| vivona |
For various reasons, I record all audio at 16bit-44.1K. With my previous SB Live, when I reached clipping the recorded waveform was at 0VU (Ful-scale or 100%) as expected. I recently upgraded to an Audigy 2 ZS. Now, clipping occurs at -.5dB, which is about 95% of full scale. This only happens at 16bit-44.1K. If I record at 24/44.1k, or 16/48k, the peaks clip at 100%. This problem only happens with WDM drivers.
Here is my theory of what is happening. In the SB Live, the initial ADC sampling rate was 16b-48k. If you wanted to record at 16/44.1, the recording sample rate converter (SRC) only had to downsample the rate, but the resolution stayed at 16bit. But the Audigy initially samples at 24b-48k. So a 16b-44.1k recording has to undergo a downsample in sample rate as well as a change in bit resolution. In that process, about .5dB of volume is lost. That shouldn't happen. I think it is somehow related to the WDM driver because I tried an older driver and the clipping point dropped to 90% of full-scale. I figure if the amount of volume loss changed with the driver, it can be fixed in the driver.
I have been communicating with Creative support, but they can't duplicate the problem. That's where you can help. I need to convince them that other users see the same thing.
Simply launch Creative Wave Studio, select File/New and set the recording at 48,000, 16-bit. Also open the Creative Surround Mixer. Now feed any audio source (wave payback, microphone, line-in) and start recording. Adjust the record level until the VU meter is full scale, and increase the level some more to be sure you are reaching clipping regularly. All we are doing here is to establish the clipping level at 16/48.
Stop recording and zoom into the recorded waveform as tightly as you can and note that in the middle between the right and left channels the peaks almost touch the same line. They are clipping at 100% of the possible level.
Now, go back to File/New and select 44,100 and 16-bit, leave the record level the same and record again. Stop recording and look at the peaks. You should see that the peaks now are lower. If you have the same problem as I have, the bottom peaks of the left channel don't like up with the top peaks of the right channel. In essence, the clipping point is less than 100%.
Report your results here. Did you see a difference in the clipping point when changing 16bit recordings from 48k to 44.1k?
I know that recordings are not normally done with clipping. This test is done with clipping merely to illustrate the difference in the clipping point. The problem with a clipping point less than 100% is that you can't rely on your VU meter clipping indicator to warn you of clipping, and that is why I have a problem with this.
This test should only take you a few minutes. Won't you help me right now before you close this message and move on? I would really appreciate it. |
|
|
| vivona |
Over 130 views of my original posting and no responses. That is disappointing. Surely there are Audigy 2 users reading this forum. Any suggestions on how I could get some responses?
Thanks. |
|
|
| SY |
I'm interested in trying this, I use an Audigy2ZS, but your request came when I was swamped with stuff to do. And anything involving software or computers takes me three times as long as a normal person.
I will get to it as soon as I can- it's a pretty interesting question. |
|
|
| alfsch |
excuse me, but why worry about 0,5db ?????
i have a audigy2, the real problem is, the emu10k2 works at 48khz fixed rate, so all other sampling rates will have errors (maybe except 96k).
95% clipp seem the last problem to worry about.
but your milage may ...
i modded it a little, to get direct input to the 24/96 adc, which is the best analog input on this sound card. |
|
|
| vivona |
Why worry about .5dB? When recording, most recording software VU meters have a peak indicator that will let you know if any peaks made it to clipping. That is very helpful when setting audio record levels. You can't always see the fastest clips because of the refresh rate of the VU meter, but the clip indicator will catch them. With the peaks dropped by 5% I can no longer have the benefit of the peak clip indicator. I have to set the recording level lower than normal to be sure no peaks get to 95%, then record, then search the peaks to be sure none were clipped. That is time consuming.
I prefer to stay with 16-bit, 44.1k for various reasons, so changing to 48k is not an option for me. It would really be helpful to get verification that other Audigy users have the same problem. I need to determine if the problem is with my system, or with the card. I have received a few reports of the same problem from users at another forum, but I thought an audio forum like this would get me a lot more responses, particularly from people that understand audio.
If I can convince Creative that this is a widespread problem, they will be more motivated to actually look at the issue instead of tell me to buy a bigger power supply or turn down the line-in level. My power supply is only loaded to 50%, all voltages are within tolerance, and the 95% clipping happens with wave sources so the line-in level is not the problem. |
|
|
| fmak |
[QUOTE]Originally posted by vivona
[B]For various reasons, I record all audio at 16bit-44.1K. With my previous SB Live, when I reached clipping the recorded waveform was at 0VU (Ful-scale or 100%) as expected. I recently upgraded to an Audigy 2 ZS. Now, clipping occurs at -.5dB, which is about 95% of full scale. This only happens at 16bit-44.1K. If I record at 24/44.1k, or 16/48k, the peaks clip at 100%. This problem only happens with WDM drivers.
--------------------------------------------------------------------------------
I have posted several times about the specmanship of Creative porducts - They DO NOT measure up to scratch and clipping is one of the issues.
Nevertheless, all the magazines rave about the,m and people keep buying them.
Me, I avoid all Creative products, having used quite a few of them for only short periods because noyt only do they not perfrom well compared to other high quality soundcards but they pollute the computer with rubbish software.
Thet can't even get their update website working properly! |
|
|
| Jocko Homo |
Well, can't comment on the technical aspects, but I agree on the ******* s/w. Installed one on a machine once, I can not get rid of the stinking "thunder" that blasts out of the speakers, everytime it boots up. Glad it wasn't my machine!
Even though the sound card was changed, and everything was "uninstalled".
See what good hype, er, ah.....marketing will get you.
Jocko |
|
|
| pinkmouse |
| Mind your language Jocko... |
|
|
| vivona |
I went with the Creative product because I also use SoundFonts in my MIDI work. I don't want to shell out $$ for a hardware MIDI module and have tried softsynths, but find the sound I get from my Utopia Live SoundFont to be the best. I don't know of other cards that support SoundFonts.
As an added bonus, the Audigy came with a FireWire port for my DV camera video capture, so I saved a PCI slot. I read a lot of Audigy reviews and they all were positive.
I am not having any other issues with the Audigy other than the premature clipping at 44.1k. I did not install all the Creative MediaSource stuff.
If I can verify that the clipping issue is a design defect in the Audigy, then I know I have to either live with it, or try another card. If I find other Audigy users that don't have the problem, I can explore the differences in their setups and find why our cards behave differently. Hopefully SY and others will go ahead and run the simple test. |
|
|
| vivona |
| I installed a new Creative X-Fi which can be set to not resample. Problem solved. 16/44 behaves just like it should and the RMAA test results are excellent. |
|
|
| alfsch |
what is X-Fi ? new software, driver...?
i used the kx-project software some time.
(now i use linux, alsa driver..) |
|
|
|
