Seeking ME for an edited piano recital with audible artifacts

[wmgan]

Dear all,

This is my first time posting on this forum, and indeed, my first time seriously investigating professional mastering options.

I'm a student at the Bard College Conservatory of Music majoring in piano performance (and math and computer science), and I recently gave my senior recital, in which I performed Busoni's editions of Bach's D-minor Concerto, BWV 1052 with a string quintet (i.e. piano with five string instruments), and of the Goldberg Variations, BWV 988 (solo piano).

Both the concert and the dress rehearsal were recorded, and I decided to edit between the two to create a performance that was more satisfactory to me than either of the two. There are two problems:

(1) Due to different acoustic conditions (the presence/absence of an audience, the piano being ever so slightly more in tune in one performance than the other, and the possible repositioning of the mics), the piano has a slightly different timbre in each of the recordings (more so than the strings, though you can hear differences in the reverb). This means that you can hear the piano suddenly changing in tone at several of the edit points.

(2) Due to tempo differences between the two performances (I have a tendency to play everything so much fast than I intend to when playing in front of an audience), I was sometimes compelled to digitally speed up or slow down portions of one recording to match the tempos of the other. I tried to avoid doing this as much as possible, and to do it only by small amounts, but occasionally, this resulted in an audio artifact I would describe as a "wobble".

I uploaded two short samples of my edited recording to my web space so that you can hear what I mean:

http://www.amazming.com/temp/concerto_sample_edited.wav
http://www.amazming.com/temp/goldberg_sample_edited.wav

Each of these samples illustrate both problems, and are probably the most distracting examples (which is why I picked them). They are un-EQ'd, uncompressed 32-bit float WAV files at 44.1kHz.

I attempted to fix these problems myself by running the whole edited recording through a low-pass filter (since the higher frequencies seemed to be contributing most to the tonal inconsistency of the piano), adding reverb to compensate for the differences in reverb (but more reverb seemed desirable anyway since the mics were placed fairly close to the stage), and using stereo widening and an exciter to invigorate the sound of the strings. A quick fix that seemed adequate (barely) for the short-run duplication order I made so that I would have copies to give/sell to faculty members and friends who had been asking for them -- but nowhere close to what I would feel comfortable handing out to potentially influential contacts in the near future. (For those who are concerned about the ethics of handing out a recording labeled "live" that had in fact been edited -- the tray artwork I've designed clearly states that this is an edited recording.)

So, I'm asking the mastering engineers in this forum -- especially those who've had experience with classical (specifically piano) recordings -- if you could listen to the two samples I linked to above and give me a realistic estimate for what you can do to eliminate the aforementioned problems and otherwise improve the sound of this recording (I would love for the strings to sound more lively -- the exciter I used produced almost artificial-sounding results), and how much you would charge for such a project.

I have already taken a look at some of your studios' websites, but it has been difficult to get a pricing estimate since it doesn't really conform to the "how many songs" model. I have two pieces (a 23-minute piece for piano and strings, and a 40-minute piece for piano solo), each of which I will be splitting into three tracks on the CD. My budget is $200-300, but if I really love the results I may be able to stretch it, so please do PM me anyway even if you would have to ask for a lot more than that.

I'm sorry this message turned out so long. I'll stop now and wait for replies/PMs. I'll be happy to post more samples or the complete recording if anyone asks for them.

Thank you,
Ming.

[trmchenry]

PM'd

The timbre matching would most likely be easiest to do before any other processing is done.

You also have to be extremely careful when using any widening. The concerto example is too wide.

I'm just afraid that any processing you're doing might be more than what the recording needs.

[TranscendingM]

You should probally have the raw (before) clips of these 2 examples up as well. This way one can discern what the original situation was sonically and what your editing did for it.

[trmchenry]

Originally Posted by TranscendingM

You should probally have the raw (before) clips of these 2 examples up as well. This way one can discern what the original situation was sonically and what your editing did for it.

My thoughts exactly. The fidelity of the recording sounds nice but I can't help but think there's a better process for cleaning everything up from start to finish.

[wmgan]

Originally Posted by trmchenry

You also have to be extremely careful when using any widening. The concerto example is too wide.

Actually, the samples I posted were the final edits BEFORE I ran them through the filter chain. No widening, no EQ, no compression, nothing. The only processing I did was to digitally change the tempo of small portions of the recordings, as I mentioned in my original post (and converting it to 32-bit floating-point format, but that conversion should be lossless).

Originally Posted by TranscendingM

You should probally have the raw (before) clips of these 2 examples up as well. This way one can discern what the original situation was sonically and what your editing did for it.

Great idea! I just checked the samples I originally posted and realized that the Goldberg sample had a glitch where the levels suddenly spiked randomly -- so sorry about that! (Adobe Audition's "Save as Selection" command is evidently flawed.) This has been fixed in the re-uploaded version. Also, I decided to post a longer Concerto sample. Here are clips (of the exact same measures of music) of the unedited, unprocessed recordings of both the concert and dress rehearsal:

http://www.amazming.com/temp/concert...le_concert.wav
http://www.amazming.com/temp/concert..._rehearsal.wav
http://www.amazming.com/temp/concerto_sample_edited.wav

http://www.amazming.com/temp/goldber...le_concert.wav
http://www.amazming.com/temp/goldber..._rehearsal.wav
http://www.amazming.com/temp/goldberg_sample_edited.wav

(The filenames should be self-explanatory.)

The funny thing is -- having only listened to the edited version since I finished the editing a week ago, and going back now and listening to the original source recordings -- I realize that some of the artifacts I attributed to my editing and to the digital time-scaling were, in fact, audio imperfections that were present in the original recording. Huh.

Do listen to these newly posted clips and let me know what you think. Also do tell me if you hear anything that bothers you in the edited version that was clearly caused by the editing. If enough people don't feel like there's a problem, or if there isn't anything that can be done to substantially improve the sound of the edited version, perhaps I should leave well enough alone and focus on other things, like future projects.

[By the way, I got my short run order from cdpostershop.com today, and the quality is top notch. The quality of the color printing on the disc, in particular, exceeded all my expectations. I'm really glad I chose them over diskfaktory.com -- especially after what I read elsewhere in this forum about them stealing another studio's picture -- and will definitely be returning to them for future short-run orders. I cannot recommend them highly enough.]

[cameron1]

I also got some short run disc and poster from Diskfaktory. They were very professional on deliver my stuff on time. The quality of my CDs were great and the posters were OK( for the price i paid). I did know they had mastering there, but they seem to have all artist services under 1 roof. Next time, I'd check them out. Overall, they've done a good for me and couple friends....

[TranscendingM]

With the concerto examples I was surprised about something. First of all, they aren't immensely different. But here's my take on them. In the unedited version, the piano's attack sounds more natural to me. The initial attack doesn't sound overly "excited" like in the "edited". But the strange thing is, in the edited version the strings sound more mellow. Mostly in the staccato you can hear that. Eh, may be it's just a performance thing or may be certain parts I heard as such were parts you edited together and had slightly different feels. With the Goldberg, for me at least, I heard more distortions here and there like at about 9 secs. Not sure what it is, could be the "wobble" thing you mentioned but sounds more pronounced in the edited version. Overall though, for me, I wouldn't say the acoustic space between joined parts is staggering in difference.

[trmchenry]

The biggest difference I heard between the two is the overall wideness. The tones are damn near identical beyond that. You do hear a TINY bit more of the ambient reflections in the dress rehearsal copy. Honestly, I don't know that the average listener would ever pick up on something that minute. I'd be very hesitant to correct it with reverb. The concert performance is a bit more dramatic/dynamic sounding.

[wmgan]

Originally Posted by TranscendingM

In the unedited version, the piano's attack sounds more natural to me. The initial attack doesn't sound overly "excited" like in the "edited". But the strange thing is, in the edited version the strings sound more mellow.

My best explanation would be that, in the portions where I sped up the audio (it's almost always the dress rehearsal recording that got sped up by about 5% to match the tempo of the concert recording), the corresponding changes in attack and decay times altered the overall tone of the recording enough to make it sound jarringly different. That may be the "excited"-ness you're talking about.

The most jarring (to me) jump in timbre in the concerto is in that section where the strings drop out and the piano has a solo for awhile -- as the piano begins its descending sequence, there is an edit from one recording to another, and you can hear a dramatic shift in timbre.

Originally Posted by trmchenry

Honestly, I don't know that the average listener would ever pick up on something that minute.

That's kind of what I'm hoping. I've given out a couple copies of my CD, and will ask for feedback. Knowing where the edits definitely doesn't help my objectivity, and I've only listened to these on (pretty good) headphones because my laptop's speakers are crap and I don't have a good pair of monitors -- the headphones probably amplify all the audio imperfections.

Thank you to those of you who PM'd me -- I'm sorry I haven't been able to answer most of your PMs yet. This just happens to be an incredibly busy time for me, with finals and all -- plus I'm getting married tomorrow. But I'll try to respond to everyone in the next couple days.

[trmchenry]

Sounds like a busy time indeed.

When things settle down and you have some extra time try taking the effected versions I sent to you and redo the particular edit you're talking about and see if it's still as jarring.

Personally, I find edits easiest to do on headphones because they do seem to amplify those imperfections. If edits, fades, and transitions sound good on headphones they are ready to print.