IBD - nVidia downgraded on patent license cliff with Intel

[RampantAndroid]

Integrated motherboard sound is not even in the ballpark of "being as good as sound gets." That said, it's pretty obvious that the sound card is dead. Today people who care about sound quality use external amps and DACs so those components are free from electrical interference from the rest of the electronics in the computer. Add-in sound cards just don't make that much sense.

Eh, the Asus Xonar ones do, if you don't want an external amp/DAC setup.

 

[HurleyBird]

Eh, the Asus Xonar ones do, if you don't want an external amp/DAC setup.

EM shielding only goes so far, and even then you're using the same power source to power the sound card that you are using for your CPU, GPU, etc. Low EM interference and clean power are so insanely critical to extracting good sound quality but PC enclosures are about the worst environments possible for that.

 

[tviceman]

Do you mean Intel would have to stop production of products already on the market that use the cross licensed tech? I cant imagine Intel would ever make such an agreement. It would put them totally at the mercy of nVidia forever if that were true.

I'm not sure how it works, but everyone and their mom pays a licensing fee to Qualcomm for baseband modems. I think commonly used IP, or IP that is heavily relied upon, has to be negotiated "at fair prices" otherwise it goes to court to determine what is fair. The current licensing agreement was brought about because Nvidia sued Intel over locking them out of the chipset business, and Intel settled out of court by granting Nvidia licensing fees for GPU IP. I don't know if there were licensing fees prior to the settlement, but settling in that manner sets a precedent. It's likely, IMO, that a licensing renewal will occur and will be AT LEAST as much as the current licensing deal in place. If Intel is selling MORE chips that use the IP in question, Nvidia could use that to leverage a higher licensing agreement.

Personally, I would *LOVE* to see Intel and Nvidia team up. Intel grants Nvidia access to it's fabs, Nvidia grants Intel full use of it's graphics IP to integrate into it's CPU's and SOC's. It would kill off Nvidia's low end GPU's, but it would put Nvidia smack square into the mobile game AND it would give Nvidia a massive boost in future GPU performance with Intel's fab. It's a nerd's wet dream, but wouldn't it be nice!

 

[RampantAndroid]

EM shielding only goes so far, and even then you're using the same power source to power the sound card that you are using for your CPU, GPU, etc. Low EM interference and clean power are so insanely critical to extracting good sound quality but PC enclosures are about the worst environments possible for that.

You can do a lot of filtering on power, that as long as you have enough power on the correct voltage rails, you're fine. The nicer ASUS cards require a molex for powering the amp section, and if you isolate to a rail not driving the GPU, you're fine.

EM interference is one thing, but shielding will go far enough, frankly. At lower impedance I think you'd hear less of it, too. More at higher impedances as you need a higher voltage (I'm thinking of the 600 ohm DT880s, for example. Probably electrostatics too, though they need a different amp) you might hear more. A lot of money you put into headphone amps is snake oil, let's be honest (past a point. I've built DACs and amps, and I get some of the complexity and amount that a single op amp can affect the overall sound, but there's no reason to buy a $2k headphone amp. Many times, you can build amps equal to the more expensive Macs and whatnot for a fraction of the cost.

I'd like to see more speaker setups meant for PCs take in a digital signal, however.

 

[stahlhart]

Moving to General Hardware.
-- stahlhart

 

[Cerb]

How long has Intel been "coming at the graphics market"?

I've yet to meet an Intel graphics product I would use for even mid-range gaming.

90s: D:
00s, Pre-GMA: D: (but, hey, for single monitor office PC on the cheap...).
GMA: don't be so cheap, get a Geforce 6200 or something (unless it's a price-point PC...).
HD graphics: nice, but can you do a little better? X, Y, or Z isn't working well enough.
Haswell: why is anyone still selling $50 video cards? Who's buying them? You don't even need the cheap Quadro, now!

Midrange gaming isn't their target (yet). After many years, they finally displaced what used to be low-cost high-volume parts from nV and AMD as unnecessary. In ye olde days, a good laptop would have a Radeon or Savage in it. Then Radeons or Geforces. Now, Intel is fine. I want a dGPU for gaming, sure, but it's fine for the desktop. Get 2-4x the performance, and the added bandwidth coming with commodity DDR4, and Intel IGP will be fine for 2-3 4K monitors. System RAM technology will need at least another generational jump, along with large on-package caches becoming standard, for them to target all but the lowest-end of gaming.

That's where it hurts both other companies (no more Intel+AMD for the mass market), but nV especially, since they don't have any integrated PC solutions anymore (I hate using that word, but I can't think of a better way to say it, ATM). Intel can eat at their markets from the bottom, by including something just good enough.

 

[Cerb]

At lower impedance I think you'd hear less of it, too. More at higher impedances as you need a higher voltage (I'm thinking of the 600 ohm DT880s, for example.

That doesn't work with impedance, unless the amp is bad enough that it's allowing the headphone cables to act as antennas, or if the headphone impedance affects the feedback behavior a great deal (shouldn't be happening). You could have a ~50 Ohm headphone that sound about as loud as the same voltage as a 400 Ohm one (I have just that ). Even with lower sensitivity as dB/V (usually XdB@1V), that will need the buffer section to be what is picking up and/or rectifying the interference. It can come in anywhere from the DAC's output pins to the pins going to the jack(s) (the buffer's output itself shouldn't be a problem, though). From the DAC up to the buffer's inputs or feedback path, or if on the virtual ground somewhere (especially if the buffer's reference is not carrying the headphone ground currents), impedance at the output won't make much, if any, difference. The line-level traces, and bipolar inputs they go to, aught to be more of a concern for both shielding and coupling prevention.

 

[RampantAndroid]

That doesn't work with impedance, unless the amp is bad enough that it's allowing the headphone cables to act as antennas, or if the headphone impedance affects the feedback behavior a great deal (shouldn't be happening). You could have a ~50 Ohm headphone that sound about as loud as the same voltage as a 400 Ohm one (I have just that ). Even with lower sensitivity as dB/V (usually XdB@1V), that will need the buffer section to be what is picking up and/or rectifying the interference. It can come in anywhere from the DAC's output pins to the pins going to the jack(s) (the buffer's output itself shouldn't be a problem, though). From the DAC up to the buffer's inputs or feedback path, or if on the virtual ground somewhere (especially if the buffer's reference is not carrying the headphone ground currents), impedance at the output won't make much, if any, difference. The line-level traces, and bipolar inputs they go to, aught to be more of a concern for both shielding and coupling prevention.

Tried to reply sooner...but we were all banned

You cannot leave the volume dial in the same place and up the impedance of the headphone and keep the same voltage, assuming everything else remains the same. I get that impedance is measured with an alternating current instead of DC and that it's z, not r. Regardless, higher impedance requires a higher voltage to drive it. A higher voltage is more likely to show off any noise you have on your source signal.

 

[Cerb]

Tried to reply sooner...but we were all banned

You cannot leave the volume dial in the same place and up the impedance of the headphone and keep the same voltage, assuming everything else remains the same. I get that impedance is measured with an alternating current instead of DC and that it's z, not r. Regardless, higher impedance requires a higher voltage to drive it.

But in reality, the bolded does not remain the same. As a user, dB@1V is a much more useful metric than per mW, for this very reason. Just like high-impedance drivers being, "hard to drive," it's a generalization that is too often wrong.

A higher voltage is more likely to show off any noise you have on your source signal.

How so? Noise coming in with the signal is going to be amplified the same regardless of whether it's 500uV or 5mV, and is correctly measured by -dB. Common (often common mode, too) noise within an amp tends to be more audible at lower volumes, being nearly constant in voltage regardless of output. Higher voltage is only going to show more noise if something is creating nonlinear voltage noise with or in the amp.

 

[dark zero]

I am wondering... Intel Skylake iGPU is using nVIDIA technology this time? I say that since their iGPU improved dramatically