Skip to main content ↑ ↓

Notice

Please note that most of the software linked on this forum is likely to be safe to use. If you are unsure, feel free to ask in the relevant topics, or send a private message to an administrator or moderator. To help curb the problems of false positives, or in the event that you do find actual malware, you can contribute through the article linked here.
Topic: 24-bit recording with M-Audio Audiophile 2496 (Read 8802 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

24-bit recording with M-Audio Audiophile 2496

I have purchased and installed 2 M-Audio 2496 cards into my computer. The problem is that Windows Wave interface refuses to give me 24-bit data when recording from these cards. Data comes back in 16-bit resolution with 2 bits of noise. Actual noise level could be somewhere at -85 dB. The S/PDIF input gives 24-bits but also with 1 bit containing noise.

Is it the Windows mixer to blame and what can be done about it? Interestingly that after installing an E-MU 0404 on another computer running the same software, 24-bit input was possible.

The system is WinXP SP1. I can try to reboot in Win98, but that's no solution since it won't be possible to reboot in the future whenever I'll need to record.

24-bit recording with M-Audio Audiophile 2496

Reply #1 –
WinXP has been past SP1 for a long time. I believe there were some general audio problems that were eventually solved; I don't know anything in particular about this sound card and SP1.

I also am unable to guess just what you mean about "16-bit resolution with 2 bits of noise" or "S/PDIF input gives 24-bits but also with 1 bit containing noise." What tells you that you don't get 24 bit input from the ADC?

The Windows mixer is not involved in any way.

I run Win98, but I've seen nothing to suggest differences under WinXP except the I/O device labels in various recording programs.

My noise floor Peak Amplitude values are -85dB, but I think the RMS Average of -98dB is more significant in this case. Occasional peaks are not representative of total noise.

24-bit recording with M-Audio Audiophile 2496

Reply #2 –
Quote
I also am unable to guess just what you mean about "16-bit resolution with 2 bits of noise" or "S/PDIF input gives 24-bits but also with 1 bit containing noise."

If I zoom in my 24-bit recordings of silence all the way, I see the "waveform" quantized between values 256, 0 and -256 (on occasions the values are 255, -1, -257). S/PDIF input never gives -inf level, instead each sample randomly takes value 0 or -1.

Quote
The Windows mixer is not involved in any way. I run Win98, but I've seen nothing to suggest differences under WinXP except the I/O device labels in various recording programs.

My impression was that under Win98 more of audio paths are handled by the driver. Volume control is part of the driver. The driver does mixing of multiple streams and rejects playback of unsupported ones.

Instead on XP they manage to visualy plug both hardware and software synths in one "Midi Out" bus, downconverting of hi resolution audio happens behind the back of the user (without throwing an error), all Windows audio output arrives to the sound driver as a single stream. So it's likely that audio applications can no longer talk directly to the driver, but receive and send the data from/to Windows. I read somewhere on HA that Windows mixer applies dithering when downconverting from its internal format. This is what seems to happen here, in addition to a wrong assumtion that I don't have the capability to process 24-bit.

24-bit recording with M-Audio Audiophile 2496

Reply #3 –
Are you using recent M-Audio drivers or the umpteen year old ones that come in the soundcard package? I know that some Manufacturer's announced drivers (long ago) that bypass WinXP problems but, not being an XP user, I'm not informed about many specifics. However, since the Audiophile 2496 has been in use for so long, and widely tested to operate correctly, the chances of something new being discovered about it seems unlikely. What does RMAA say about yours?

In what manner are you ‘recording silence'? The soundcard's noise floor will vary depending upon its environment. I think I've seen Average RMS values as high as -95dB, but they are usually more like -98dB to -99dB. Since that is a bit higher than the Manufacturer's specs, it is probably due to my never having worked under the laboratory conditions in which they derive their numbers.

Anyway, it isn't possible to record actual digital silence with any soundcard. There are quieter ones than the Audiophile but none approaching 24 bits of actual noise-free resolution. The sample values you are reporting do seem indicative of a 16 bit recording, however, if they are consistent over all noise floor recordings. In what application are you recording?

24-bit recording with M-Audio Audiophile 2496

Reply #4 –
That really doesn't sound unreasonable.  You're not going to actually record silence with any 24 bit ADC on the market.

24-bit recording with M-Audio Audiophile 2496

Reply #5 –
Try the ASIO drivers.
"You can fight without ever winning, but never win without a fight."  Neil Peart  'Resist'

24-bit recording with M-Audio Audiophile 2496

Reply #6 –
After that darn REVO_SD.EXE had shutdowned my machine by force, I decided to try Win98. And as I already suspected, everything worked fine there (24 bits, no extra noise). I installed VxD driver 4.13.01.0047.

Next I updated my WinXP installation from 5.10.00.5048 to 5057. No change.  The reason I first installed original drivers is to do everything by the book the first time, so that if the system doesn't work I don't have to question my actions. I know one case where an updated didn't work at all (Intel740 PV4.0).

I 'acquired' Sound Forge 8 to try the ASIO input and it indeed worked properly. Thank you for the tip, Dreamliner77.

24-bit recording with M-Audio Audiophile 2496

Reply #7 –
Replies to AndyH-Ha:

Quote
The soundcard's noise floor will vary depending upon its environment. I think I've seen Average RMS values as high as -95dB, but they are usually more like -98dB to -99dB.

The RMS level of ADC noise turned out to be at -98 dB. But in order to see and measure it I needed 24-bit representation, and not some on-off fluctuation of the least significant bit in 16 bps.

Quote
Anyway, it isn't possible to record actual digital silence with any soundcard.

But a properly configured S/PDIF input should be able to.

Quote
What does RMAA say about yours?

Results. The same as Sound Forge 6, the application I use for recording: noise floor on the spectrum graph at -120 dB. I checked "Save resulting WAV files" and they were the same 16-bit. Measurement of any low level distortions is impossible, because they're beyond the range of 16 bits. I am satisfied with the performance of AP2496 (it passed the Udial test, noise floor is low), the problem is that some bug in the software won't let me get the data from AP2496.

24-bit recording with M-Audio Audiophile 2496

Reply #8 –
S/PDIF simply passes digital data, it isn't really recording, any more than sending a file over ethernet is recording. If the source is sending a file of digital zeros, digital zeros is what you should receive, but it doesn't say anything about the quality or noise floor of the soundcard, except that S/PDIF works (or doesn't; it is pretty much a 'yes' or 'no' question).

24-bit recording with M-Audio Audiophile 2496

Reply #9 –
I am not arguing about that. Strictly speaking: with Windows Wave interface the S/PDIF i/o does not work.

24-bit recording with M-Audio Audiophile 2496

Reply #10 –
I understanding about resampling screwing up proper S/PDIF operation, not because the soundcard and its drivers don't work properly, but because the data is resampled before entry to the S/PDIF process. This card is widely used, however, and it seems rather unlikely that so many people are satisfied with it if it can't work properly under WinXP, the most widely used OS version. Have you contacted M-Audio, to find out what they have to say about operating without OS interference?

24-bit recording with M-Audio Audiophile 2496

Reply #11 –
Dumb question- are you saving 16 or 32 bit wave files?

24-bit recording with M-Audio Audiophile 2496

Reply #12 –
I am not implying that AP2496 is of deffective design or something like that. I'm just saying that this device didn't work as expected in my configuration, and now I want to find out what can be done to change it: without drastic measures such as reinstalling the operating system. I'm quite desperate because I have 3 these cards, which I ordered as some of the last pieces of hardware working properly with Win98. Otherwise I'd just dump the M-Audio and reorder an E-MU.

I have described my problem to M-Audio UK, but have not received a reply.

I suppose other XP user's could have SP2, to which maybe M-Audio has a workaround.

BTW: Is it normal to get clicks once in a while when playing hi-res data (88/96 kS)? They appear running both VxD and WDM drivers, and with different DMA buffer lengths. I think clicks are more frequent when recording something simultaneously, but I'm unsure about this.

Ultranalog :: I am creating 24-bit files. This is however irrelevant, as the sound data gets tampered with before saving to a file.

24-bit recording with M-Audio Audiophile 2496

Reply #13 –
Quote from: j7n on
Replies to AndyH-Ha:

Quote
The soundcard's noise floor will vary depending upon its environment. I think I've seen Average RMS values as high as -95dB, but they are usually more like -98dB to -99dB.

The RMS level of ADC noise turned out to be at -98 dB. But in order to see and measure it I needed 24-bit representation, and not some on-off fluctuation of the least significant bit in 16 bps.


This is normal.

24-bit recording with M-Audio Audiophile 2496

Reply #14 –
I think the most common cause of clicks is the inability to access the data from disk or write the data to disk. The most common reason is some other activity taking over too many resources, starving the audio feed. WinXP has a great many background processes, largely unneeded in many people's lives. If these background tasks (occasionally) overwhelm disk access, or RAM usage, or use too many CPU cycles, audio suffers. People seriously concerned with recording or mixing and mastering make an effort to optimize their systems to avoid such problems.

Application software can also be responsible for such problems. Especially such things as always on anti-virus programs, automatic updates that periodically check to find out if there is anything to make newer and better, anything that runs on a schedule, etc. etc..

My 233MHz K6 machine, with 64 meg of memory, running Win95, has no problem with high resolution audio, but the machine has nothing else running on it.

24-bit recording with M-Audio Audiophile 2496

Reply #15 –
There seem to be many posts on the Web about 24-bit audio under WinXP in general, i.e. not specific to M-Audio hardware. An user here explains it and suggests installing Warped Wave Wrapper. I tried to install it, but my WinXP refused to call the driver's configuration dialog and said there was a problem with my new 'hardware'. Some driver developers have created and published workarounds to this WinXP problem. I guess, E-MU also might have one.

Quote
I think the most common cause of clicks is the inability to access the data from disk or write the data to disk. WinXP has a great many background processes, largely unneeded in many people's lives. If these background tasks (occasionally) overwhelm disk access, or RAM usage, or use too many CPU cycles, audio suffers.

I don't think it's what happens here. My computer can playback normal 48/16 tracks without interruption (DirectSound, 1000 ms, 1024 samples DMA) while I launch huge DirectX games and movies. I agree about the unneeded processes and have nLite'd my system accordingly. But as I said above, clicking happens also under Win98/VxD which has been started 10 minutes ago without launching any servers. It sounded like some timing errors during communication between CPU and the AP2496 and were apparent during playback of 1 kHz sine wave.

Again no clicking with Emu. I think I'll save up for another E-MU 0404 then. 

24-bit recording with M-Audio Audiophile 2496

Reply #16 –
About clicking: do you have AMD? In such case, there are known issues with cool'n'quiet. When I had AP2496, I disabled the feature and clicks went away.

24-bit recording with M-Audio Audiophile 2496

Reply #17 –
No, I have: Intel 915P, Pentium 4 2666 MHz w/o Hyperthreading.

24-bit recording with M-Audio Audiophile 2496

Reply #18 –
Try 32-bit recording. You can convert it to 24-bit later. I'm afraid 24-bit settings my not work correctly.

24-bit recording with M-Audio Audiophile 2496

Reply #19 –
When I had 'clicking' noise from my 2496 record/playback at high sample rates, it turned out to be a resource sharing problem.  As WinXp doesn't let you easily do manual IRQ allocation (and switching the soundcard to a different socket didn't help), I just turn off the conflicting device (my aftermarket wireless card) when I'm using the 2496.
 

24-bit recording with M-Audio Audiophile 2496

Reply #20 –
I installed E-MU 0404 in this computer, so that it now has 1 emu and 1 audiophile. The emu board now works properly while 2496 continues clicking. It sounds like single omited or duplicated samples; audible not only on test signals but also during music where there is no treble.

Indeed the audiophile now sits on a single IRQ 18 together with an USB controller. However, as I said, I also tried booting the same hardware into Windows 98. Had unneeded peripherals such as serial ports, onboard Realtek, three out of four USB controllers disabled and every device had its own IRQ. Clicking persisted.

The E-MU manual says that 24 bit perfect recording is possible only when sample rate of the Windows application is set to one which the PatchMix runs at (or thinks running at while the card is slaved to incoming SPDIF). Otherwise Windows mixer kicks in. I suppose M-Audio doesn't make this exception.

These M-Audio are good cards and will get installed into Win98 machines.