[Roller]

Quote:

I meant PCIe vs usb, and I'm going to use the titanium HD for headphones and speakers until I can afford a good DAC and amp

 
You're going to have to spend significantly more than what you paid for the Titanium HD in order to get a better DAC. The money you intend on using for a newer DAC would be better put on better headphones/speakers and a more powerful amp to drive them.

 

[gopanthersgo1]

You're going to have to spend significantly more than what you paid for the Titanium HD in order to get a better DAC. The money you intend on using for a newer DAC would be better put on better headphones/speakers and a more powerful amp to drive them.

Yeah, I think eventually I'm getting an AudioGD NFB-1.32 for a DAC, and call the DAC done until I get to summit-fi

 

[stv014]

Quote:

I meant PCIe vs usb

 
That should not be a real issue for latency either. Poor latency is caused mainly by software issues. Although with USB you can in most cases at least avoid having to use bloated and often buggy proprietary drivers.

 

[MrMateoHead]

Quote:

 
I assume you took the 2 W from the datasheet of the amplifier IC, rather than the specifications of the sound card ? What the card can actually output depends on the details of the implementation, and if the manufacturer did not provide sufficiently detailed specifications, it can only be determined with measurements. By the way, the "2 W power" is meant as the total power in stereo mode, so it is really 1 W per channel, and that assumes the chip has adequate heat-sinking. Into 600 Ω, the maximum power output of the TPA6120A2 is ~80 mW at the most commonly used +/- 12 V power supply voltage.
 
 
It is the same amplifier as above, and the maximum power output is "only" 80 mW because it is specified into 600 Ω. At a lower load impedance, it is higher, but you normally do not need more for dynamic headphones. It is actually a fairly typical implementation of the TPA6120A2, using a 2x12 V power supply, and 10 Ω output resistors. The FiiO E9 (same amplifier IC and same voltage) is comparable, while the other FiiO amplifiers (E6, E7, E10, E11, and E17) are in fact less powerful, but have lower output impedance.

stv - Can you help me understand how to read into certain specs then, because I have been going nuts trying to figure out what honest power outputs are of a variety of sources. Your assumption was correct in that I took specs from Texas Instruments rather than from the particular soundcard. I assume that companies like Asus are not going to skimp however, if they are buying a high performance part. There aren't specs for my Realtek 888 chipset as implemented into my laptop, but power output is rated as "1 volt RMS" at 32 ohms. There is no amperage specs I can find, but I suspect the chipset can make about 1 watt, based on how frickin loud it can get with my HD 595s. At impedance of 16 Ohms, the amp seems to get noisier (I can hear static through my in-ears). The Schiit Asgard is a known 1 watt amplifier, but specs "1 volt RMS" at .1% THD or something like that, 8-600 ohms. It should be more powerful than a typical soundcard, but I can't get enough spec to know for sure.
 
I do not understand why ratings are so bad in the head-fi world. My Senn HD 595s are rated something like 112 Db @ 1 Volt. My new headphones will be 93.4 Db, but this could be at 1 V or 1 mW, it isn't specified. The rules of impedance helped me figure out the Texas Instruments chip, however I still find it interesting that such a huge power cut at high impedance could still yield adequate volume on high quality phones. I am looking to get sorted out.

 

[stv014]

Quote:

 
stv - Can you help me understand how to read into certain specs then, because I have been going nuts trying to figure out what honest power outputs are of a variety of sources.

 
Well, the following simple equations show the relation between power (P), voltage (V), current (I), and resistance (R). Impedance (Z) is similar to resistance in that it is the ratio of voltage and current, but it is for AC signals and depends on the frequency, and can be a complex number if the voltage and current are not in phase. For now, to keep things simple, you can ignore the difference between resistance and impedance.
 
R = V / I
P = V * I = V * V / R = I * I * R
 
The voltage (and also current) of an AC signal can also be expressed as RMS (root mean square, the equivalent DC level that would produce the same amount of power on average), peak (the maximum absolute difference from the 0 level), and peak to peak (the difference between the maximum and minimum level). Usually, when not specified otherwise, an AC voltage is meant as the RMS voltage of a sine wave. Here is how to convert between these:
 
Vp-p = 2 * Vpeak = 2.8284 * Vrms
 
So, once you know the maximum RMS voltage or current, you can calculate the power as shown above. For high impedance headphones, the maximum output is usually voltage limited, but with some amplifiers and low impedance loads, it can also become current limited. Another factor to take into account is the output impedance of the amplifier, which, if not insignificant compared to that of the load, can reduce the voltage and thus also the power as shown below:
 
V = Vout * R / (R + Rout)
 
In the best case, you can find a graph like this for your device:

If one is available, just find the voltage where the distortion starts to rise steeply for the load impedance nearest to whatever you intend to use. So, for example, the FiiO E10 is capable of a maximum of ~1.5 Vrms output driving a 32 Ω headphone.
 
Unfortunately, the manufacturers rarely release such graphs, so you have to calculate the maximum output from whatever limited information is available. It is safe to assume that if an amplifier is capable of outputting a certain voltage into a known impedance, then it can also output at least as much (and often more) into a higher impedance. On the other hand, if the power/voltage is only specified for a high impedance, then it is not guaranteed at all that the same voltage will also be available when driving a low impedance load, because of the total (open loop) output impedance and possibly current limiting. You can actually see this on the graph above.
 
For the TPA6120A2, assuming the usual +/- 12 Volt power supply (~10 Volt peak ~= 7 Vrms output) and 10 Ω output impedance, here are some simple example power calculations:
 
16 Ω: V = 7 * 16 / (16 + 10) = 4.31 Vrms, P = 4.31 * 4.31 / 16 = 1.16 W
32 Ω: V = 7 * 32 / (32 + 10) = 5.33 Vrms, P = 5.33 * 5.33 / 32 = 0.89 W
600 Ω: V = 7 * 600 / (600 + 10) = 6.89 Vrms, P = 6.89 * 6.89 / 600 = 0.079 W
 
To convert the calculated power or voltage to maximum sound pressure level (SPL), you need to know the sensitivity of the headphones. This is usually specified in dB at either 1 mW power, or 1 Vrms voltage. At innerfidelity.com, you can also find measurements that show the amount of voltage and power needed for 90 dB SPL. Use one of the following formulas:
 
SPL =  10 * log10(P * 1000) + [dB/mW]
SPL =  20 * log10(V) + [dB/Vrms]
SPL = 10 * log10((P * 1000) / [mW/90 dB]) + 90
SPL = 20 * log10(V / [Vrms/90 dB]) + 90
 
It depends on personal preferences and the dynamic range of the music how much is enough, but if you get less than 100 dB, that is likely underpowered, while 120 dB is more than enough for basically everyone, and 110 dB is a reasonable target for most.
 
Quote:

The Schiit Asgard is a known 1 watt amplifier, but specs "1 volt RMS" at .1% THD or something like that, 8-600 ohms.

 
That is a different type of specification, it is the total harmonic distortion (THD) at the specified output voltage. Basically, it quantifies (in a rather limited way) the sound quality, rather than the maximum power output.

 

[stv014]

Quote:

I do not understand why ratings are so bad in the head-fi world. My Senn HD 595s are rated something like 112 Db @ 1 Volt. My new headphones will be 93.4 Db, but this could be at 1 V or 1 mW, it isn't specified. The rules of impedance helped me figure out the Texas Instruments chip, however I still find it interesting that such a huge power cut at high impedance could still yield adequate volume on high quality phones. I am looking to get sorted out.

 
The sensitivity of Sennheiser headphones is indeed often specified in dB/V, and very likely that is the case for the HD595, too (112 dB/mW would be an unrealistically high value). The specs are also sometimes too optimistic.
 
For most headphones, with the exception of some very inefficient planar magnetics, you do not actually need that much power. 80 mW is in fact enough for the majority of full size dynamic headphones.

 

[mojorific]

I recommend the Asus Xonar Phoebus (I also wrote up a quick review on this site for it).  I use it to drive my Senn HD 650's, and it does quite well with the higher gain options and Dolby sound.
 
I am also looking at external DAC's, because I would like to know if that would produce a better sound or not.  For now though, i'm very happy with the sound quality from my HD 650's using the Phoebus.

 

[dlwilladsen]

I have the HT Omega Claro Plus and have had no problems with it.  It might be discontinued and, if so, you could find it for a bargain  It would seem to be quality-made and the drivers work for XP, Win7, and Win8.  Mine is connected to a Klipsch 5.1 setup, but do most of my listening with USB out to a good DAC to headphone amp to phones.