that have both RCA/XLR and you will see the results are not consistently better for XLR in terms of THD+N and even just SNR. For two channel systems, if balanced is available, I would always use them. I like the click (not all do though), and the theoretically better noise rejection whether that makes any difference or not.

One benefits often cited is that XLR, if from fully balanced (end to end) amps, you will theoretically have less even harmonics because of the cancellation. Even so, it may not result in audibly better sound quality, as it depends on how low the distortion is to begin with. Also, some people may actually prefer to have some low order even harmonics, and some thought that would make the sound "warmer", and/or more "airy" so those people may not prefer the cancellation of even harmonics.

 

[Vasr]

I think you picked up the wrong formula. The 28 dB or 34 dB in their specs are voltage gain and I am sure you agree with that.

Voltage gain = Vout/Vin, or in dB form, That is Voltage gain = 20 X log(Vout/Vin) dB

You are using the same formula dude!

You are calculating what the power should be given a gain number and input voltage reversing that formula to calculate Vout and then using the load to calculate power. This assumes a constant gain at any load.

I calculated what the gain is at max power at a given load using that formula given the input voltage by calculating what the Vout would be into a specific load using Vout = P/R.

As far as I know, the latter is how amp manufactures report gain/sensitivity in their specs (typically from max power into 8ohms). That is when you would find your calculation consistent regardless of which way you went. Max power can be limited by current capability and so the gain at max output could be different from the gain at no load (as measured by Vout).

Otherwise, for every amp the power developed into 4ohms should be exactly twice that of power into 8ohms using your calculation. This isn't true for almost every amp (with a few exceptions). So, you could claim almost every amp spec is dubious with the same logic because the amp does not output as much power as your calculations would indicate for the lower load supported (unless they are lying about the power specs) using the reported gain number typically calculated for max power at 8ohms.

What seems to have confused you is that ATI appears to have calculated the gain for the max power it can develop at 4ohms. That amp would have 24db unbalanced gain and 18db balanced gain at max power into 8 ohms. May be the marketing felt that magical 28.8-29 number would look better! Of course, this gain in the spec wouldn't be consistent with the stated power into 8ohms if you worked backwards to find power output. So, no perfect solution.

In practice, most people buying an amp don't look at the gain number. What they want to know is the max power output and for the few that venture into matching pre-amps and amps or doing gain staging with variable gain controls, what the input sensitivity is (along with noise and distortion numbers). As long as they are honest and correct, amp voltage gain is just a computed entity for all practical purposes that neither makes or breaks the amp.

You are right. My bad. The gain formula is right, I was using the wrong calculation for getting the Vout from power to verify with the gain spec as you pointed out.

Leaving aside how gain changes with 8ohm or 4ohm at max power, here is my take on the specs you have posted.

For 2000 and 3000 series, the voltage gain should be 28 for RCA and 28-6 = 22 for XLR before "normalization". Not the 34 and 28 that is in the specs.

My guess is that they took the numbers in reverse taking 28 for XLR in error and added 6 to it for RCA. Or they de-rated the power spec later for some reason (even if unlikely)

For the later series, the RCA gain does come out to 29 as written in the spec and if their "normalization" means they are simply attenuating the XLR signal by half before the amp section, the gain would remain the same for XLR relative to what the amp section sees. The XLR would still need to send 2x input to the unit to avoid being attenuated too much.

From a consumer perspective all they would need to know is the input sensitivity number and send 1x over RCA and 2x over XLR. That would be consistent with their power rating at 8ohms (but not at 4ohms which is common).

 

[audioBliss]

An advantage of XLRs is that it's quite cheap and easy to buy good cables. With RCA I've found all store bought cables at a resonable price basically useless. The connection pressure is too low, the cable connection to the connector is not solid enough, cables deliver uneven signals between them etc. I had to resort to building my own with Neutrik connectors and Belden cable. That fixed all my problems and it was quite cheap but time consuming. But with XLR it is very easy to get good cables with good connectors. The format makes so that you can basically not mess up too bad.

Also you much easier end up in ground loop hell with RCA or preamps with poor output voltage.

 

[Vasr]

thats too bad. so therefore over short runs no benefit

i spent all this time looking for a processor with XLR outs, for essentially no benefit! lol! is that correct?

Avoiding ground loop noise is independent of the run length. If, for example, you connect a PC to a USB DAC and then connect it to the amp.

 

[krizvi786]

Avoiding ground loop noise is independent of the run length. If, for example, you connect a PC to a USB DAC and then connect it to the amp.

ok thanks

 

[bigguyca]

Avoiding ground loop noise is independent of the run length. If, for example, you connect a PC to a USB DAC and then connect it to the amp.

If the subject is RCA cables then the level of ground loop noise will likely be dependent on the length of the cable and also on the characteristics of the shield.

 

[bigguyca]

That's just another one of those hifi myth, that XLR is always better. It may be, but I have seen measurements that show either way, for short runs of cables. Just search S&V for devices measured that have both RCA/XLR and you will see the results are not consistently better for XLR in terms of THD+N and even just SNR. For two channel systems, if balanced is available, I would always use them. I like the click (not all do though), and the theoretically better noise rejection whether that makes any difference or not.

One benefits often cited is that XLR, if from fully balanced (end to end) amps, you will theoretically have less even harmonics because of the cancellation. Even so, it may not result in audibly better sound quality, as it depends on how low the distortion is to begin with. Also, some people may actually prefer to have some low order even harmonics, and some thought that would make the sound "warmer", and/or more "airy" so those people may not prefer the cancellation of even harmonics.

Please provide more details about this claim. What are the overall details of the circuitry you are describing, In particular, where is the distortion created and where does the cancellation take place?

 

[peng]

Please provide more details about this claim. What are the overall details of the circuitry you are describing, In particular, where is the distortion created and where does the cancellation take place?

It is not a claim, it is a fact, though I do not believe it matters practically speaking. Cancellation is not total in practice either, it just help reduce some of the even harmonics presented to the amp. I am not going to speculate on how and where they are created, but certainly there are there from the preamp output. FFTs show them..

https://ieeexplore.ieee.org/document/1506979

There are lots of details in the TI article:

https://www.ti.com/lit/an/sloa054e/sloa054e.pdf?ts=1600894882794&ref_url=https%3A%2F%2Fwww.google.com%2F

On page 6:

"Reduced Even-Order Harmonic Distortion Expanding the transfer functions of circuits into a power series is a typical way to quantify the distortion products. Taking a generic expansion of the outputs and assuming matched amplifiers, we get: Vout+ = k1Vin + k2Vin2 + k3Vin3 + . . . , and Vout– = k1(–Vin)+ k2(–Vin)2 + k3(–Vin)3 + . . . . Taking the differential output Vod = 2k1Vin + 2k3Vin3 + . . . , where k1, k2 and k3 are constants. The quadratic terms gives rise to second-order harmonic distortion, the cubic terms gives rise to third-order harmonic distortion, and so on. In a fully-differential amplifier, the odd-order terms retain their polarity, while the even-order terms are always positive. When the differential is taken, the even order terms cancel. Real life is not quite this perfect. Lab testing of the THS4141 at 1 MHz shows that the second harmonic at the output is reduced by approximately 6 dB when measured differentially as compared to measuring either output single-ended. The third harmonic is unchanged between a differential and single-ended measurement."

Gene of Audioholics.com has written something about that in one of his review on the fully balanced Denon AVP, but I could not find that article.

 

[peng]

You are using the same formula dude!

No I did not, I used the right formula.

You are calculating what the power should be given a gain number and input voltage reversing that formula to calculate Vout and then using the load to calculate power. This assumes a constant gain at any load.

Actually it looks to me you are the one doing the "reversing" sort of calculation by using the 4 ohm, or 8 ohm rated power output to "back" calculate the gain based on the 1.6 V that is given as "input sensitivity". The trouble is, how do you know their basis/definition of that 1.6 V input sensitivity, it could be for 28.18 V output, 30 V, 50 V, could be anything, or could be for the 8 ohm rated 200 W, or 4 ohm rated 300 W. So you tried it with the 4 ohm rated 300 W and thought the calculated gain was 33.4, close enough to the 34 dB specified for the AT2000, but then how about the 28 dB specified for the AT4000 that is also rated 300 W for 4 ohm and the same input sensitivity 1.6 V?

And besides, you have an error in your formula, if you had the formula right, your results would been 26.7 dB. More to follow...

I am using the simple formula, by definition Voltage gain = 20 Log (Output V/Input V) in dB, or just Output V/Input V in multiples. Okay, I also have to make an assumption, that is, assume the 1.6 V is defined as the input voltage for the amp output to reach 200 W into 8 ohm.

I calculated what the gain is at max power at a given load using that formula given the input voltage by calculating what the Vout would be into a specific load using Vout = P/R.

As mentioned earlier, you have an error in that formula and so I can see why your calculations (the 33.4 calculated gain) in your post#587 is incorrect because Vout = P/R is wrong, it should be Vout = square root of PXR, if you had used the right formula, your calculated gain in post#587 would have been 26.7 dB, not 33.4 dB. Nothing serious there, you just forgot the square relationship, P = V^2/R, that is V²/R in case you don't like the ^ symbol spreadsheet formula uses.

As far as I know, the latter is how amp manufactures report gain/sensitivity in their specs (typically from max power into 8ohms). That is when you would find your calculation consistent regardless of which way you went. Max power can be limited by current capability and so the gain at max output could be different from the gain at no load (as measured by Vout).

That's an interesting point, but I have to disagree, but that's only after I lose sleep thinking about it, thanks to you..
Here's my counter point, a well designed power amp such as the AT2000, as long as it is operating within its designed limit, the voltage gain should in fact be constant, or very close to being constant, because hifi audio power amplifiers are linear amplifiers such that the output voltage will vary linearly with the input voltage, until it approaches it maximum/limit and starts to clip.

For example, the AT4000 is rated 200 W 8 ohm and 300 W 4 ohm, gain is 28 dB. That should be the same whether you load it with an 8 ohm or 4 ohm test resistor when operating within its linear range.

The input sensitivity is given as 1.6 V so using the formula Vout = Vin X (10^(28/20)) = 1.6 X 25.119 = 40.19 V, and P = (V^2)/R = 40.19X40.19/8 = 201.91 W, so the AT4000's specs of 28 dB gain and input sensitivity 1.6 V would work out, but the AT2000's 34 dB gain and input sensitivity 1.6 V clearly wouldn't. Will do the 4 ohm rated 300 W scenario later..

Otherwise, for every amp the power developed into 4ohms should be exactly twice that of power into 8ohms using your calculation. This isn't true for almost every amp (with a few exceptions). So, you could claim almost every amp spec is dubious with the same logic because the amp does not output as much power as your calculations would indicate for the lower load supported (unless they are lying about the power specs) using the reported gain number typically calculated for max power at 8ohms.

That is a different and separate issue. It has to do with the current and heat dissipation capability of the amp when driving low impedance loads. The gain of the AT4000 is still 28 dB, but using the same formula:

P = Vout^2/R, that is Vout = square root (P X R)
Vout = Vin X 10^(28/20), that is Vin = Vout/ (10^(28/20)

Substituting the power output, P = 300 W and load resistance, R = 4 ohms,
Vout = square root (300 X 4) = 34.64 V
Vin = 34.64/25.119 = 1.379 V

So you can see that the gain is still the same for both 8 ohm and 4 ohm resistor load, but if we assume ATI's specified input sensitivity voltage is 1.6 V to output the 8 ohm rated 200 W, then it will take only 1.379 V to output the 4 ohm rated 300 W.

Just to check: Gain = 20 X log(Vout/Vin) = 20 X log(34.64/1.379) = 20 X log(25.1189) = 28 dB, everything works out!!

Obviously voltage gain and input sensitivity are directly related, but there is a subtle difference. Gain is always defined as Output voltage/Input voltage, whereas input sensitivity is generally defined as the input voltage required to output 28.18 V (as we discussed before), or the input voltage required for the amp to output its rated power, and we both know the rated power outputs are different for different test load resistance .

So if we use the rated power for 4 ohm load, the input voltage will be lower than that required for an 8 ohm load. (the voltage gain remains the same as we are not considering output beyond "rated", that is, within the linear range only). Once the amp started to clip, then all bets are off and distortions will start to rise sharply.

What seems to have confused you is that ATI appears to have calculated the gain for the max power it can develop at 4ohms. That amp would have 24db unbalanced gain and 18db balanced gain at max power into 8 ohms. May be the marketing felt that magical 28.8-29 number would look better! Of course, this gain in the spec wouldn't be consistent with the stated power into 8ohms if you worked backwards to find power output. So, no perfect solution.

I don't really think so, and if I am confused, it is only because of the inconsistency and lack of clarity in the two specification sheets I attached earlier. The fact is, as you can see from the attached ATI spec sheets, the AT4000 and AT2000 amps are both rated 200 W 8 ohms and 300 W 4 ohms, yet one has 34 dB gain RCA input and the other 28 dB gain RCA input, the gain for XLR input for both are 28 dB, and input sensitivity are also the same 1.6 V for both.

So how can you reconcile the difference for the RCA input gain? 34 dB vs 28 dB is very different!

In practice, most people buying an amp don't look at the gain number. What they want to know is the max power output and for the few that venture into matching pre-amps and amps or doing gain staging with variable gain controls, what the input sensitivity is (along with noise and distortion numbers). As long as they are honest and correct, amp voltage gain is just a computed entity for all practical purposes that neither makes or breaks the amp.

Thank goodness, at least and at long last we finally have a lot to agree on.

Just realize I missed your post as we were cross posting, it would have saved me this super lengthy post. Sorry!

 

[lashto]

Looks like it will be coming to the EU in 2021.
Thinking about trying one but the return policy is not so good here, just the mandatory 14 days. That might not be enough, especially if they only deliver from the UK warehouse.
Anyway, still good news and hopefully it'll happen sooner than 2021.