mm lossy?

[*dJ*]

I just did a check for a show I got some time ago...

This is the result:

SA

FA


From the SA, I would say it's lossy, but what are those vertical stripes going to max. frequency?

I was told this was first recording of the taper and the microphone level was set to high, so the show has distortion, but could that explain the (abscence of?) stripes??? Maybe FA can tell more, but don't know really how draw conclusions form a FA.

Lineage:
aud>mz-r909>ecm-ds70p>cdr1

I have a flac copy of a cdr (think gen 1 or 2)

edit:- I can post some samples if needed
- There aren't real blocks???

[Five]

the fa here is displaying incorrectly... try clicking the cursor around after you open the fa window next time.

this was recorded on minidisc and the sa looks about right

altho md is lossy it is acceptable to most traders (and okay to seed here)

[*dJ*]

Originally Posted by Five

the fa here is displaying incorrectly... try clicking the cursor around after you open the fa window next time.

this was recorded on minidisc and the sa looks about right

altho md is lossy it is acceptable to most traders (and okay to seed here)

If I understand it correct the [almost] absence of the frequencies above 13 kHz doesn't make it real lossy [as opposed to MD lossy], so how can I/you tell this (or in general any) show is mp3 sourced?

Is it, not mp3 sourced, because the purple spikes going >20 kHz? does the FA give this information? Do you trust the line-age given and conclude from there on? -- I am a bit puzzled about this (and wouldn't mind to learn a bit more about it).

Also made a new FA from the same show (different track) it has a steep[er] drop-off... should I worry?

[Five]

well what is fairly easy to tell is lossy vs lossless sources. the steep cutoff in the fa and blocks visible in the zoomed in sa are the giveaways. pure black in the high frequencies of the sa is also an artifact of lossy encoding in the lineage.

when it comes to md it can be difficult to really tell the diff since it is lossy just like mp3 et al. the blocks often look a little more hyphen-like.

it is possible sometimes that an md recording has also been converted to mp3 this is one of the most difficult things to spot. your zoomed in sa is 2ms would be better if that was 2sec ( 0:02.000 in the view/length stat). If you want to upload a sample via yousendit.com post the link here and I can download it and have a look on my end.

One thing that is unfortunate about the show in question here is that it seems to have been recorded at a quite low quality setting and as a result the sound will be worse than the average mp3, which cuts off at 16kHz

[*dJ*]

well what is fairly easy to tell is lossy vs lossless sources. the steep cutoff in the fa and blocks visible in the zoomed in sa are the giveaways. pure black in the high frequencies of the sa is also an artifact of lossy encoding in the lineage.

I got that, but often the SA/FA aren't that clear-cut..

when it comes to md it can be difficult to really tell the diff since it is lossy just like mp3 et al. the blocks often look a little more hyphen-like.

I think this show is more hyphen-like, but that is more based on my feeling, but then why is MD sourced okay to most traders. Isn't it as lossy as mp3?

I know Hi-MD's do support less compressed and uncompressed (?) PCM formats, but I assume you are talking about the older generation of MD's?

it is possible sometimes that an md recording has also been converted to mp3 this is one of the most difficult things to spot.

that sucks..

your zoomed in sa is 2ms would be better if that was 2sec ( 0:02.000 in the view/length stat).

I see, must have been sleeping

If you want to upload a sample via yousendit.com post the link here and I can download it and have a look on my end.

I made 3 for you:
http://s46.yousendit.com/d.aspx?id=1...902TXCEBF0CKD1
http://s46.yousendit.com/d.aspx?id=2...J1IJN6D82XZ1D4
http://s46.yousendit.com/d.aspx?id=1...0U0EYWKKIGC2NG

Freq's of the las one go up on the end & beginning of the sample (If you make an SA you will see what i mean), those are the spikes I was talking about.

One thing that is unfortunate about the show in question here is that it seems to have been recorded at a quite low quality setting and as a result the sound will be worse than the average mp3, which cuts off at 16kHz

True, the sound isn't great..

As I said microphone level set to high and show is distorted. I just don't know what to do with it.

Should I consider this as lossless format, but note that a mp3 recording of this show had been better (although the higher freq's would have been more smoothed out in FA?)

[ssamadhi97]

Originally Posted by *dJ*

I know Hi-MD's do support less compressed and uncompressed (?) PCM formats, but I assume you are talking about the older generation of MD's?

Yes.

Originally Posted by *dJ*

I made 3 for you:
http://s46.yousendit.com/d.aspx?id=1...902TXCEBF0CKD1
http://s46.yousendit.com/d.aspx?id=2...J1IJN6D82XZ1D4
http://s46.yousendit.com/d.aspx?id=1...0U0EYWKKIGC2NG

Freq's of the las one go up on the end & beginning of the sample (If you make an SA you will see what i mean), those are the spikes I was talking about.

Seems the beginning and the end of the last sample are from an entirely different source and recording.

Originally Posted by *dJ*

Should I consider this as lossless format, but note that a mp3 recording of this show had been better (although the higher freq's would have been more smoothed out in FA?)

Actually, looking at the beginning of sample 1, this 1) clearly looks upsampled from 11025Hz sampling frequency (note the slight vertical mirroring), and 2) I suspect it was mp3 compressed before (note the blocky holes in the spectrum).

Relevant SA screenshot attached.

[Five]

^what he said

[*dJ*]

Damn!! That sucks.. (although I am still very happy to have this show!)

I remember the guy who send this show to me ripped it from CDR to Flac (using NERO) and he was a bit surprissed he could rip whole show in 4 minutes. Don't know if it matters (possible not), but maybe it wasn't ripped secure and that's reflected in FA/SA

Another member of this board also has this show (not from me), maybe he has a better version. Will send him a PM asking if he could post some samples as well, so we can find out more about this one (maybe I can upgrade)

[*dJ*]

While in wait for some more samples I decided to play around with some other shows I got.

I found this show, I am almost sure it's lossy (big drop and really blockish SA), but I was wondering where the missing parts of the carrier did go.. is this just due to the compression used?

[Five]

that is weird... it is lossy for sure, & sourced from tv/fm

perhaps somebody here can come up with a theory about what happened to the missing bits of the carrier

[range_hood]

Lossy encoding of an FM signal of a common tuner. The carrier is strong enough to be encoded. Some better tuners fade out that stripe so you get weaker ones that would get dropped entirely by the encoder.

[ssamadhi97]

Originally Posted by Five

perhaps somebody here can come up with a theory about what happened to the missing bits of the carrier

Considered inaudible by the encoder and discarded.

Probably because the pilot tone was a little bit weaker for a bit, or because the encoder decided that it's temporarily masked (drowned) by a stronger music signal in the frequencies below.

[*dJ*]

Originally Posted by ssamadhi97

Considered inaudible by the encoder and discarded.

Probably because the pilot tone was a little bit weaker for a bit, or because the encoder decided that it's temporarily masked (drowned) by a stronger music signal in the frequencies below.

music is indeed strong in that area, didn't noticed that before, thanks!!

I got another one:

lineage:
Core Sound Binaurals(Bass Rolloff @ 90Hz) » Sony MZ-R37 MD » DAT » Fostex D5(Optical) » m-audio DiO 2496 » CD Wave 1.94.4 beta » .FLAC » Goldwave(Convert to .wav(16 bit), normalize, split tracks) » FlacFrontend (Encode Level 6 and align sector boundries) » YOU!




The drop can that be due to MD?
Or the transfers? e.g. why wav>flac>wav>flac ?
Also i don't get the DAT (maybe taper means data??)

[*dJ*]

additional info:

Special shoutout to <edit name> for the digital MD transfer.

[Five]

yes, that looks exactly like a digitally-transferred MD recording. this is very distinctive spectral signature, with lego building-like "skyline", very much like I get when I transfer from my own MD recorder.