vivona
Dec 10 2005, 10:00
On my Audigy 2 ZS card, regardless of the recording source (wave, MP3, line-in, etc.) the recorded waveform does not extend beyond 95% of full scale. While recording, I can see the VU meter stopping just before full scale. It behaves as if there were a hard peak limiter in the circuit. No matter how high you move the recording or source mixer slider, the waveform hits 95% and clips there. This problem only happens at sampling frequencies of 44.1kHz or lower. At 48kHz or higher, I can easily get peaks to 100% the way it does with all other soundcards I have used.
I did all tests at 16-bit because that is what my recording software supports.
It appears that this problem may be related to the internal sample rate conversion where 44.1kHz and lower are downsampled from 48kHz. I have not seen this problem with other Soundblaster cards (Live, Live Value) that do the sample rate conversion, so it must be something with the Audigy.
Has anybody else noticed this problem?
AndyH-ha
Dec 10 2005, 13:57
Built in limiting has been a trademark of CreativeLies cards since the dawn of their time. There are other threads around here about this.
B.Fink
Dec 13 2005, 06:20
Having seen this post I quickly tested my Soundblaster X-Fi card.
Fortunately this one does not have this problem - both at 44.1 and 48 kHz clock setting in creation mode. Tested with Soundforge.
vivona
Dec 21 2005, 22:15
The key to testing your card is to record the test at 16-bit. The problem doesn't happen with higher bit rates. First record at 48k-16bit and adjust your recording level where clipping occurs in the recorded waveform. Then, without changing your recording level, record the same thing at 44.1k-16bit and look at the resulting waveform. If you have the problem, you will see the clipping that occurred at 48k shows up about -.5dB below 0VU when recording at 44.1k.
The problem is not that 44.1k is clipping at -.5dB. The problem is that all audio is first recorded at 48k. If you want a lower sample rate, the 48k recording is downsampled by the card's sample rate converter (SRC) to the lower rate, in this case 44.1k. When the 48k audio hits clipping and you record at 44.1k, the SRC introduces a .5dB loss, so the clipping appears at a point .5dB below 0VU.
This problem makes the clipping indicator on your recording software useless when recording at rates lower than 48k. It represents a defective algorithm in the SRC that is not preserving the amplitude of the waveform. As far as I know, this problem is with the Audigy. I did not see it with a SoundBlaster Live card.
I am communicating with Creative support to isolate the cause of the problem. It would be very helpful if members of this forum run this simple test and report their results here.
Thanks!
Drenholm
Dec 22 2005, 01:59
Also, the Live! 24-bit seems to introduce delays when recording at 44100Hz. 48/96kHz are fine. I suppose the SRC you mention is related too and introduces the delay.
Acid8000
Dec 22 2005, 02:08
This problem with recording is present with my Creative Sound Blaster PCI 128 (Ensoniq 5880). Fortunately I don't do very much recording. Playback is fine when testing using udial (all maxed).
B.Fink - Did you test at 16-bit? Changing to the X-Fi may ultimately be one of my solutions, but I have to be sure the X-Fi doesn't have this problem before I shell out $130 for an X-Fi.
How about other Audigy users? This test should only take you a few minutes while you leave this message open. It would be very helpful if you would make the test and respond here. X-Fi users could also report if the have the problem.
Remember to test at 16-bit and using WDM drivers only.
cliveb
Jan 10 2006, 03:53
QUOTE(Acid8000 @ Dec 22 2005, 09:08 AM)
This problem with recording is present with my Creative Sound Blaster PCI 128 (Ensoniq 5880). Fortunately I don't do very much recording. Playback is fine when testing using udial (all maxed).
This may be a different problem. Most (all?) Creative cards that use the Ensoniq 137x chipset (PCI64, PCI128, original SB Live) have some sort of analogue saturation problem at about -2dB (exact level varies from card to card). You can easily see that's it's an analogue problem from the "ragged" shape of the clipped wavetops. If it were genuine digital clipping, the clipped wavetops would be ruler-flat.
vivona
Jan 10 2006, 07:49
cliveb - First, let me mention that your excellent WaveRepair program is one of the reasons why I want to stay with 16/44. For those that aren't aware of the program, visit www.waverepair.com and check it out.
I see the ragged shape of the clipping with my Audigy and, at first, thought it was an analog problem. I asked Creative if they had some sort of a limiter in the circuit and they said no. But, I also see the same ragged clipping with wave playback as my source and I believe that source stays in the digital realm. The clipping that happens at 44.1k occurs at the exact recording level that causes clipping when recording at 48k, so it appears that the sample rate converter has artifacts that cause the level to drop in an irregular fashion, so we see the passed-along clipping with a ragged edge.
I have tried every driver version I could get my hands on, including the KX drivers and nothing eliminates the problem. At this point I am seriously considering scrapping the three-month old Audigy (sell on eBay) and buying another soundcard. Because I want SoundFonts, I may try the X-Fi because Creative claims you can run it in 44.1k mode without resampling. I have been trying to verify that the X-Fi doesn't have the problem.
cliveb
Jan 10 2006, 11:34
QUOTE(vivona @ Jan 10 2006, 02:49 PM)
I see the ragged shape of the clipping with my Audigy and, at first, thought it was an analog problem. I asked Creative if they had some sort of a limiter in the circuit and they said no. But, I also see the same ragged clipping with wave playback as my source and I believe that source stays in the digital realm. The clipping that happens at 44.1k occurs at the exact recording level that causes clipping when recording at 48k, so it appears that the sample rate converter has artifacts that cause the level to drop in an irregular fashion, so we see the passed-along clipping with a ragged edge.
Interesting... I wonder if the clipping in the 137x chipset is also an artifact of SRC, and I've been assuming it was an analogue problem all this time? Mind you, for the SRC to drop the level by almost 2dB seems unbelievable, even for Creative. To test this theory I'd have to do a recording at 48kHz to see if it can get to 0dB. But my PCI128 is in the attic somewhere, in a box of old PC bits and pieces. If I ever get the time and sufficient interest I'll dig it out and do the test.
QUOTE(vivona @ Jan 10 2006, 02:49 PM)
At this point I am seriously considering scrapping the three-month old Audigy (sell on eBay) and buying another soundcard. Because I want SoundFonts, I may try the X-Fi because Creative claims you can run it in 44.1k mode without resampling. I have been trying to verify that the X-Fi doesn't have the problem.
Why not just keep the Audigy for the SoundFonts and buy a second decent soundcard (eg. M-Audio, Terratec) for recording? If you want straightforward analogue recording without anything being fiddled with, I'd steer clear of any Creative cards (especially those X-Fi things which I am led to believe do all manner of unknown DSP stuff to the signal as it passes through).
vivona
Jan 10 2006, 12:20
Interesting idea, keep the Audigy for SoundFonts. That would also save me having to buy a PCI IEEE-1394 card which the Audigy has and I would lose if I took it out.
Are there any downsides to having two soundcards at the same time? I haven't researched this arrangement. Do you end up with both showing up in your recording software with ASIO and WDM drivers, or would I disable the Audigy WDM drivers in device manager? If I do disable the Audigy WDM drivers, would that affect SoundFont playback?
cliveb
Jan 10 2006, 13:03
QUOTE(vivona @ Jan 10 2006, 07:20 PM)
Are there any downsides to having two soundcards at the same time? I haven't researched this arrangement. Do you end up with both showing up in your recording software with ASIO and WDM drivers, or would I disable the Audigy WDM drivers in device manager? If I do disable the Audigy WDM drivers, would that affect SoundFont playback?
There shouldn't be any problem having two soundcards enabled at the same time. Most recording software has a way for you to select which soundcard to use. (You said you're a Wave Repair user: if you're recording with Wave Repair then just go to File | Options | Playback/Recording | Configure Soundcard to set up which card to use).
AndyH-ha
Jan 10 2006, 13:11
Generally there is no problem with multiple soundcards in one computer. There can be certain difficulties if you try to use different soundcards together, in a coordinated fashion, such as expecting them to synch with each other, but not generally otherwise.
The idiot-proof kind of software frequently ties use to whatever soundcard you set-up as default in the OS multimedia properties. Therefore you have to make a decision which soundcard will rule from that viewpoint. However, most serious audio programs allow you chose which card will be used for what, independent of OS defaults.
Earlier OS versions could have resource conflicts, especially IRQ, if the computer contains enough devices, irrespective of whether those devices are soundcards or anything else. I don't believe that is often a problem with WinXP.
If you're going for a second card, I'd advice you take a look at the E-MU cards (for example the 0404). The 0404 is not expensive and a very good piece of hardware (plus you get all sorts of nice software with it).
The advantage of these cards is - yes this time it is an advantage - that E-MU is part of Creative and therefore they are less likely to cause conflicts with other Creative Cards is my theory. I have a 0404 and a Soundblaster Live! living happily together in my system, except for the fact that the Live! is only used as a MIDI controller.

[edit]By the way, the main reason I bought a 0404 is because I got strange clipping artifacts when making recordings on my Live! and now I understand this was probably due to the resampling...[/edit]
Acid8000
Jan 10 2006, 19:21
QUOTE(cliveb @ Jan 11 2006, 03:34 AM)
QUOTE(vivona @ Jan 10 2006, 02:49 PM)
I see the ragged shape of the clipping with my Audigy and, at first, thought it was an analog problem. I asked Creative if they had some sort of a limiter in the circuit and they said no. But, I also see the same ragged clipping with wave playback as my source and I believe that source stays in the digital realm. The clipping that happens at 44.1k occurs at the exact recording level that causes clipping when recording at 48k, so it appears that the sample rate converter has artifacts that cause the level to drop in an irregular fashion, so we see the passed-along clipping with a ragged edge.
Interesting... I wonder if the clipping in the 137x chipset is also an artifact of SRC, and I've been assuming it was an analogue problem all this time? Mind you, for the SRC to drop the level by almost 2dB seems unbelievable, even for Creative. To test this theory I'd have to do a recording at 48kHz to see if it can get to 0dB. But my PCI128 is in the attic somewhere, in a box of old PC bits and pieces. If I ever get the time and sufficient interest I'll dig it out and do the test.
I guess I could do some sort of test eventually. I don't really have the time nor the interest at the moment, honestly. I hardly do any recording. Something I've wondered though is whether the card even resamples 48000 Hz material. This seems to be the case.
This is a "lo-fi" version of our main content. To view the full version with more information, formatting and images, please
click here.