Topic: Korg MR1, WTF??? SOS

[WiFiJeff]

I took my MR1 for a ride this afternoon, setting up a pair of Josephson C617s with Gefell large diameter caps (MK102.1), baffled with GuySonic's "LiteGuy" going into an SD722, with the unbalanced analog tape-out from the 722 going to the Korg.  Both recordings sound great, but when I went to compare them in Wavelab I got the following:  the 722 recording at 24/192 looks like a normal tape, but the MR1 DSF recording run through AudioGate to be 24/192 wav, though it sounds fine, comes out looking fat and fuzzy, even the silent parts have a fuzz that runs to almost 50% of the signal though I hear nothing.  Looks like some sort of low frequency artifact, but when I tried exporting the DSD from Audiogate at 16/44.1 it looks normal.  At 24/96 the fuzz is faintly there, the silent parts not quite a nice line like the direct 722 tape and the 16/44.1 export.

Anybody got a clue what this is?  Or how to get rid of it?  Anyone else try exporting to 24/192?

Jeff

[WiFiJeff]

Here's one more weirdness.  When I downsample and dither the 24/192 file to 16/44.1 in Wavelab, it comes out looking normal, as well.

Jeff

[Nick's Picks]

hmmm...i haven't tried converting to 24bit anything yet.

[it-goes-to-eleven]

Maybe try a brief recording with the korg and no mics and see if you can repeat?  I'm sure korg will be confused if you send them a problem case that involves the 722...

If you can post a tiny sample of the 24/192, I'd be curious to see it.  I'd suggest an FFT to see what freq it is (I like baudline for that but it is linux only).  My guess is that something with the high freq filtering isn't right in the conversion to 192.

[newblue]

Isn't there a DSD filter option in the MR-1 menu?  Could be related to this.  Is there a similar option on AudioGate?  You could try using DiscWelder Chrome II to do the DSD conversion to PCM to see if you get the same results (pm me if interested).

[Colin Liston]

Isn't there a Korg "rep" registered on here?  Can't he help with any or all of these questions?

[WiFiJeff]

Isn't there a DSD filter option in the MR-1 menu?  Could be related to this.  Is there a similar option on AudioGate?  You could try using DiscWelder Chrome II to do the DSD conversion to PCM to see if you get the same results (pm me if interested).

Yes, you have a choice of three.  I left it at the default, #1, which seems to do the most hf cutting.  I don't see such an option in AudioGate.

I ran the Korg tonight in my pocket with a DPA4060 set into the MMA6000 preamp, same gunk (inaudible) when I convert to pcm 192/24.  By the way, it runs really HOT in a pocket, sort of adds an edge to a classical
cello recital to be anticipating I'll burst onto flame at any minute.  It is certainly high frequency stuff, since a quick and dirty eq at 20kH down 15 dB gets rid of a whole lot of it.  I have DiskWelder Steel, Chrome is a rather pricy upgrade but if it fixes this it may be worth it to me, I'll have a look.

Jeff

[Colin Liston]

I ran mine tonight for MMW.  Ran V3>analog out> MR1 and everything is distorted or brickwalled.  The levels on the MR1 were never over -12.  I was never overloading on the V3 so I assume I feed the MR1 to hot a signal.   
 I was also sending a digi signal to the JB3 and MT and this sounds fine. 

[WiFiJeff]

I ran mine tonight for MMW.  Ran V3>analog out> MR1 and everything is distorted or brickwalled.  The levels on the MR1 were never over -12.  I was never overloading on the V3 so I assume I feed the MR1 to hot a signal.   
 I was also sending a digi signal to the JB3 and MT and this sounds fine. 

Where did you set the levels?  The DPA mics I use are pretty hot, I used 20 dB of gain on the MMA6000, and set the line-in internal gain at 0 dB, this is about the same gain I need for the Sonosax MiniR82 set to hit 0 dB at 0 dBu.  I can't imagine brickwalling unless you set the line-in levels way down to adjust for the very hot V3 output, or hit mic in by mistake.

Jeff

[Nick's Picks]

when i used mine the other night, I ran line in -12db coming from an apogee MMP.  I had the gain there running fairly hot, and the "output" setting at 0 (a nice feature of that preamp).
My levels were pretty accurate.  man, those suckers are FAST on the Korg.  watch those peaks.  the levels can look like they are -6db all the time, but the peaks are 0 and over.

[Church-Audio]

Isn't there a DSD filter option in the MR-1 menu?  Could be related to this.  Is there a similar option on AudioGate?  You could try using DiscWelder Chrome II to do the DSD conversion to PCM to see if you get the same results (pm me if interested).

Yes, you have a choice of three.  I left it at the default, #1, which seems to do the most hf cutting.  I don't see such an option in AudioGate.

I ran the Korg tonight in my pocket with a DPA4060 set into the MMA6000 preamp, same gunk (inaudible) when I convert to pcm 192/24.  By the way, it runs really HOT in a pocket, sort of adds an edge to a classical
cello recital to be anticipating I'll burst onto flame at any minute.  It is certainly high frequency stuff, since a quick and dirty eq at 20kH down 15 dB gets rid of a whole lot of it.  I have DiskWelder Steel, Chrome is a rather pricy upgrade but if it fixes this it may be worth it to me, I'll have a look.

Jeff

Hey Jeff if you can send me the file I can take a look at it I only need about 10 seconds worth. I can analyze the audio and calculate the noise floor.

[Colin Liston]

and everything is distorted or brickwalled.

Bummer!  I'd suggest testing... stereo > v3 > mr1..   Run the stereo output low (if possible, you may have to use a headphone out) and step the gain up on the v3 gradually to see when problems begin...  If you were also running the mt, that would provide a reference for the total gain, etc.. I'm sure folks would be interested in those recordings.

I had the gain set to 35 on the V3, and the MR-1 set at -20.  I was also running digi out and was more concerned with that than with the analog out.  How hot is the analog out on the V3?  Isn't there an internal jumper to take some of the heat off the analog outputs?

[live2496]

I took my MR1 for a ride this afternoon, setting up a pair of Josephson C617s with Gefell large diameter caps (MK102.1), baffled with GuySonic's "LiteGuy" going into an SD722, with the unbalanced analog tape-out from the 722 going to the Korg.  Both recordings sound great, but when I went to compare them in Wavelab I got the following:  the 722 recording at 24/192 looks like a normal tape, but the MR1 DSF recording run through AudioGate to be 24/192 wav, though it sounds fine, comes out looking fat and fuzzy, even the silent parts have a fuzz that runs to almost 50% of the signal though I hear nothing.  Looks like some sort of low frequency artifact, but when I tried exporting the DSD from Audiogate at 16/44.1 it looks normal.  At 24/96 the fuzz is faintly there, the silent parts not quite a nice line like the direct 722 tape and the 16/44.1 export.

Anybody got a clue what this is?  Or how to get rid of it?  Anyone else try exporting to 24/192?

Jeff

Hi Jeff,
It sounds to me like the conversion to PCM is not applying any anti-alias filtering.

I don't know too much about DSD, but PCM requires filtering to remove frequencies that will wrap. What makes me think that this is it, is when you mentioned sample rate conversion to 44.1 which sounds fine. And applying eq at 20kHz makes it sound ok also.

What happens when you take the 192kHz file and sample rate convert to 96Khz ?
(All SRC requires filtering as does any A>D pcm sampling)

Gordon

P.S. The other possibility is that DC offset is also involved. Typically I think, low frequency filtering solves this problem, but if present in your audio, it could be coming back as an alias high frequency after conversion to pcm. (Just a guess, of course.)

[WiFiJeff]

Hi Jeff,
It sounds to me like the conversion to PCM is not applying any anti-alias filtering.

I don't know too much about DSD, but PCM requires filtering to remove frequencies that will wrap. What makes me think that this is it, is when you mentioned sample rate conversion to 44.1 which sounds fine. And applying eq at 20kHz makes it sound ok also.

What happens when you take the 192kHz file and sample rate convert to 96Khz ?
(All SRC requires filtering as does any A>D pcm sampling)

Gordon

P.S. The other possibility is that DC offset is also involved. Typically I think, low frequency filtering solves this problem, but if present in your audio, it could be coming back as an alias high frequency after conversion to pcm. (Just a guess, of course.)

There is a setting in AudioGate to do a DC cut, this does not seem to have any effect on the exported wav.  I get rid of the gunk both by exporting from DSD directly to 16/44.1 and by exporting to 24/192 and then downsampling and dithering to 16/44.1. It would be nice to have a DSD to wav export engine that took care of this rather than just trying to throw out everything, otherwise I don't see the advantage to DSD as a "future proof" solution.

Jeff

[eric.B]

and everything is distorted or brickwalled.

Bummer!  I'd suggest testing... stereo > v3 > mr1..   Run the stereo output low (if possible, you may have to use a headphone out) and step the gain up on the v3 gradually to see when problems begin...  If you were also running the mt, that would provide a reference for the total gain, etc.. I'm sure folks would be interested in those recordings.

I had the gain set to 35 on the V3, and the MR-1 set at -20.  I was also running digi out and was more concerned with that than with the analog out.  How hot is the analog out on the V3?  Isn't there an internal jumper to take some of the heat off the analog outputs?

from the grace manual

INPUT ATTENUATOR JUMPERS Jumpers J3 and J4 select the 20dB input attenuator for channel 1. Jumpers J14 and J15 select the 20dB input attenuator for channel 2. These should be set for –20dB when the V3 is to be used with professional line level input signals.


this, however, is an input attenuator which will also reduce the output digi signal..   so no, I dont think there is internal jumpers for the analog output..