Wierd sound

Discussion forum for Amadeus users

Moderator: Martin Hairer

Post Reply
RSchmidt
Posts: 1
Joined: Wed Oct 17, 2007 12:01 pm

Wierd sound

Post by RSchmidt »

Dumping a cassette recording onto Amadeus I often get a strange sound that's similar to a beep. It's an electronic sound that is not coming from the tape. It reminds me of "cripple ware" that is reminding me to register.

Any ideas? I'm using this in a school setting and I need to figure this out.

Thanks.

Macoustics
Posts: 14
Joined: Sat Oct 20, 2007 12:06 am

Buffer problems?

Post by Macoustics »

I'm experiencing the same thing.

I record LPs in 24 bits, 96 KHz, from a M-Audio FireWire AudioPhile interface (buffer set to max - 3072 samples). I record straight to AIFF-files with Amadeus Pro.

During a 20 minute LP side, there's VERY often 1-2 or more short "Brrrp"-sounds. It's clearly different from any surface noise or pops of the record itself, and double-checking reveals nothing on the record.

To me it sounds like a short hickup caused by some buffer problem.

My first attempt at recording with Amaedeus Pro resulted in a "wall of semi-white noise" after approx. 2 minutes. After buying the software I tried another recording, which was fine - except for a short "dangdang" type of echo: a short fragment of the sound was repeated.

I use a brand new MacBook Pro Core 2 Duo with 4 GB RAM and a 7200 RPM harddisk - there's power in abundance.

I turned off ALL power saving - especially the harddisk sleep. I thought my recordings would be fine.

But no - there's every once in a while this totally annoying distortion - from a very short "crrrp" to a longer "Brrrrrp" - the longer it gets, the more it resembles "cellphone distortion".

Is this Amadeus Pro's fault? Why is it IMPOSSIBLE to record sound whitout getting digital artifacts? Isn't there ENOUGH power in this computer to ensure that the recording gets written to disk without any artifacts???

Believe me - I've recorded on my old Pentium 4 PC using Wavelab and the very same M-Audio FireWire Audiophile. Not a single hickup. Ever. Not ONE.

On my brand new Mac - I get hickups almost EVERY recording.

Is Mac OS X incapable of anything resembling realtime recording, or is there a bug in Amadeus Pro?

There is NO way I should have to listen through each and every recording in order to check it for "malfunction".

I should be able to trust the result.


Please help someone - this drives me (and I guess RSchmidt too) crazy.

Surely there must me more than just two people who has experienced this? What can we do about it?

Macoustics
Posts: 14
Joined: Sat Oct 20, 2007 12:06 am

Found another one...

Post by Macoustics »

I see we're not the only ones, RSchmidt:

jjoiv writes:

"I have been using GarageBand for a while to record a podcast and have had no problems. Today I tried Amadeus Pro and everything seemed fine, but after about 2 minutes, the sound disintegrated and became overrun with static. It looks like I will have to rerecord both programs. Ay idea why this happened? I am using a mixer connected to a MacBook Pro via usb. I made sure it was selected in the preferences before I stated."
(http://www.hairersoft.com/forum/viewtopic.php?t=431)

This is EXACTLY what I'm talking about.

Macoustics
Posts: 14
Joined: Sat Oct 20, 2007 12:06 am

Should be reliable

Post by Macoustics »

I did the math - Even recording in the highest quality (24 bit/96 KHz) only amounts to 0,55 MBps.

The harddrive of a modern day Mac can at least chew through 30-50 MBps, so that shouldn't be a bottleneck at all.

Even recording 6 channel 24/192 would be no more than 3,3 MBps.

A buffer of 3072 samples at 96 KHz equals 0.032 seconds. This should be plenty.

What causes the flow of data to be interrupted for more than 0.032 seconds? Especially with no power saving active?
Last edited by Macoustics on Sat Oct 20, 2007 11:59 am, edited 1 time in total.

User avatar
Martin Hairer
Site Admin
Posts: 1975
Joined: Wed Nov 08, 2006 11:49 am
Contact:

Wierd sound

Post by Martin Hairer »

I record LPs in 24 bits, 96 KHz, from a M-Audio FireWire AudioPhile
interface (buffer set to max - 3072 samples). I record straight to
AIFF-files with Amadeus Pro.
That may be the problem. When you record directly to AIFF, it really
creates *two* copies of the recording. One contains the "raw" sound
data, and then there is the AIFF file. The reason why there are two
copies is that if writing the file format takes up lots of CPU (like
Mp3 or AAC), there is a risk that the computer falls behind with the
encoding and looses some data.

Now, as you said, two of these streams are still only about 1.1MBps
worth of data which is panuts for a moder harddrive. The problem is
that this data gets streamed to two different files, so the harddisk
has to continuously switch between one and the other. This is quite
time-consuming (seek times haven't improved that dramatically) and
may well cause the problem that you saw. Now I just changed slightly
the way Amadeus Pro switches between the two files in order to
optimise the load. It may be that the version at http://
www.hairersoft.com/Downloads/AmadeusPro.zip behaves better. If not, I
would suggest not to use the "Record to file" function when recording
at very high qualities. Regards,

Martin

HairerSoft
http://www.hairersoft.com/


_______________________________________________
Amadeus forum mailing list
Unsubscribe / change settings at http://two.pairlist.net/mailman/listinfo/forum_list

Macoustics
Posts: 14
Joined: Sat Oct 20, 2007 12:06 am

Ah - I see!

Post by Macoustics »

I see - That could very well be the case.

After all - when I think about it, 0.032 seconds isn't all that much of a buffer after all. 32 ms could very well be spent in harddisk operations. I get the feeling that Mac OS X has more overhead than Windows in some regards.

There probably isn't a bug as such in Amadeus Pro - it's just related to writing multiple files.

I started recording straight to AIFF because i had the application freeze up a couple of times, and when I restarted it I wasn't able to recover any raw data from the files put into the trash. Maybe I didn't do it right - there was a bunch of files, and I just tried opening the first one of them (if I remember correctly).

[Edit] - I now realize I hadn't made the correct settings when importing the raw data. My 24/96 recording should have been interpreted as 96000 Hz / 24 Bits - Little Endian / Stereo / 72 Bytes header. That's why I got white noise instead of the recording back...

By the way - the "wall of static" that both jjoiv and myself experienced - what is that??? I can understand a short "Brrrp", but can the program end up being "constantly lagging behind"? I mailed you short clips of the different sounds - this was one of them.

Keep up the good work.

Post Reply