Digital I/O on these CMedia chips usually is totally OK... PLL based clock so not ultra low jitter, but modern DACs should be able to cope with that just fine. Your RMAA results don't look bit-perfect but one should be able to sort that out. (Which OS?)
The part that is complete and utter rubbish is the integrated converters and analog stages. Analog measurements of a CMI8768/70 based card's output look dreadful. The chip could barely match run-of-the-mill AC97 codecs of its day (now 5+ years ago). When pitted against any decent HDAudio codec, it's looking really, really sad.
Here are the results of a few RMAA runs on the Realtek ALC262 (C version) in my "new" FSC Celsius W360, estimated early 2008 vintage. I installed a trusty Terratec DMX XFire 1024 and a Creative SB Live! 24-Bit in PCI slots #2 and #4, respectively, so I could get a better idea of how the Realtek chip performed. Note some ground loop related noise when using the dedicated cards' ADCs.
Download
OS = Vista. Sample rate switching can get quite annoying there, but if your applications don't support WASAPI in exclusive mode, you have to do it in order to avoid resampling (which is VERY obvious on the FR graph).
I don't have many complaints with this particular chip. So periodic passband ripple in the digital filter barely matches vintage CD players (newer chips eventually sorted that out), noise floor does rise with signals approaching fullscale, and the ADC just about reaches its SNR spec but is still surpassed by the old CS4294 on the XFire (it has slightly lower analog distortion though). DAC performance seems excellent for such a chip, and all the sample rates tested (44.1/48/96) worked flawlessly.
Did I mention that the front headphone out sounds just fine? It's a capacitor-coupled, estimated 47..100 ohms affair, so it's not a good match for extremely critical cans (my SE420s do sound noticeably thinned-out, while HD590s gain just a bit of extra low-end heft). I plugged in my HD580s which are my "computer cans" anyway, et voilà, great sound. The volume control usually stays under 10%. (Which kind of idiot thought that a linear volume control was a good idea?) Even on the very sensitive Shures, not a trace of hiss or other unwanted noises.
The part that is complete and utter rubbish is the integrated converters and analog stages. Analog measurements of a CMI8768/70 based card's output look dreadful. The chip could barely match run-of-the-mill AC97 codecs of its day (now 5+ years ago). When pitted against any decent HDAudio codec, it's looking really, really sad.
Here are the results of a few RMAA runs on the Realtek ALC262 (C version) in my "new" FSC Celsius W360, estimated early 2008 vintage. I installed a trusty Terratec DMX XFire 1024 and a Creative SB Live! 24-Bit in PCI slots #2 and #4, respectively, so I could get a better idea of how the Realtek chip performed. Note some ground loop related noise when using the dedicated cards' ADCs.
Download
OS = Vista. Sample rate switching can get quite annoying there, but if your applications don't support WASAPI in exclusive mode, you have to do it in order to avoid resampling (which is VERY obvious on the FR graph).
I don't have many complaints with this particular chip. So periodic passband ripple in the digital filter barely matches vintage CD players (newer chips eventually sorted that out), noise floor does rise with signals approaching fullscale, and the ADC just about reaches its SNR spec but is still surpassed by the old CS4294 on the XFire (it has slightly lower analog distortion though). DAC performance seems excellent for such a chip, and all the sample rates tested (44.1/48/96) worked flawlessly.
Did I mention that the front headphone out sounds just fine? It's a capacitor-coupled, estimated 47..100 ohms affair, so it's not a good match for extremely critical cans (my SE420s do sound noticeably thinned-out, while HD590s gain just a bit of extra low-end heft). I plugged in my HD580s which are my "computer cans" anyway, et voilà, great sound. The volume control usually stays under 10%. (Which kind of idiot thought that a linear volume control was a good idea?) Even on the very sensitive Shures, not a trace of hiss or other unwanted noises.
Last edited:
Hi,
Os is Win7 ultimate. Unfortunately the rmaa (v6) software doesn't give me any options that would work so it was mme only. The Dogbert driver won't allow 24 bit and Windows will only let me record at 44.1Khz for some reason too. I set Windows to 44/16 in the sound devices settings for both playback and record. Playback allows up to 96/16 but record seems stuck at 44/16.
I was also very surprised by the sound quality from this card. I really expected it to be junk but it's not. Even though it tests very poorly, it sounds quite good, so that's why I thought I'd start this thread.
Anyway, if you know how I can improve the windows settings, or the output, I'd be very grateful for any advice.
About the only thing I can think of to do now is to change the resistors/capacitors on the spdif output to try to make that wave a little squarer.
Os is Win7 ultimate. Unfortunately the rmaa (v6) software doesn't give me any options that would work so it was mme only. The Dogbert driver won't allow 24 bit and Windows will only let me record at 44.1Khz for some reason too. I set Windows to 44/16 in the sound devices settings for both playback and record. Playback allows up to 96/16 but record seems stuck at 44/16.
I was also very surprised by the sound quality from this card. I really expected it to be junk but it's not. Even though it tests very poorly, it sounds quite good, so that's why I thought I'd start this thread.
Anyway, if you know how I can improve the windows settings, or the output, I'd be very grateful for any advice.
About the only thing I can think of to do now is to change the resistors/capacitors on the spdif output to try to make that wave a little squarer.
Hi,
Just as an after thought, one of the reasons this PC sounds so good is the low latency. Here's a photo of the screen to show you what I mean :
cheers
Just as an after thought, one of the reasons this PC sounds so good is the low latency. Here's a photo of the screen to show you what I mean :
An externally hosted image should be here but it was not working when we last tested it.
cheers
I'm not a programmer, so I can't tell you precisely, but I expect it has to do with timing and jitter.
That is the usual answer. The next two questions
* how does it affect timing and jitter when the data delivery to the soundcard is asynchronous and clocked by the card itself?
* did you do a blind test to confirm the premise "it sounds different"?
always go unanswered.
So the usual answer is always wrong ? Hmm...
No doubt you know the difference between phase and period jitter, and no doubt you understand lost bytes. So what is your question exactly ? Are you asking me if the card is asynch and how this works on a pci bus ?
I haven't done blind tests but I have compared it to the other two PCs I have. So yes, lower latency makes a difference.
I expect a lot of your questions go unanswered because you don't seem to be asking genuine questions. I'm only answering in the hope that it'll wind you up. Hopefully, you'll understand from this that I DON'T LIKE YOU. Probably the feeling is mutual and you are only posting to try to wind me up.
Last edited:
No, I am asking about how low latency can affect the resultant sound quality apart of timely undelivery.
So you compared one PC to another PC and concluded low latency makes a difference? The blind test assumes all the rest stays unchanged, not using a different hardware alltogether.
I have been always genuinly interested in finding out if people can really hear the difference in latency they claim to hear. That is a very fair question to ask. I have never seen a blind test result of someone making the claim. My blind test revealed no difference, but I never claimed to hear the difference in the first place, thus rendering my blind test irrelevant.
Hopefully, you'll understnd from this that I DON'T LIKE YOU.
I could not care less about that. This is public forum and claims should be either accompanied by "IMO" or "I think" or supported by reasonable evidence.
Well, that would be your problem. I try to keep my posts at strictly factual level.
Apology accepted !
?? You should re-read your and my posts in this thread first. I have nothing to apologize to you for, sorry. And this is still being a factual post
There are a lot of guides out there. Try this :
Performance Tweaks - Guru3D.com Forums
It's intended for gamers with much higher bit rates, but the same principles apply.
Cheers
Haha... yeah... most of them are.
But disabling hpet, switching the processor on 100%, and disabling everything you don't need is a good way to start. I had to add a fan (12cm 190mA) to my silent mobo so I also added a 3W 75R res in series with the fan so even at full speed I can't hear it.
Having another source with a similar quality is a good way to go to know if it's progress or just perspiration.
But disabling hpet, switching the processor on 100%, and disabling everything you don't need is a good way to start. I had to add a fan (12cm 190mA) to my silent mobo so I also added a 3W 75R res in series with the fan so even at full speed I can't hear it.
Having another source with a similar quality is a good way to go to know if it's progress or just perspiration.
Adding a resistor in series with a motor is a good way to ensure it stalls when it becomes worn, or slowed by dust/hair/thickened lubricant.
Thanks - and if it stalls I won't be able to hear it.... but the temps will rise, so that's another good reason to keep the hardware monitor running. I tried a pwm led controller but it didn't like that. A pwm motor controller was fine though. In the end, the resistor was the simplest solution, and i think 50r would have been fine.
- Status
- Not open for further replies.