You are not logged in. Please login or register.
RME User Forum → DIGICheck → Oversampled Digicheck meters at 192 kHz...
I discovered today that the oversampled meter would not detect ANY overs at 192 kHz. Perhaps because it already samples at 192 kHz and so there are no intersample peaks to judge? I was looking for a way to detect overloads of my A/D converter and apparently Digicheck would not detect them at 192 K.
BK
I see oversampled overs at 192 kHz, no problem. Seems to be your ADC doing this.
I see oversampled overs at 192 kHz, no problem. Seems to be your ADC doing this.
Thanks, MC.
Could be the ADC! What would it be, that when the ADC clips it doesn't produce any high frequency intersample products? How can that be produced?
Anyway, what is the oversampling ratio of the meters, or rather the meter sample rate at 192 kHz sampling? At 44.1 kHz sampling? Is it the same ratio always?
A clipped ADC usually does not produce many intersample peaks. Logically, as the clipped stage is followed by a steep anti-aliasing filter.
The oversampling ratio is always the same.
If you want to see it working just generate a 0 dBFS 10 kHz square wave @ 192 kHz with any software, most will draw it perfectly, means wrong, then check with the OVS option on. I got +2.9 dBFS this way.
A clipped ADC usually does not produce many intersample peaks. Logically, as the clipped stage is followed by a steep anti-aliasing filter.
The oversampling ratio is always the same.
If you want to see it working just generate a 0 dBFS 10 kHz square wave @ 192 kHz with any software, most will draw it perfectly, means wrong, then check with the OVS option on. I got +2.9 dBFS this way.
Thank you so much, MC, for that information! Yes, I knew that the ADC does not produce many intersample peaks, even when clipped because unlike digital clipping the analog filter itself does not produce aliasing products. So, how would you evaluate when an ADC clips? Use the consecutive overs technique? Can you please put consecutive OVER counting into the Digicheck software also, for those who are transferring from ADC?
BK
That statement confuses me a bit. Fire up two identical Level meters side by side, same settings, same source. Now enable OVS on one, and not on the other one. On the one without OVS you can see via F2 that 3 samples are default. Now raise your input level and see what happens. I am pretty sure that the OVS meter will detect Overs immediately, while the consecutive one will not. Maybe you did not realize that any level above 0 dBFS is an Over? Even 0.1 dBFS, and this is easy to see as the bar turns to white color then.
That statement confuses me a bit. Fire up two identical Level meters side by side, same settings, same source. Now enable OVS on one, and not on the other one. On the one without OVS you can see via F2 that 3 samples are default. Now raise your input level and see what happens. I am pretty sure that the OVS meter will detect Overs immediately, while the consecutive one will not. Maybe you did not realize that any level above 0 dBFS is an Over? Even 0.1 dBFS, and this is easy to see as the bar turns to white color then.
Hi, MC.
I'm not sure what F2 means. I didn't know there was a consecutive sample option in the unit. Anyway, I've been using the K-Meters presets and with the Cranesong HEDD-192 ADC I did not detect ANY intersample peaks or signal levels "above" full scale no matter how hard I pushed the ADC. At least at 192K. I'll check this out next week when I'm doing another analog tape transfer.
BK
F2 means the key F2 on your keyboard. It opens the Level Meter setup, which includes the consecutive sample count for Overs entry.
F2 means the key F2 on your keyboard. It opens the Level Meter setup, which includes the consecutive sample count for Overs entry.
Thanks, MC. It would seem to me that if the HEDD ADC does not produce intersample peaks for some reason, using the consecutive sample count method, however inaccurate it may be, would be a useful backup. But I agree when both methods are available, the OVS approach will give meaningful results.
-----------------
By the way, is there a way to receive emails when posts are made to subscribed topics in this forum?
By the way, is there a way to receive emails when posts are made to subscribed topics in this forum?
In the "Post Reply" screen (not in the quick reply window) you have options for "Never Show Smileys" and "Subscribe to this topic" below the text window. I receive E-mail notifications for any topics I subscribe to. I previously requested that RME add this "Subscribe" check box to the Quick Reply screen for convenience...
I previously requested that RME add this "Subscribe" check box to the Quick Reply screen for convenience...
and it has been there too, for a long time... at least on my screen :-P
looks like this, right beneath the posts on any forum page - subscription is always just one mouse click away:
-----
Pages: 1 ___ Index ? DIGICheck ? Oversampled Digicheck meters at 192 kHz... ___ Post reply
Subscribe to this topic
Quick post
Write your message and submit
-----
(haha, nice that I was able to contribute to a conversation between Bob Katz and MC )
Randyman... wrote:I previously requested that RME add this "Subscribe" check box to the Quick Reply screen for convenience...
and it has been there too, for a long time... at least on my screen :-P
looks like this, right beneath the posts on any forum page - subscription is always just one mouse click away:
-----
Pages: 1 ___ Index ? DIGICheck ? Oversampled Digicheck meters at 192 kHz... ___ Post reply
Subscribe to this topic
Quick post
Write your message and submit
-----(haha, nice that I was able to contribute to a conversation between Bob Katz and MC :cool:)
You've helped! Thanks. Can we request a function called "automatically subscribe" to any topics that we post in? And by the way, I am overwhelmed by MC's knowledge, I am only an apprentice in his eyes!
I have no personal experience with the HEDD unit. From your description it seems that they included some kind of soft-clipping before the ADC, or process the data with additional filtering on the digital domain. Really don't know. In both cases it would be complicated to detect Overs easily, with both consecutive and oversampling, at 192 kHz.
I have no personal experience with the HEDD unit. From your description it seems that they included some kind of soft-clipping before the ADC, or process the data with additional filtering on the digital domain. Really don't know. In both cases it would be complicated to detect Overs easily, with both consecutive and oversampling, at 192 kHz.
Interesting, then I will stay below -1 dBFS digital sample peak for all these transfers. I've measured the distortion and headroom of the ADC using Spectrafoo and it is excellent, better than many other ADCs and it is sonically forgiving to overs, but there is no explicit overload processing or filtering going on to my knowledge.
Bob, you do not have to lose 1 dB IF you are using DIGICheck. Simply set the count for consecutive samples to 0. In that case the OVR warning appears when you reach 0 dBFS, but not when the level is just a fraction of a db below (exact number must be in the online help of the Level Meter Setup).
Bob, you do not have to lose 1 dB IF you are using DIGICheck. Simply set the count for consecutive samples to 0. In that case the OVR warning appears when you reach 0 dBFS, but not when the level is just a fraction of a db below (exact number must be in the online help of the Level Meter Setup).
Thanks, MC. I will do. Though this is strictly a protection issue as I am not the least bit concerned about a loss of resolution or signal to noise ratio judged by a peak meter. 1 dB down is completely trivial for raw transfer.
BK
MC wrote:Bob, you do not have to lose 1 dB IF you are using DIGICheck. Simply set the count for consecutive samples to 0. In that case the OVR warning appears when you reach 0 dBFS, but not when the level is just a fraction of a db below (exact number must be in the online help of the Level Meter Setup).
Thanks, MC. I will do. Though this is strictly a protection issue as I am not the least bit concerned about a loss of resolution or signal to noise ratio judged by a peak meter. 1 dB down is completely trivial for raw transfer.
BK
And of course the object is to see a -1 dBFS reading and no OVR reading, because when you reach 0 dBFS exactly you have no idea how much it went over. Of course it is almost impossible for it not to have gone over if you get an OVR, since how many analog tape sources have exactly one sample's length of full peak level?
BK
RME User Forum → DIGICheck → Oversampled Digicheck meters at 192 kHz...
Powered by PunBB, supported by Informer Technologies, Inc.